Which AI companies serve manufacturing, and what they do
Eight layers, from the sensor on a machine out to the suppliers and customers around a factory.
The layers, from the machine outwards
Software for a factory can sit at very different distances from the machine. At one end it reads a sensor and acts within seconds; at the other it helps a planner, a worker, or a whole chain of companies decide over weeks or years.
- horizon
- milliseconds to seconds
- in
- spindle vibration sampled at 10 kHz, motor current, coolant temperature on a CNC mill
- out
- tool-wear alert about 20 minutes before the tool breaks; a soft sensor estimating melt viscosity from screw torque
- decision
- maintenance swaps the tool at the next part change instead of after a crash
- core stack
- ML 9 · edge / IIoT 5 · data eng 1 (of 15)28 open tech roles at 13 companies: ML / data science 32%, OR 0%
data and standards
- reads / writes
- reads sensor signals and PLC tags · writes edge alerts and time series into the historian (LX)
- standards
- OPC UA, MQTT, MTConnect, umati, SECS/GEM in fabs, IPC-CFX in electronicsopen shared standardMostly an open standard: any tool can read machine data through OPC UA or MQTT, but what each value means, and where to put the sensor, is worked out by hand in every plant.
- horizon
- seconds to minutes
- in
- kiln inlet temperature, oxygen in the flue gas, fuel feed rate, free-lime lab result every two hours
- out
- new setpoints for fuel rate, fan speed and kiln speed every few minutes, written to the control system
- decision
- the controller trims fuel by a small step while keeping clinker quality inside spec
- core stack
- ML 8 · control 7 · physics + ML 2 (of 20)35 open tech roles at 14 companies: ML / data science 23%, OR 3%
data and standards
- reads / writes
- reads DCS tags, analysers, lab results · writes setpoints and recipes into the DCS or PLC
- standards
- OPC UA write access, PackML, ISA-88 / BatchML recipes, SECS/GEM remote commandsmixed: open shared standarda big vendor's systemSplit: sending a new setting to a machine has an open standard, but whoever owns the plant's control system (the DCS) decides who may send one.
- horizon
- per part
- in
- a camera image of every painted panel, the pressure curve of every injection-moulding shot
- out
- pass or fail per part with a defect class (scratch, inclusion, short shot); predicted scrap rate for the shift
- decision
- the part goes to the rework lane; the process engineer is told which cavity drifts
- core stack
- CV 16 · ML 1 (of 17)19 open tech roles at 8 companies: ML / data science 21%, OR 0%
data and standards
- reads / writes
- reads images, per-part curves, measurements, CAD with PMI · writes pass or fail and defect class to the PLC or MES
- standards
- QIF for metrology, STEP AP242 for the product definitionmixed: open shared standardno shared formatSplit: measurements and 3D design data have open formats, camera images and defect classes have none.
- horizon
- shift to day
- in
- run, idle and down states with reason codes from 40 machines, batch records, shift logs, quality results from the MES
- out
- OEE per line and a ranked cause list, for example "most micro-stops on line 3 follow a change of film supplier"
- decision
- the plant manager changes the changeover procedure and checks the effect next week
- core stack
- edge / IIoT 6 · ML 2 · data eng 2 (of 13)7 open tech roles at 9 companies: ML / data science 0%, OR 0%
data and standards
- reads / writes
- reads machine states with reason codes, MES records, shift logs · writes OEE and downtime reports, sometimes downtime records into the MES
- standards
- ISO 22400 KPIs, VDMA 66412-1, PackML state model, ISA-95 / B2MMLmixed: open shared standardno shared formatOn paper only: standards define the OEE formula, but each plant decides which stops count and how to label them.
- horizon
- days to weeks
- in
- open orders with due dates, routings, machine capacity, stock levels, supplier lead times from the ERP
- out
- a two-week sequence per machine, purchase-order proposals, a delivery date to promise a customer
- decision
- the planner accepts a rush order and moves two low-priority jobs back a day
- core stack
- OR 12 · data eng 9 · ML 7 (of 35)31 open tech roles at 23 companies: ML / data science 13%, OR 6%
data and standards
- reads / writes
- reads ERP master data (BOMs, routings, orders, stock, lead times) and MES order status · writes schedules and plans back to ERP or MES, purchase-order proposals
- standards
- ISA-95 / B2MML schedule schemas, EDIFACT DELFOR at the company boundarymixed: a big vendor's systemno shared formatMostly the vendor's: orders, stock and parts lists sit in each ERP vendor's own data model. Rules such as which product may follow which have no format; they are in the planner's head.
- horizon
- weeks to years (training, documentation); seconds when a worker asks
- in
- 2,000 pages of manuals and SOPs, past 8D reports, audit checklists, a worker's question typed on a tablet
- out
- a step-by-step instruction with photos; an answer with a page citation such as "25 Nm, manual p. 112"; training gaps per person
- decision
- the shift lead assigns a trained operator; the changeover is done right the first time
- core stack
- workflow software 6 · RAG 4 · agents 3 (of 18)24 open tech roles at 12 companies: ML / data science 0%, OR 0%
data and standards
- reads / writes
- reads manuals, SOPs, 8D reports, checklists, some MES context · writes work instructions, training and audit records
- standards
- none foundno shared formatNowhere: the know-how is in documents and in people. No standard format for work instructions turned up.
- horizon
- weeks to months (Catena-X DCM: weekly buckets up to 24 months)
- in
- supplier capacity and lead time by tier, customer forecasts and call-offs, transit times, and a scenario such as "supplier X loses a plant for six weeks" or "demand at customer Y rises 20%"
- out
- simulated stock, service level and cost across the whole chain under each scenario: which parts run short, in which week, at which company
- decision
- dual-source a part, hold four extra weeks of stock, or reserve next year's capacity with a supplier
- core stack
- data eng 3 · OR 2 · ontology 1 (of 7)49 open tech roles at 6 companies: ML / data science 14%, OR 10%
data and standards
- reads / writes
- reads customer forecasts and call-offs by EDI, supplier capacity where shared, own ERP · writes scenarios, buffer and sourcing decisions, capacity reservations sent on to suppliers
- standards
- EDIFACT DELFOR, X12 830, VDA 4905 / 4984, Catena-X CX-0128, GS1, EU battery passportmixed: open shared standardno shared formatCustomers have sent forecasts to suppliers in open formats since 1988. The one format found for suppliers to send back their capacity is Catena-X CX-0128, car industry only. Scenario models have none.
Few do this. 7 companies on the map show, on a page we read, that they model suppliers or customers and play scenarios over them; 4 of them are large planning or supplier-network vendors. Most tools sold as "supply network" work one hop out (procurement, supplier portals, risk alerts) and sit in L4. Doing this well needs data from companies that do not share it, so for a small manufacturer it is still hard to do today.
7 companies with this as their main layer →- horizon
- not a decision level; continuous ingestion
- in
- PLC tags over OPC UA and MQTT, historian archives, MES and ERP tables
- out
- one clean, named data stream per machine (a unified namespace) and a place to run and monitor models at the edge
- decision
- IT and OT decide which plant data leaves the site and where a model runs
- core stack
- edge / IIoT 13 · data eng 5 · ML 2 (of 22)30 open tech roles at 14 companies: ML / data science 20%, OR 0%
data and standards
- reads / writes
- reads PLC and sensor tags over OPC UA or MQTT, historians, MES and ERP tables · writes the historian, unified namespace or data lake every layer reads
- standards
- OPC UA, MQTT, Sparkplug, Asset Administration Shell, ISA-95 hierarchyopen shared standardOpen standards for moving the data, none for describing it: no shared model says what each data point in a plant means.
Which manufacturers show up most
65 of 147 companies have at least one tie to a large manufacturer, 111 ties in all (49 invested, 4 JV, 32 founder from, 1 acquired by, 25 partner). The list ranks the manufacturers by how many companies on the map they are tied to; a company tied in two ways counts once in the total and once under each kind.
A tie is one of six kinds with a large manufacturer or industrial group (more than 10,000 employees or more than USD 5bn revenue) or its venture arm or subsidiary: it invested, a joint venture, the company is a spin-off, a founder worked there, it acquired the company, or a formal partnership with a joint announcement or a co-built product. Customers alone do not count. Left out: utilities and energy suppliers, trading houses, mining and oilfield services, research institutes, fabless chip and outsourced-assembly designers, multi-company funds, and partner claims that rest only on a logo, a marketplace or ecosystem listing, or the startup's own partner page.
- Siemens 6: 2 invested, 2 founder from, 2 partner
- Bosch 4: 1 JV, 3 founder from, 1 partner
- Omron 3: 2 invested, 1 partner
- Saudi Aramco 3: 3 invested
- ABB 2: 2 invested, 2 partner
- Cisco 2: 2 founder from
- Continental 2: 1 invested, 1 JV
- Dell 2: 2 invested
- Emerson 2: 1 invested, 1 founder from
- Fortive 2: 1 invested, 1 partner
- Foxconn 2: 1 invested, 1 founder from
- GE 2: 2 founder from
- Henkel 2: 1 invested, 2 partner
- Intel 2: 1 invested, 1 partner
- Mercedes-Benz 2: 2 founder from
- Nortel 2: 2 founder from
- Rockwell Automation 2: 2 invested, 1 partner
- Toyota 2: 1 invested, 1 partner
60 more firms appear with one company each
AB InBev (founder from), Airbus (invested), Aisin (invested), Amoco (founder from), ArcelorMittal (invested), AUO (invested), Boeing (invested), Canon (invested), CETC (invested), Chevron (invested), China Baowu (partner), CRRC (invested), Delta Electronics (founder from), Ecolab (invested), Festo (founder from), Flex (founder from), Ford (founder from), Fujitsu (invested, partner), Galanz (founder from), Heraeus (acquired by), Honeywell (invested), Huawei (partner), Hyundai Motor (founder from), Infineon (partner), Juniper (invested), Kawasaki Heavy Industries (partner), Keyence (founder from), Koch (founder from), Konecranes (partner), Lenovo (invested), LG Display (founder from), LG Energy Solution (invested), Lockheed Martin (partner), Meidensha (invested), Nippon Steel (partner), Northrop Grumman (founder from), NSK (invested), OPPO (partner), Panasonic (partner), Phoenix Contact (invested), Porsche (invested), Renault (founder from), Repsol (invested), Rizhao Steel (invested), S-OIL (invested), Samsung (invested), Sandvik (partner), Schaeffler (JV), Seagate (partner), Sensata (invested), Shimadzu (invested), Shin-Etsu Chemical (founder from), Sony (invested), Tata Steel (founder from), Texas Instruments (founder from), Tyson Foods (invested), Weir Group (partner), Yara (invested), Yokogawa (invested), ZF (JV)
Where the standard lives
A decision's data format lives in one of three places: an open standard, a big vendor's system, or no shared format. Where an open standard exists, a new tool can plug in quickly. Where one vendor owns the record or the control interface, such as the DCS (distributed control system) that runs a process plant's valves, pumps and burners, that vendor tends to keep the margin. Where nothing exists, every customer is a custom integration project.
Terms used here (62), one line each
Systems, data and shop-floor words
- DCS
- distributed control system: the computer system that runs a process plant's valves, pumps and burners.
- PLC
- programmable logic controller: the small industrial computer that controls one machine.
- MES
- manufacturing execution system: tracks orders and what each line is producing.
- ERP
- enterprise resource planning: the company's main business system for orders, stock, purchasing and parts lists.
- historian
- a time-series database of sensor readings from the plant.
- tag
- one named data point on a machine, such as a temperature or a motor current; roughly a variable.
- setpoint
- the target value a controller holds, such as an oven temperature or a line speed.
- recipe
- the set of machine settings for making one product.
- OEE
- overall equipment effectiveness: one score for how much of the possible output a line actually produced.
- reason code
- the label recorded for why a machine stopped, such as no material, changeover or breakdown.
- changeover
- switching a machine from one product to the next; the time it takes is lost output.
- EDI
- electronic data interchange: the decades-old message formats companies use to send each other orders and forecasts.
- call-off
- a customer's message telling a supplier how many of each part to deliver and when, usually with a forecast for the weeks after.
- BOM
- bill of materials: the parts list of a product.
- routing
- the sequence of steps and machines a product goes through.
- unified namespace (UNS)
- one shared tree of named live plant data that every system publishes to and reads from, like a message bus with agreed topic names.
- edge
- computing on site, next to the machine, instead of in the cloud.
- CAD with PMI
- the 3D design model, with product and manufacturing information (tolerances, surface finish) attached.
- SOP
- standard operating procedure: a written step-by-step instruction.
- 8D report
- a standard eight-step problem-solving report for a quality defect.
- APC
- advanced process control: software that keeps adjusting a process plant's setpoints.
- WMS, TMS
- warehouse and transport management systems.
- fab, 300 mm
- a fab is a chip factory; 300 mm is the diameter of the silicon wafers its tools handle.
- OEM, tier 1
- in car making, the vehicle maker, and a supplier that delivers to it directly.
- LLM
- large language model: the kind of AI behind chat assistants, able to read unstructured documents.
- CNC
- computer numerical control: a machine tool, such as a mill or lathe, driven by a program.
- MRP
- material requirements planning: the ERP calculation of what to buy or make, and when.
- IT, OT
- information technology (office and business systems) and operational technology (the systems that run the plant).
- CV
- computer vision: software that interprets camera images.
- RAG
- retrieval-augmented generation: an AI language model answering from documents it looks up first.
- IC, front-end
- IC: integrated circuit, a chip. Front-end tools are the ones that build the circuits on the wafer.
Standards named in this section and on the layer cards
- OPC UA
- a common protocol for reading and writing machine data, with typed data models.
- MQTT
- a lightweight publish-subscribe message protocol; it does not say what a message means.
- Sparkplug
- fixed topic names and payload format on top of MQTT, so devices and apps understand each other.
- umati
- a shared OPC UA data model for machines, machine tools first.
- Asset Administration Shell
- a standard digital description of a machine or product, also used for EU product passports.
- SECS/GEM, GEM300
- the chip industry's standard for how fab tools talk to the factory's host computer; GEM300 is the suite for 300 mm tools.
- IPC-CFX, Hermes
- messaging standards for electronics assembly lines; Hermes covers handing a board from one machine to the next.
- PackML
- one shared set of machine states (such as held, blocked, starved) and tags, first for packaging machines.
- ISA-88, BatchML
- a standard model for batch recipes and batch production records.
- ISA-95, B2MML
- a standard model of what passes between the ERP and the MES (orders, schedules, materials); B2MML is its XML and JSON schemas.
- QIF
- an XML format for measurement plans and measurement results.
- STEP AP242
- a neutral 3D design format that carries tolerances with the geometry.
- ISO 22400, VDMA 66412-1
- international and German definitions of factory KPIs, OEE among them.
- EDIFACT, DELFOR
- the UN's family of EDI messages; DELFOR is its delivery forecast message.
- X12 830
- the North American EDI message for a rolling forecast from buyer to supplier.
- VDA 4905, VDA 4984, Odette
- the German and European car industry's EDI formats for delivery call-offs; VDA 4984 replaces 4905 and is built on DELFOR.
- Catena-X, CX-0128
- the car industry's shared data network; CX-0128 is its standard for exchanging demand and capacity (DCM).
- GS1, GTIN
- the body behind the retail barcode; the GTIN is the product number it encodes.
- EU battery passport, DPP, ESPR
- machine-readable product records required by EU law: for batteries first (Regulation 2023/1542), other products under ESPR (digital product passports).
Who publishes them
- SEMI
- the global industry association of semiconductor equipment and materials makers; it publishes SECS/GEM.
- VDA
- the German Association of the Automotive Industry.
- VDW, VDMA
- the German Machine Tool Builders' Association, and the German Engineering Federation.
- OMAC
- the Organization for Machine Automation and Control; it publishes PackML with ISA.
- ISA
- the International Society of Automation, which writes ISA-88, ISA-95 and PackML's technical report.
- MESA
- MESA International, an association of MES vendors and users; it publishes B2MML.
- IPC
- the Association Connecting Electronics Industries; it publishes IPC-CFX.
- Odette
- the European car industry's organisation for EDI messages and their transport.
- CEN-CENELEC JTC 24
- the European standards bodies' joint committee writing the harmonised standards for the passports.
- EMO Hannover
- a trade fair for machine tools.
OPC UA (reading and writing machine data), MQTT (messaging between devices), SECS/GEM (chip factories' tool interface), EDIFACT / VDA EDI (orders and forecasts between companies), GS1 (product numbers and barcodes)
for whoever builds on top Anyone can plug in on the same terms, so the interface itself earns its owner little. Chip factories show the upside: with GEM300, connecting a new tool went from "weeks of custom interface coding" to plug and play. community.criticalmanufacturing.com
the ERP's data model (ERP: the company's main business system for orders, stock and purchasing), the DCS, and who may write to it (DCS: the control system that runs a process plant)
for whoever builds on top Owned, not agreed. A new tool goes through the owner of the record to plug in, and that owner tends to end up with the optimisation built on top of it.
rules of thumb on the shop floor (L5), which product may follow which on a machine, and how long the switch takes (L4), which defect is which in a camera image (L2)
for whoever builds on top Each deployment starts by collecting the data by hand, so delivery leans on services or on reading documents. Whoever gets a format adopted would own it, but only if competitors adopt it too.
Layer by layer
For each layer: which standards already exist, and in plain words where the agreement lives.
| layer | standards that already exist | where it lives |
|---|---|---|
| L0 Machine & sensor | OPC UA (protocol for reading and writing machine data), MQTT (lightweight publish-subscribe messaging), MTConnect , umati (shared OPC UA data model for machines), SECS/GEM in fabs (chip industry's standard for talking to fab tools), IPC-CFX in electronics (messaging on electronics assembly lines) | open shared standard Mostly an open standard: any tool can read machine data through OPC UA or MQTT, but what each value means, and where to put the sensor, is worked out by hand in every plant. |
| L1 Machine configuration & control | OPC UA write access (protocol for reading and writing machine data), PackML (shared machine states and tags), ISA-88 / BatchML recipes (standard model for batch recipes), SECS/GEM remote commands (chip industry's standard for talking to fab tools) | mixed: open shared standarda big vendor's system Split: sending a new setting to a machine has an open standard, but whoever owns the plant's control system (the DCS) decides who may send one. |
| L2 Output & product | QIF for metrology (XML format for measurement plans and results), STEP AP242 for the product definition (neutral 3D design format with tolerances) | mixed: open shared standardno shared format Split: measurements and 3D design data have open formats, camera images and defect classes have none. |
| L3 Line & plant monitoring, causal analysis | ISO 22400 KPIs (international definitions of factory KPIs), VDMA 66412-1 (German definitions of factory KPIs, OEE among them), PackML state model (shared machine states and tags), ISA-95 / B2MML (how business and production systems exchange data) | mixed: open shared standardno shared format On paper only: standards define the OEE formula, but each plant decides which stops count and how to label them. |
| L4 Planning | ISA-95 / B2MML schedule schemas (how business and production systems exchange data), EDIFACT DELFOR at the company boundary (UN EDI message for delivery forecasts) | mixed: a big vendor's systemno shared format Mostly the vendor's: orders, stock and parts lists sit in each ERP vendor's own data model. Rules such as which product may follow which have no format; they are in the planner's head. |
| L5 Knowledge & people | none found | no shared format Nowhere: the know-how is in documents and in people. No standard format for work instructions turned up. |
| L6 Supply chain simulation beyond the company | EDIFACT DELFOR (UN EDI message for delivery forecasts), X12 830 (North American EDI message for rolling forecasts), VDA 4905 / 4984 (car industry EDI for delivery call-offs), Catena-X CX-0128 (car industry network; shares demand and capacity), GS1 (product numbers and barcodes), EU battery passport (EU-required digital record per battery) | mixed: open shared standardno shared format Customers have sent forecasts to suppliers in open formats since 1988. The one format found for suppliers to send back their capacity is Catena-X CX-0128, car industry only. Scenario models have none. |
| LX Infrastructure & data | OPC UA (protocol for reading and writing machine data), MQTT (lightweight publish-subscribe messaging), Sparkplug (fixed topics and payloads on top of MQTT), Asset Administration Shell (standard digital description of a machine), ISA-95 hierarchy (how business and production systems exchange data) | open shared standard Open standards for moving the data, none for describing it: no shared model says what each data point in a plant means. |
Each standard links to a page it was checked on; hover for the body and the year. Every acronym is spelled out under "Terms used here" above.
What drives standardization
Three things push an industry onto a standard: a big buyer that will not buy without it, a law, or a group of companies in one industry agreeing among themselves. In 5 of the 7 cases below, the push came from buyers. The three clearest:
Chip factories told equipment makers: every new 300 mm tool must speak the same language (SEMI GEM300), or we will not buy it. Connecting a new tool went from weeks of custom code to plug and play.
evidence "By the early 2000s, all major IC manufacturers were requiring 300 mm equipment to comply with this GEM 300 standard suite." Integration went from "weeks of custom interface coding" and "one-off adaptations for each vendor" to plug and play. A vendor page states 100% of front-end tools support GEM.
who Major IC manufacturers, through SEMI standards
US grocers and food makers agreed on one product number and one barcode for every product, used by every company that handles it.
evidence "Over 1 billion products now carry GS1 barcodes that are scanned billions of times every day" (GS1, 2023).
who US grocery industry leaders (retailers and manufacturers); Uniform Code Council as administrator
Car makers including BMW, Mercedes-Benz and Volkswagen set up a shared data network with their suppliers, with common formats for demand and capacity, quality and battery passports. BMW then made joining it part of how it buys.
evidence 1,069 OEMs and suppliers exchange data through Catena-X (21 September 2026). "Since April 2025, Catena-X registration has been an integral part of the BMW Group's procurement process." DCM-certified solutions from SAP, o9 and CatX Services (January 2025).
who Catena-X Automotive Network e.V. (BMW, Mercedes-Benz, Volkswagen Group, Porsche, Renault, Volvo, suppliers, SAP and others)
4 more cases
German car makers, through their association VDA, agreed on one message format for telling suppliers how many of each part to deliver to which plant, with a forecast attached.
evidence "Automotive OEMs require every tier 1 supplier to use an electronic data interchange (EDI) system" (Orderful, an EDI vendor). VDA 4905 has been replaced by VDA 4984, built on EDIFACT DELFOR.
who VDA (vehicle makers' association), Odette International
German machine tool builders agreed on one OPC UA data model, so any machine tool can talk to any factory software.
evidence EMO Hannover 2019: 70 partners from ten countries connected 110 machines to 28 software services.
who VDW (machine tool builders), later VDMA, with the OPC Foundation
Large buyers of packaging machines, P&G and Nestlé among them, pushed one shared set of machine states and tags, so a line reads every machine the same way.
evidence P&G gave OMAC its PackML Implementation Guide and software template royalty-free; Food Engineering (October 2024) says P&G, Nestlé, Arla and Boeing "drove the effort for wider international acceptance"; a Nestlé engineer chaired OMAC in 2022.
who OMAC with ISA; Procter & Gamble, Nestlé, Arla Foods, Boeing
en.wikipedia.org digitaledition.foodengineeringmag.com automationworld.com
EU law requires a machine-readable passport for each battery in scope from 18 February 2027: identity, performance, recycling and sustainability data, read by QR code.
evidence Not yet in force. BMW already runs a digital battery passport use case on Catena-X (December 2025). Batteries are the first DPP group under ESPR; steel, textiles, tyres and others follow in the 2025 to 2030 working plan.
who European Union; CEN-CENELEC JTC 24 for harmonised standards
single-market-economy.ec.europa.eu bmwgroup.com oneclicklca.com
What should be standardized next
- L6 Suppliers telling customers how much they can make
A customer can send its suppliers a forecast in a standard format, and has done so for decades. A supplier has no standard way to send back how much it can make in the coming months, except in car making, through Catena-X.
Forecast formats: VDA 4905 (German car industry, edition 3 dated October 1988), today also EDIFACT DELFOR and X12 830, the UN and North American EDI messages. Capacity going back up: one standard found, Catena-X CX-0128, weekly figures up to 24 months ahead, automotive only. SAP, o9 and CatX Services presented certified software for it in January 2025.
This is the gap this map was built to show: 7 of 147 companies sit on L6, 4 of them established.
ecom-specialist.com seeburger.com comparatio.com catenax-ev.github.io catena-x.net
- L4 Rules for what can run after what
The ERP holds the orders, parts lists and machine calendars. Rules like "no black run after white without a two-hour clean" (an illustration) are written down nowhere a scheduling tool can read them.
ISA-95's schemas (B2MML) cover a production schedule, what a plant is able to make, and a work calendar (MESA's GitHub repository). No standard was found for setups, changeovers and the order in which products may run.
- L3 One way to count lost time
VDMA 66412-1 (2009) and ISO 22400 (from 2014) define OEE on paper.
- L5 Shop-floor know-how
Know-how like "machine 2 runs slow on Mondays" (an illustration) lives in people. No format standard for work instructions or shop-floor knowledge turned up.
Who sits where
Each company appears once, on its main layer; tick the box to also see it, faded, on its secondary layers. The chip style says how large the company is and how public its customer record is, by the rules in the legend.
Few do this. 7 companies on the map show, on a page we read, that they model suppliers or customers and play scenarios over them; 4 of them are large planning or supplier-network vendors. Most tools sold as "supply network" work one hop out (procurement, supplier portals, risk alerts) and sit in L4. Doing this well needs data from companies that do not share it, so for a small manufacturer it is still hard to do today.
Layer × who closes the loop
Each dot is one company at its main layer. The vertical position says who acts on the output: the machine itself at the top, a human reading a report at the bottom.
Companies by layer
| Amiral Technologies | L0 | DiagFit, software that learns normal behaviour from equipment sensor time series and flags anomalies and fault diagnoses for predictive maintenance. I-care Group bought the company's assets (DiagFit plus research and sales teams), announced 2026-09-17, to fold DiagFit into its I-see platform. | any manufacturing | MLwhyAll four papers on their scientific publications page are about machine learning anomaly and fault detection on industrial time series without past failure data (blind mode), and the careers page names data scientists as a core team; no open roles to weigh. hiring 0 open tech roles amiraltechnologies.com | detailfunding EUR 2.8m, as of 2021-10-29 (the October 2021 round led by Kreaxi with Accurafy4 and BACS-Innov; no all-time total found) satt.fr Raised EUR 2.8m in 2021, then I-care Group bought its assets (DiagFit plus research and sales teams) in an announcement dated 2026-09-17, to fold DiagFit into I-care's I-see platform; price undisclosed. sources satt.fr interestingengineering.com | - | Customer track record | reads sensors | 16 LinkedIn 11-50 range, 16 members listed | amiraltechnologies.com linkedin.com interestingengineering.com |
| Falkonry | L0L3 | Time Series AI platform that ingests PLC and sensor data for multivariate anomaly detection, pattern recognition and root-cause analysis at the edge and in the cloud. | any manufacturingsourceNamed customers are steelmakers (Stelco, Ternium, Nucor, AM/NS Calvert), the US Navy and Air Force, and energy producers (Chesapeake, Civitas). falkonry.com | ML · streaming · edge / IIoTwhyIts nine technical papers are all self-supervised or unsupervised time series anomaly detection (for example convolutional variational autoencoders), run on high-speed PLC and sensor streams ingested via Modbus and OPC UA at the edge; of 3 open roles, 1 is ML and AI engineer. hiring 3 open tech roles: ML 1, software 1, solutions / deployment 1 falkonry.com | USD 10k to 100kdetaildeal size USD 50k · per 12-month contract, 50 machines or line segments (plus USD 20k one-off activation). AWS Marketplace list price for Falkonry Clue: USD 50,000 for a 12-month contract covering 50 entities (machines, line segments or process areas) plus a one-time USD 20,000 activation fee; 24-month terms advertised at up to 50% savings aws.amazon.com funding USD 20.5m, as of 2024-11-14 cbinsights.com Public list price works out to about USD 1,000 per monitored machine or line segment per year plus a USD 20k setup fee, on USD 20.5m raised over 11 rounds since 2012. sources aws.amazon.com cbinsights.com | Fortive (invested) · Siemens (invested)detailFortive · invested Fortive joined the USD 4.6m Series A (2018) as a new strategic investor; its VP Strategic Development took a board observer seat. news.crunchbase.com | Customer track record | reads sensors | 30 LinkedIn 11-50 band (midpoint used) | falkonry.com falkonry.com linkedin.com |
| Feelit Technologies | L0 | Printed nanotech sticker sensors with a wireless edge device and analytics that watch pipes, valves, steam traps, pumps and heat exchangers for leaks, thinning and deformation. | any manufacturingsourceSolution page lists industries served as OEM manufacturers (automotive, industrial equipment, tooling, robotics), oil and gas, food and beverage, chemicals, pharma and infrastructure. feelit.tech | edge / IIoT · MLwhyThe core is a printed nano-ink sticker sensor invented by its nanotech founders, paired with a wireless edge device and failure-prediction analytics per its homepage; the careers page lists 1 MLOps engineer and 2 sales engineers as its technical openings. hiring 3 open tech roles: ML 1, solutions / deployment 2 feelit.tech | detailfunding USD 17.4m, as of 2025-09-16 cbinsights.com USD 17.35m raised over 7 rounds, the latest an USD 8.3m Series A+ in September 2025 led by Alea Capital with Henkel Ventures; 25 employees per Calcalist; hardware sensors plus edge devices plus analytics put margins below pure software, and no price or revenue is public. | Henkel (invested, partner) · Continental (invested)detailHenkel · invested Henkel Tech Ventures invested in January 2021 (Series A) and Henkel Ventures continued in the USD 8.3m Series A extension of September 2025 led by Alea Capital Partners. henkel.com Henkel · partner Feelit's CEO called Henkel a strategic partner at the 2021 investment, a collaboration opening oil and gas opportunities; Henkel is also a deployment site. henkel.com | Customer track record | reads sensors | 32 LinkedIn 11-50 range, 32 members listed | feelit.tech linkedin.com |
| Timeseer.AI | L0LX | Sensor data quality platform that detects, validates and fixes bad OT time-series before it reaches analytics or AI. | process industriessourceIndustries section lists utilities, life sciences, chemicals, oil and gas, and sensor manufacturers. timeseer.ai | data eng · MLwhyCore is a data-quality layer for OT time series: more than 100 prebuilt sensor-agnostic quality metrics (variance drift, broken correlations, stale or missing data) plus remediation such as imputation and metadata enrichment, per Verdantix, with AI anomaly detection on top per its homepage. | USD 10k to 100kdetaildeal size USD 49k · per year, entry Scale tier. AWS Marketplace list prices for a 12-month contract: Scale tier USD 49,000 with 4,000 tokens included (capped at 15,000), Enterprise tier USD 99,000 with 7,000 tokens included; more tokens by private offer aws.amazon.com funding USD 6m, as of 2022-01 verdantix.com competes with AVEVA, Cognite, Hitachi Vantara Public list price is USD 49k to 99k per year by token tier on AWS Marketplace, sold as SaaS through the Azure and AWS marketplaces; founded by the former TrendMiner CEO, with a USD 6m seed in 2022. | GE (founder from)detailGE · founder from Co-founder and CRO Niels (surname not given on the page) is a former Account Executive at GE, before senior sales roles at TrendMiner. timeseer.ai | Customer track record | reads sensors | - | timeseer.ai linkedin.com |
| CyberInsight (天泽智云) | L0L3LX | Edge intelligence hardware (iEC, iGW gateways) and the EdgePro platform for equipment condition monitoring, anomaly detection, health management and remaining-life prediction on industrial machines. | any manufacturingsourceRecord names wind power and other renewables, rail, and discrete manufacturing equipment OEMs as its markets. pitchhub.36kr.com | ML · ontology · agentswhyOf 13 open technical roles on Liepin, 6 are industrial-AI or time series algorithm engineers or leads, 2 are large-model or multimodal algorithm roles and 1 is an industrial knowledge modelling engineer; the current site leads with a rotating-equipment health diagnosis Agent built on existing DCS data with drift-warning and online-learning trend algorithms, plus an LKM industrial knowledge hub and an industrial AI data annotation platform. hiring 13 open tech roles: ML 6, LLM / AI 2, software 3, solutions / deployment 1, other 1 liepin.com | detailNo round amounts are public (seed, angel and a 2020 A+ from Foxconn per 36Kr); the IoT ONE profile puts annual revenue in a USD 10m to 100m band, and the company's own about page describes delivery by a team of process engineers, equipment experts and annotation engineers alongside product. | Foxconn (invested) · Lenovo (invested)detailFoxconn · invested Foxconn Industrial Internet (FII) invested in 2019, holding 20 percent; 36Kr lists Foxconn Technology Group in the A+ round of February 2020. Co-founder Liu Zongchang joined FII in 2019 as chief data officer and became its rotating CEO in June 2025. tianyancha.com Lenovo · invested Lenovo Capital (联想创投) invested in the December 2016 seed round and again in a May 2018 share transfer alongside IDG and Gaorong, per 36Kr PitchHub. pitchhub.36kr.com | Early | reads sensors | 120 | pitchhub.36kr.com asiagrowthpartners.com |
| Elmodis | L0L3 | Hardware and software that monitors industrial electric machines by sampling motor current and voltage at high frequency and deriving performance and health KPIs with ML. | any manufacturingsourceIndustries menu lists utilities, manufacturing and OEM, energy, and electric transportation; it is organised by asset type (pumps, fans, motors, wind turbines). elmodis.com | edge / IIoT · MLwhyHomepage says edge devices compute hundreds of KPIs in real time from motor current and voltage plus process data and send them to the cloud, so the core is high-frequency electrical signal acquisition at the edge with ML-derived health and performance insights. hiring 0 open tech roles elmodis.com sources elmodis.com elmodis.com | detailCB Insights lists USD 5.15m raised over 4 rounds with the last round in 2019, and Meidensha (Tokyo-listed) invested in 2021 as a strategic distribution partner for Japan and Asia. | Meidensha (invested) · Intel (invested)detailMeidensha · invested Elmodis's homepage says Meidensha Corporation of Japan became a major partner bringing investment and access to Asian markets; a database lists a venture round in March 2021. Meidensha size: 10,302 employees as of 2026-03-31 (meidensha.com/corporate). elmodis.com meidensha.com | Early | reads sensors | 14 LinkedIn 11-50 range, 14 members listed | elmodis.com linkedin.com |
| Groundup.ai | L0L3 | IoT sound and vibration sensors with an AI platform (GINA) for anomaly detection and root-cause analysis on industrial machines, sold as cognitive maintenance. | any manufacturingsourceCustomers span beverage bottling (Coca-Cola Bottling Saudi Arabia), materials manufacturing, the Singapore Navy and critical infrastructure. groundup.ai | ML · edge / IIoT · agentswhyThe product is anomaly detection on sound, vibration and thermal data from its own sensors, with GINA V2 described as agentic AI that diagnoses and prescribes. hiring 2 open tech roles: embedded / edge 1, other 1 groundup.ai | detailfunding USD 6m, as of 2025-04 e27.co Public sources give only funding (USD 1.8m seed in March 2023 and USD 4.25m Series A in April 2025, about USD 6.05m in total) and a claim of 30+ business customers in 5 countries. | - | Early | reads sensors | 30 LinkedIn 11-50 band (midpoint used) | groundup.ai thesaasnews.com linkedin.com |
| IPercept Technology | L0 | AI health monitoring for CNC machine tools from a retrofit motion sensor, detecting wear and misalignment before breakdowns. | machinery & metalworking onlysourceHealth monitoring for CNC machine tools in machining shops, with aerospace and automotive users. tech.eu | edge / IIoT · MLwhyThe core is a patented retrofit device on the moving part of a CNC machine that turns micrometre-scale motion into data (Motion Pattern Method) with no controller link, plus diagnostics of bearings, ballscrews, guides and spindles. hiring 1 open tech role: software 1 ipercept.io | detailFunding is EUR 5m seed (May 2025) plus USD 16.5m Series A (Sep 2026) in mixed currencies, with 20+ staff and 30+ enterprise customers; the Quant Gulf partnership for 200 machines is reported with up to EUR 10m revenue potential. | Konecranes (partner)detailKonecranes · partner Series A announcement names Konecranes among industrial service companies deploying IPercept's technology as the basis of their own service offering. arcticstartup.com | Early | reads sensors | ? | tech.eu vestbee.com |
| MOVUS | L0L3 | FitMachine magnetically mounted vibration/temperature sensors and the PlantOS platform that turn machine data into prioritised maintenance recommendations. | any manufacturingsourceCustomers span metals and mining (JSW Steel, Nyrstar), food and beverage, industrial manufacturing, water utilities and ports. movus.com.au | edge / IIoT · MLwhyThe product is FitMachine clip-on vibration and temperature sensors with a gateway feeding PlantOS, which detects faults weeks ahead and ranks maintenance recommendations by production impact. hiring 0 open tech roles movus.com.au | under USD 10kdetaildeal size AUD 10.2k · per year for 10 machines (2018 offer). October 2018 offer: AUD 85 per machine per month including 10 FitMachine sensors, a gateway, dashboard, installation and training; 10 machines for 12 months is AUD 10,200. prwire.com.au funding $ 4.8m, as of 2018-07 (article gives no currency code) startupdaily.net The only public price is a 2018 offer of AUD 85 per machine per month with sensors, gateway and installation bundled; the homepage still pitches it as less than a cup of coffee a day, and no funding after the 2018 4.8m Series A or any revenue figure was found. | - | Early | reads sensors | 30 LinkedIn 11-50 band (midpoint used); older press figure 16 FTE plus 13 interns in July 2018 | movus.com.au startupdaily.net linkedin.com |
| Nanoprecise Sci Corp | L0L3 | Wireless vibration/acoustic/temperature sensors plus AI software that predicts failures of rotating industrial equipment and schedules maintenance. | process industriessourceServes mining, oil and gas, metals, cement, chemicals, pharma, food and beverage and pulp and paper. betakit.com | ML · edge / IIoTwhyBetaKit describes AI-powered predictive maintenance built on internet-connected sensors that gauge equipment health, and says the planned 50 hires focus on AI among other functions; the own website and careers page were behind a bot wall and no ATS listing was found, so hiring could not be counted. sources betakit.com | -detailrevenue USD 10m, 2024 betakit.com funding USD 53m, as of 2025-03 betakit.com BetaKit (March 2025) reports 2024 ARR of USD 10m, up 114 percent year on year, a 2025 target of USD 20m, churn under 5 percent, net revenue retention over 115 percent, improving gross profit, USD 53m raised in total and a valuation over USD 100m, sold as subscription SaaS; the revenue figure is ARR, not audited revenue. sources betakit.com | Honeywell (invested) · NSK (invested) · Henkel (partner) · Sensata (invested)detailHoneywell · invested Honeywell Ventures participated in the USD 10m Series B led by Export Development Canada. ec-ma.com NSK · invested Bearing maker NSK Ltd. participated in the USD 10m Series B led by Export Development Canada. ec-ma.com Henkel · partner Henkel Adhesive Technologies announced a strategic partnership with Nanoprecise in November 2023 to digitalise industrial maintenance for process industries. henkel.com | Early | reads sensors | 127 press: https://betakit.com/nanoprecise-closes-series-c-to-predict-industrial-equipment-failures-with-ai (127 staff, March 2025, planning +50); LinkedIn 51-200 band | betakit.com linkedin.com |
| Neuron Soundware | L0L2 | nEdge sensor boxes plus nGuard software that detect machine faults and end-of-line product defects from sound and vibration with edge AI. | any manufacturingsourceHomepage says it can monitor any machine, from power generation equipment, pumps, compressors and mills to CNC machines, robots and 3D printers, with a dedicated energy industry page. neuronsw.com | edge / IIoT · MLwhyProduct is nEdge edge computers with 4 to 32 acoustic/vibration sensor inputs running AI on the device plus nGuard software with ML models for anomaly detection and PdM, per their 2024 sales deck; no open roles were listed to weigh. hiring 0 open tech roles startupjobs.cz | detailrevenue CZK 12.7m, 2018 oenergetice.cz Their 2024 sales deck prices a CNC use case at EUR 3,600 hardware per machine plus a licence fee of EUR 0.008 per product and claims 140+ installations; the only public revenue figure is about EUR 0.5m (CZK 12.7m) for 2018. | Airbus (invested) · Repsol (invested)detailAirbus · invested Tracxn lists Airbus among its investors; StartupYard says it was invited to the Airbus Innovation Lab in 2016. Round and amount not given. | Early | reads sensors | 15 LinkedIn 11-50 range, 15 members listed | neuronsw.com linkedin.com cc.cz |
| OnePredict (원프레딕트) | L0 | GuardiOne predictive-diagnostics products (Turbo, Substation, Motor, Cyclone) that read equipment sensor signals to predict the timing and cause of failures in rotating machinery, transformers and motors. | any manufacturingsourcePredictive diagnostics for power generation, petrochemicals, utilities, semiconductors and batteries. onepredict.ai | ML · data engwhyOf 6 open tech roles on their Greeting careers board, 3 are backend/full-stack/DevOps, 1 data scientist, 1 data engineer and 1 embedded developer; the products are the PDX platform for anomaly diagnosis and fault prediction plus the Cyclone data platform that prepares manufacturing data for AI. hiring 6 open tech roles: ML 1, data eng 1, software 3, embedded / edge 1 onepredict.career.greetinghr.com | detailfunding KRW 51.8bn, as of 2026-09 hellot.net Revenue is not public, but Korean press in September 2026 reports H1 2026 revenue at about 1.6 times full-year 2025 and new orders up 5.6 times year on year, with deployments described as moving from one-off demonstration projects to repeated rollouts ahead of a planned KOSDAQ listing. sources hellot.net byline.network | LG Energy Solution (invested) · S-OIL (invested)detailLG Energy Solution · invested New investor in the KRW 30bn Series C, March 2022 (round led by LB Investment). news.mtn.co.kr S-OIL · invested S-OIL (refiner majority-owned by Saudi Aramco) invested KRW 2bn in total in 2019 and 2020 and holds about 5 percent; it also uses OnePredict's predictive maintenance in its refinery and petrochemical plants. energykorea.co.kr | Early | reads sensors | 92 THE VC profile (thevc.kr/onepredict): 92 employees as of 2026-09, pension-registration based | onepredict.ai news.mtn.co.kr thevc.kr |
| Rotomate | L0 | Condition-monitoring AI that reads existing vibration sensors plus CMMS and ERP context and returns prioritised maintenance decisions instead of raw alerts. | process industriessourceSite names no industry list; named customers are Metsa Group, SSAB and Aurubis in pulp and paper, steel and copper. rotomate.ai | ML · agentswhyThe site describes analysing about 1,000 vibration signals in 10 to 15 minutes to catch developing bearing faults, then an AI colleague that answers natural-language questions with reasoning, ranks machines using maintenance history and process context, and sends work tickets to the CMMS; no open roles were listed to weigh. hiring 0 open tech roles thehub.io | detailfunding EUR 2.1m, as of 2026-06 tamradar.com Pre-seed company (EUR 2.1m, team of 10) selling software on top of customers' existing vibration sensors and CMMS/ERP, so no hardware in the deal; no pricing, contract value or revenue is public. | - | Early | reads sensors | 10 press: https://www.tamradar.com/funding-rounds/rotomate-pre-seed-2-1m (team grew from 2 to 10) | rotomate.ai tamradar.com |
| Samotics | L0 | SAM4: condition and energy monitoring of motor-driven equipment from current and voltage measured in the electrical cabinet. | process industriessourceIndustries page lists chemicals, metals and mining, oil and gas, pulp and paper, water, data centres and airports. samotics.com | ML · edge / IIoTwhyThe core is Electrical Signature Analysis: split-core current transformers in the motor control cabinet with signal conditioning feed fault-detection models (site claims 95.5% recall on confirmed faults), and ABB embeds the same ESA in its drives; the careers page lists no tech roles, only a revenue operations and an information security role. hiring 0 open tech roles careers.samotics.com | over USD 1mdetaildeal size GBP 7m · six-year framework contract (Southern Water). Southern Water signed a GBP 7m, six-year framework agreement in September 2025 covering 1,458 critical assets at about 600 sites (about GBP 1.2m per year) samotics.com Large multi-year utility frameworks (GBP 7m over six years with Southern Water, extensions with Welsh, Anglian and Yorkshire Water) across 80+ customers and 7,000+ monitored assets; delivery is cabinet hardware plus software plus analyst review of ambiguous findings. | ABB (partner, invested)detailABB · partner Long-term strategic partnership since 2022; ABB Motion Services launched drives with Samotics ESA technology embedded at ADIPEC, November 2025. samotics.com | Early | reads sensors | 100 company about page says approximately 100 specialists; LinkedIn 51-200 range | samotics.com samotics.com linkedin.com |
| Viking Analytics | L0 | MultiViz: unsupervised-ML vibration analysis and condition monitoring that triages machines for maintenance analysts. | any manufacturingsourceSite says MultiViz monitors machine types across industries; the record names process and discrete plants. vikinganalytics.se | MLwhyMultiViz builds unsupervised, explainable per-sensor models of normal machine behaviour ('modes') and runs detector algorithms for vibration symptoms such as impacting and harmonics without failure history, per the product site; the careers page says there are no open positions. hiring 0 open tech roles vikinganalytics.se | detailSoftware sold directly (Sodra group-wide framework agreement, June 2026) and through industrial partners that embed it, such as Weir's NEXT platform, ABB and sensor vendor CTC; no contract values, pricing or revenue are published, and the last disclosed round was a EUR 3m Series A in January 2023. | ABB (invested, partner) · ArcelorMittal (invested) · Weir Group (partner)detailABB · invested ABB Technology Ventures took a minority stake and joined the Series A (January 2023, round co-led by Finindus and Industrifonden). new.abb.com ABB · partner Commercial agreement with ABB to monitor ABB customers' switchgear, announced with the Series A; earlier frame agreement with ABB in February 2021 and winner of the ABB Electrification startup challenge 2020. vikinganalytics.se ArcelorMittal · invested Series A co-led by Finindus, described in the announcement as backed by ArcelorMittal. reliabilityweb.com Weir Group · partner Weir partnered with Viking Analytics to add its analytics to Weir's digital offering for mineral processing plants (news item dated 2025-09-29). vikinganalytics.se | Early | reads sensors | 25 LinkedIn 11-50 range | vikinganalytics.se vikinganalytics.se linkedin.com |
| AISing (株式会社エイシング) | L1L0 | Sells the AiiR edge-AI algorithm library: AiirMST learns and corrects machine control in real time on microcontrollers, AiirMSAT does on-device anomaly detection for predictive maintenance. | any manufacturingsourceCompany page says it serves manufacturing (product and FA control) through to energy and infrastructure, with use cases split into manufacturing DX and energy and infrastructure DX. aising.jp | control · MLwhyCore is the AiiR algorithm family that learns and predicts in real time on edge devices and microcontrollers (AiirMST control correction, AiirMSAT on-device anomaly detection, AiirDNN, AiirDTW); both open tech roles are ML (AI researcher, data scientist). hiring 2 open tech roles: ML 2 aising.jp | detailBusiness is AI solution development and implementation consulting alongside licensing of its own edge algorithms; the recruit page's data scientist and consultant roles run PoC and development projects for large manufacturers; Speeda/INITIAL shows a JPY 700m Series B in 2020, revenue paywalled. | Omron (partner) | Customer track record | writes to the machine | - | aising.jp initial.inc mynavi-agent.jp |
| Aixtal (アイクリスタル株式会社) | L1LX | Nagoya University process-informatics startup whose MetaFactory digital twin and cloud service analyse line data with physics simulation plus ML to propose optimal manufacturing conditions across whole process chains. | any manufacturingsourceClients span semiconductors (image sensors, wafers), automotive ceramics, chemicals, moulding and casting. aixtal.com | simulation · physics + ML · MLwhyProcess informatics that combines physical simulation and CAE with ML to set manufacturing conditions (MetaFactory digital twin of whole lines); 13 open roles include 2 data scientist/ML, 1 CAE engineer, 1 manufacturing process engineer and 2 software. hiring 6 open tech roles: ML 2, simulation 1, software 2, other 1 aixtal.com (13 roles listed; the other 7 are product manager, 2 sales, 4 back office) | detailfunding JPY 720m, as of 2025-09 thebridge.jp Revenue comes from consulting, PI education and the MetaFactory product together (63 client companies and 140 projects by 2026-04 per own site, about 2.2 projects per client), which reads as project-led; no prices or revenue published. sources aixtal.com thebridge.jp | Toyota (partner)detailToyota · partner Joint announcement (2025-10-08): Toyota Motor and Aixtal developed an algorithm that automatically finds optimal sensor positions on a 3D digital twin of a Toyota plant for its Telemotion autonomous vehicle-transport system. aixtal.com | Customer track record | reads MES / ERP | ? | aixtal.com aixtal.com thebridge.jp |
| Fero Labs | L1L3 | Explainable ML software (Diagnostics, Simulator, Production) that finds root causes and recommends optimised setpoints for continuous processes in steel, chemicals and cement. | process industriessourceTargets continuous processes in steel, chemicals and cement, plus oil and gas and CPG (Gerdau, Covestro, Henkel). ferolabs.com | ML · causalwhyOwn blog says models are tree ensembles and kernel methods on tabular plant data with SHAP-style explanations and constrained setpoint optimisation; homepage sells root-cause diagnostics and cause-effect links as the core; no careers page found to weigh hiring. | detailfunding USD 24m, as of 2023-06 alleywatch.com Sold as SaaS after a 14-day free trial into a 'Ramp Year License' per homepage; Fastmarkets reports about 40% of revenue from steel and about 60% of business in the US, and Gerdau savings of about USD 3 to 6 per ton; no revenue or price figure. | - | Customer track record | reads sensors | 29 LinkedIn 11-50 band, about 29 employees listed | ferolabs.com alleywatch.com linkedin.com |
| Gigaton | L1L0 | Self-learning control software for cement kilns that forecasts the process, simulates setpoint moves on an ML twin and writes fuel, oxygen and kiln-speed setpoints to the plant's DCS or APC. | cement & building materials onlysourceProduct page and cases are all cement kilns (Heidelberg Materials, Holcim, Mannok, JK Cement, Adani); steel, glass and chemicals named only as next. gigaton.co | control · ML · simulationwhyProduct is a replacement control system that forecasts kiln behaviour with self-learning ML and simulates each action on a digital twin before writing setpoints to the DCS; the Solutions Engineer posting asks for APC, MPC and PID control strategy work. hiring 3 open tech roles: controls 1, software 1, solutions / deployment 1 jobs.ashbyhq.com. The old Carbon Re Workable board shows 0 openings.) | detailfunding more than USD 35m, as of 2026-06 enterprise.cam.ac.uk Customer value claimed at GBP 1.x m average extra profit per kiln per year (product page; the decimal was garbled in the fetched text) and USD 1 to 3m annual savings per deployment (Series A release); the Director of Deployment posting says 9 plants are live, a plant takes about six months to reach evaluation and the goal is 50 plants by end of 2027 with half the people per plant, so deployments are currently services-heavy. | - | Customer track record | writes to the machine | 32 LinkedIn 11-50 range (uk.linkedin.com/company/gigaton-technologies) | gigaton.co gigaton.co gigaton.co |
| INTERX (인터엑스) | L1L2L5 | Manufacturing AI that uses data collected from production equipment to compute appropriate operating conditions (process-condition optimisation) and to auto-judge product defects, sold as projects and a platform to automotive, semiconductor and battery makers. | any manufacturingsourceCustomers in automotive, semiconductor, display, battery, bio and food, including Hyundai Motor and Samsung SDI. hellot.net | data eng · ML · CVwhyBloter describes the product as using data collected from production equipment to compute operating conditions and auto-judge defects, and the company calls itself a manufacturing data-space platform; of 11 open technical roles on its GreetingHR page, 7 are backend, full-stack, .NET or platform software roles, with one AI agent engineer, one production simulation and optimisation developer and one AI data application engineer. hiring 11 open tech roles: LLM / AI 1, software 7, solutions / deployment 1, other 2 interxlab.career.greetinghr.com | detailrevenue KRW 8.6bn, 2025 bloter.net funding KRW 27bn, as of 2025-09-05 hellot.net Per the KOSDAQ filing coverage, revenue fell 41 percent from KRW 14.7bn (2024) to KRW 8.6bn (2025) with a KRW 8.3bn operating loss in 2025, the business today is custom AI projects of one to three years that it wants to shift to standard solutions and platform subscriptions, foreign revenue was 0.05 percent, and about 59 percent of the 2027 revenue forecast (KRW 30.5bn) is not yet contracted. | Hyundai Motor (founder from)detailHyundai Motor · founder from Founder and CEO Park Jung-yoon joined Hyundai Motor after graduating and worked in production engineering (생산기술) on production equipment, per his Forbes Korea interview. forbeskorea.co.kr | Customer track record | reads sensors | 175 THE VC profile (thevc.kr/interx): 175 employees as of 2026-09; Wanted company page lists 169 | bloter.net wanted.co.kr thevc.kr |
| Mogulinker / Mushroom IoT (蘑菇物联) | L1L0L3 | AIoT SaaS that connects factory utility stations (air compressors, chillers, pumps) and its 云智控 AI Cloud Control runs them unmanned with 10-30 percent energy savings. | any manufacturingsourceUtility-station AI for factories in general; the record names manufacturing, pharma, electronics and data centres with 6,000+ enterprises. mogulinker.com | control · edge / IIoT · MLwhyThe AI Cloud Control product runs utility stations (compressors, chillers) unmanned through cloud, edge-server and IoT-gateway layers, and of 4 distinct technical roles on its recruiting page one is a senior control-algorithm engineer (MPC, LQR, control simulation), one a vibration-diagnosis ML engineer, one an OR algorithm engineer (CPLEX, Gurobi, heuristics) and one an HVAC algorithm operations engineer. hiring 4 open tech roles: ML 1, OR / optimisation 1, controls 1, solutions / deployment 1 mogulinker.com | USD 100k to 1mdetaildeal size at least CNY 1m · per project, lower bound ("RMB million-level projects"). Own job ad for an international sales lead says the role independently leads projects at the RMB million level (百万级项目), i.e. roughly RMB 1m and up per project mogulinker.com PEdaily (2026-09) reports a C1/C2 round of about RMB 200m and two pricing models, a product model of installation fee plus annual O&M fee and a service model of energy-saving revenue share or AI hosting, delivered with IoT gateways, edge servers and inspection robots; no revenue figure or cumulative funding total was found. | Phoenix Contact (invested) · Galanz (founder from)detailPhoenix Contact · invested Phoenix Contact (菲尼克斯) took part in the Series C (Sep 2022, over RMB 100m) and made a separate strategic investment in Dec 2022; also listed on Mogulinker's own investor page. Phoenix Contact size: 22,000 employees (blog.phoenixcontact.com/about-us). pitchhub.36kr.com blog.phoenixcontact.com Galanz · founder from Founder Shen Guohui (沈国辉) was previously China market general manager at Galanz (格兰仕), per STAR Market Daily. chinastarmarket.cn | Customer track record | writes to the machine | ? | mogulinker.com pitchhub.36kr.com |
| Oden Technologies | L1L3LX | Process AI that ingests machine and sensor data from continuous and discrete lines (extrusion, cable, ink, plastics) and gives front-line operators real-time parameter recommendations and factory analytics. | process industriessourceNamed customers run extrusion, cable, copper, ink, plastics and packaging lines (Teknor Apex, Westlake, Greif, Lake Cable, INX). oden.io | ML · data eng · agentswhyFortune describes the product as historical production data combined with supervised machine learning to give operators process-setting recommendations, the site sells a Data Engine plus Process AI and Forge agents for agentic root-cause analysis, and the Ashby board lists 3 technical roles (one forward-deployed data scientist, one solutions engineer, one full-stack engineer). hiring 3 open tech roles: ML 1, software 1, solutions / deployment 1 api.ashbyhq.com | detailNo revenue or contract value is public; Oden's own Litmus partner page says pricing is based on the number of manufacturing processes and sites deployed, Fortune (2024) describes a software layer on existing factory assets with a USD 28.5m Series B and names startups Braincube and TwinThread as competitors, and only 4 roles are open in total, one of them sales. | - | Customer track record | reads sensors | 45 LinkedIn 11-50 band, 'View all 45 employees' (slug oden-technologies-ltd-) | oden.io oden.io linkedin.com |
| plus10 | L1L3L2 | Shopfloor AI (Shannon, Darwin, Hopper) that reads high-frequency machine data and recommends parameter settings and fixes for availability, bottlenecks and injection-moulding scrap. | discrete, generalsourceIndustry pages list electronics manufacturing and assembly, machinery and plant engineering, automotive, pharma and medtech, and consumer goods. plus10.de | ML · edge / IIoTwhyHopper is described as self-learning software that analyses quality-relevant process data in real time and recommends parameter changes inside validated process windows, and it rests on the DataCollector, which samples PLC signals in the millisecond range over S7, Beckhoff, Codesys, OPC UA and Modbus and ships on plus10-supplied industrial edge PCs; the careers page lists only one role, a full-stack development intern. hiring 1 open tech role: software 1 plus10.de | detailNo price, revenue or funding figure is public on its site; the money signals are channel deals, with Hopper integrated into ARBURG's GESTICA machine control (October 2026) and plus10 joining the Siemens Xcelerator marketplace (August 2026), and the DataCollector ships on plus10-supplied edge hardware. | - | Customer track record | reads sensors | 25 startbase 11-50 band (https://www.startbase.com/organization/plus10/) | plus10.de startbase.com |
| Profet AI (杰倫智能科技) | L1L3L0 | No-code AutoML and the Domain Twin platform that let factory process, quality and supply-chain engineers build predictive models from structured production data for process optimisation, parameter recommendation, anomaly detection and predictive maintenance. | any manufacturing | ML · agentswhyThe core product is a no-code AutoML platform for structured manufacturing data with 60-plus built-in algorithms (own FAQ, TechNice), now wrapped in the Domain Twin platform whose AI Studio builds governed AI agents and knowledge workflows (own site); job listings on 104 were not reachable, so no hiring evidence. | detailfunding TWD 200m, as of 2023-02 technice.com.tw Own FAQ says the platform is billed mainly by server licence and concurrent users on a rental (not buy-out) basis, with models deployable to edge; TechNice (2023) reports total funding above NT$200m, business volume doubling in 2022 and 100-plus mid-to-large manufacturing customers, but no revenue figure was found. | AUO (invested)detailAUO · invested AUO (友達) was an existing investor and invested again in the NT$170m Series A (Jan 2023) led by Darwin Ventures. meet.bnext.com.tw | Customer track record | reads MES / ERP | 50 press: https://www.technice.com.tw/uncategorized/36975/ (Feb 2023: 'from the current 50 to 100 people'); no newer figure found | profetai.com profetai.com technice.com.tw |
| All Sense Technology / Quanying (全应科技) | L1L0 | 热电云 cloud-edge platform and ADMC intelligent control that optimises and automatically regulates industrial cogeneration and thermal power plants that supply steam and heat to industrial parks. | energy & utilities onlysourceServes industrial cogeneration and thermal power plants that supply steam and heat to industrial parks. 36kr.com | control · ML · edge / IIoTwhyThe company site describes its ADMC thermal-power control system as model predictive control, real-time optimisation and real-time scheduling, and its iDiC edge controller as running AI-MPC and ML models, while 36Kr describes a cloud-edge-device architecture sold with intelligent edge controller hardware; no careers page or job listings were reachable, so hiring is null. | detailcompetes with GE, Siemens, ABB 36Kr (2019) says revenue comes from three parts, cloud platform account fees, edge controller hardware and an annual data computation service fee, and quotes a 2019 projection of over CNY 30m in contract revenue from 15+ large thermal power projects with a 300-project pipeline; later rounds are only described as tens of millions to several hundred million RMB, so no total, actual revenue or contract value is public. | Delta Electronics (founder from)detail | Early | writes to the machine | ? | 36kr.com pitchhub.36kr.com |
| AMESA (formerly Composabl) | L1 | Platform to build, train in simulation, validate and deploy autonomous agents that control industrial processes such as blending, reactors and CNC lines. | any manufacturingsourceSite case studies cover aluminium can and glass container production, snack CPG, CNC machining, thermal reactors, and pharma and chemicals. amesa.com | control · simulationwhyAMESA docs describe reward-driven (reinforcement) learning of agents trained against data-driven simulators built from historical CSV data or customer simulators in Docker, composed with Python MPC and PID controllers, with LLM calls only as an optional perceptor; no careers page or ATS board was found (amesa.com/careers returns 404), so hiring is null. | -detailfunding USD 4.2m, as of 2024-05 markets.financialcontent.com The only public money figure is the USD 4.25m seed led by Momenta (May 2024, as Composabl, with two Fortune 500 companies building agents at the time); the company advertises customer value (USD 21m a year at one refinery, USD 100m+ combined realised and projected) but discloses no pricing, contract values or revenue. | - | Early | writes to the machine | 6 LinkedIn 2-10 range (midpoint used) | amesa.com linkedin.com |
| Basetwo AI | L1L0L3 | AI agent platform that builds hybrid physics-plus-data digital twins of pharma and chemical processes to simulate, optimise setpoints, soft-sense quality and find root causes. | pharma & life sciences, chemicalssourceBuilds digital twins for pharma (Johnson & Johnson, Merck), personal care (L'Oreal) and specialty chemical processes. basetwo.ai | physics + ML · simulation · agentswhyThe homepage describes digital twins that combine mass balance, kinetics and transport with machine learning to simulate conditions outside historical data, with an AI agent doing data preparation and analysis. hiring 1 open tech role: other 1 builtin.com sources basetwo.ai builtin.com | detailbuyer engineering Basetwo sells a SaaS pilot licence (3 to 4 months) and a 12-month platform licence through AWS Marketplace with only a placeholder public price, so contract values are negotiated privately; known rounds are a USD 3.8m seed (2022) and USD 11.5m Series A (2024) per the existing record, with no revenue figure found. sources aws.amazon.com basetwo.ai | Shimadzu (invested) · Nippon Steel (partner)detailShimadzu · invested Shimadzu invested through its CVC Shimadzu Future Innovation Fund, part of the USD 11.5m Series A announced Jan 2025. shimadzu.com Nippon Steel · partner Strategic distribution agreement (Nov 2025): NS Solutions (Nippon Steel Solutions) is Basetwo's authorised distributor in Japan, with joint seminars and PoCs planned. NS Solutions is a Nippon Steel subsidiary per its GlobalData company profile. finance.yahoo.com globaldata.com | Early | reads sensors | 30 LinkedIn 11-50 band (midpoint used) | basetwo.ai basetwo.ai thelogic.co |
| Deep.Meta | L1L2 | Explainable AI and a physics digital twin for steel mills that predict slab temperatures and metallurgical defects from existing sensor data and suggest parameter adjustments. Its current product SLICE matches steel slabs to orders by dimension and grade (deepmeta.io). | metals & steel onlysourceSite describes AI optimisation and slab selection software for steel planners and producers. deepmeta.io | physics + ML · ML · ORwhyImperial describes Deep.Optimiser-PhyX as physics-based AI predicting slab temperatures in reheating furnaces, Made Smarter describes a model trained on 40 years of production data, and the current SLICE product matches slabs to orders by dimension and grade (a matching optimisation). hiring 3 open tech roles: ML 1, software 1, solutions / deployment 1 deepmeta.io | detailfunding GBP 500k, as of 2025-06 imperial.ac.uk Deep.Meta is very early: Imperial reports GBP 500k from angels and VCs plus a GBP 1.9m Innovate UK grant and a GBP 100k Manchester Prize award, Made Smarter says it had secured its first UK customer, and the product is cloud SaaS needing no hardware. | Tata Steel (founder from) | Early | reads sensors | 9 LinkedIn 2-10 range, 9 members listed | linkedin.com imperial.ac.uk deepmeta.io |
| Imubit | L1L3 | Closed-loop reinforcement-learning process control that writes moves through the plant's existing APC or DCS within engineer-set limits. | process industriessourceSite names refining and gas processing, chemicals and polymers, cement, and mining and minerals. imubit.com | control · ML · simulationwhyProduct is deep-learning process controllers trained by reinforcement learning across millions of simulations that write moves to the plant APC or DCS minute by minute (careers and home page). hiring 0 open tech roles boards-api.greenhouse.io | detailfunding USD 50m, as of 2021-08 houston.innovationmap.com buyer operations exec Imubit's public reseller agreement prices an annual recurring subscription per deployed Unit, defined as one plant or site regardless of how many process units run on it, with a 5% automatic annual increase. | - | Early | writes to the machine | 125 LinkedIn 51-200 range (midpoint used) | imubit.com linkedin.com |
| Intelecy | L1L0L3 | No-code industrial ML platform where process engineers build anomaly, forecast and recommendation models on live plant data, optionally streaming predictions back to control systems. | process industriessourceIndustries menu lists oil and gas, food and beverage, power and renewables, mining and metals, chemicals, and water. intelecy.com | ML · streaming · data engwhyProduct is a no-code ML platform for anomaly detection and forecasting on live plant data, with predictions streamed back to control systems per the homepage. hiring 0 open tech roles intelecy.bamboohr.com | -detailrevenue NOK 9.1m, 2025 21st.ai Intelecy AS operating income grew from NOK 3.0m (2023) to 6.7m (2024) to 9.1m (2025) while the annual loss widened to NOK 41.8m in 2025, so it is small and far from profitable; the only round found was USD 3.5m led by Yara in 2023. | Yara (invested)detailYara · invested Yara Growth Ventures (Yara International's venture arm) led the USD 3.5m round (2023), with GC Rieber, Aidiom and Investinor. iotm2mcouncil.org | Early | writes to the machine | - | intelecy.com linkedin.com |
| Intellegens | L1 | Alchemite, machine learning for sparse experimental data that proposes formulations and process parameters and cuts the experiments needed in R&D and process development. | any manufacturingsourceHomepage says it is for chemicals, materials, FMCG, life sciences and manufacturing. intellegens.com | MLwhyAlchemite is machine learning for sparse experimental data with uncertainty quantification, used for design of experiments and formulation optimisation per the homepage. hiring 0 open tech roles intellegens.com sources intellegens.com intellegens.com | -detailfunding USD 2.2m, as of 2023-08-30 cbinsights.com A small business: about 20 employees, GBP 684k cash and GBP 450k net assets for 2025, turnover not reported under small-company exemption; CB Insights shows USD 2.17m raised, of which USD 2.03m is a 2023 Innovate UK grant, and the site sells consultancy and training alongside the software. | - | Early | documents only | 19 LinkedIn 11-50 range, 19 members listed | intellegens.com linkedin.com |
| Kelvin | L1LX | Software that turns models and agents into governed actions on existing control systems, up to changing machine setpoints in closed loop within guardrails. | process industriessourceHomepage names energy production, continuous and process manufacturing, industrial services, power, and mining and materials. kelvin.ai | control · edge / IIoT · agentswhyThe homepage describes two-way Connectors to PLCs, RTUs, DCS and SCADA running at the edge or on premises, deterministic Engines for optimisation and control, and agents that reason over operational context, with a materials case and energy cases of closed-loop setpoint changes. | detailKelvin sells multi-year enterprise software agreements (Santos, 2023, value undisclosed) and lists a SaaS offer on AWS Marketplace priced by private offer. | - | Early | writes to the machine | 125 LinkedIn 51-200 range (midpoint used) | kelvin.ai linkedin.com |
| Limitless Labs (formerly LimitlessCNC) | L1 | An AI CAM agent inside Mastercam, NX and Creo that generates CNC toolpaths, picks tools and sequences operations, which the programmer reviews and posts to the machine. Calcalist describes the model as trained on metal-cutting physics and part geometry; it is distributed through Sandvik's Cimatron CAD/CAM. | aerospace & defence, automotivesourceHomepage says it is trusted by leaders in aerospace, defense and automotive manufacturing, built for CNC programming in those regulated environments. limitless-labs.ai | ML · physics + ML · ORwhyThe only open technical role (Senior Algorithm Developer) asks for models trained on 3D geometry (OpenCASCADE, Pytorch3D), physics-aware machining models, and graph, scheduling and constraint-satisfaction methods, matching Calcalist's description of a model trained on metal-cutting physics and part geometry; no evidence was found that an LLM sits at the core despite the 'agent' branding. hiring 1 open tech role: ML 1 limitless-labs.ai | Sandvik (partner) · Dell (invested)detailSandvik · partner Partnership with Sandvik's Cimatron CAD/CAM business (about 30,000 active users), reported with the June 2026 Series A. calcalistech.com Dell · invested Dell Technologies Capital co-led the USD 20m Series A (June 2026) with Square Peg. calcalistech.com | Early | writes to the machine | 46 LinkedIn 11-50 range, 46 members listed | limitless-labs.ai linkedin.com calcalistech.com | |
| Smart Steel Technologies | L1L4L2 | AI process-control and scheduling software for steel mills: temperature control, surface quality, casting, material allocation and schedulers. | metals & steel onlysourceAI process control and scheduling for steel mills and non-ferrous metals. smart-steel-technologies.com | OR · ML · physics + MLwhyFour of the six listed products are schedulers, planners and allocators, and the Short-Term Scheduler page says it is 'fuelled by an MILP solver'; Temperature Control combines 'built-in process expertise' with ML to give real-time recommendations. hiring 0 open tech roles smart-steel-technologies-gmbh.jobs.personio.de | detailNo revenue or price was found: Heraeus became the largest shareholder in 2023 at an undisclosed price, and Northdata holds 2019 to 2023 statements for the Berlin GmbH (HRB 203867 B) only behind a paywall; the product is sold as AI software packages that need no extra sensors or hardware, a software-margin profile. | Heraeus (acquired by)detailHeraeus · acquired by Heraeus (via Heraeus Electro-Nite) took a majority position in Feb 2023 and became the largest shareholder; Heraeus's own release says "largest shareholder", AIST says "majority position". Heraeus size: 15,181 employees at end 2024 (Heraeus financial report 2024). aist.org heraeus.com | Early | reads sensors | - | smart-steel-technologies.com heraeus.com linkedin.com |
| TVARIT | L1L2 | Industrial AI (TiA) that predicts quality and energy outcomes in foundries and metal plants and recommends process parameter changes to operators. | process industriessourceHomepage leads with metal industries (aluminium, steel) and also lists process industries: dairy, cocoa and chemical processing. tvarit.com | ML · physics + MLwhyHomepage describes a 'Hybrid AI approach' combining domain-specific knowledge equations with AI models plus transfer learning, and its Data Scientist posting asks for supervised/unsupervised learning and time-series feature engineering on hundreds of machine sensor signals; the careers page showed no listed roles when fetched, so hiring is left null. | detailfunding EUR 9.3m, as of 2023-10-31 (sum of EUR 2.8m (May 2022, Matterwave) and EUR 6.5m Series A (Oct 2023, Momenta, Futury); 2019 seed from Futury and angels undisclosed, so a lower bound) startbase.de buyer operations exec Funded with at least EUR 9.3m across a 2022 round and a 2023 Series A, describes itself as SaaS, and names Maxion Wheels, Kamax, KSM and Schunk as customers, but no revenue or contract values are public (Northdata shows revenue only behind a paywall). | - | Early | reads sensors | 125 LinkedIn 51-200 range (midpoint used) | tvarit.com linkedin.com |
| Instrumental | L2L3 | Sells a manufacturing data and AI platform for electronics assembly that combines line camera images with test and assembly data to intercept defects and find yield and throughput issues. | electronics assembly onlysourcePlatform is for electronics assembly lines across consumer, enterprise, AI compute and defense electronics (Meta, NVIDIA, L3Harris). instrumental.com | CV · data eng · MLwhyProduct combines line camera images with functional test and assembly data into a per-unit digital thread and applies visual AI and machine learning to correlate test failures with visual anomalies (TechCrunch 2022, own 2026 investment post). | USD 10k to 100kdetaildeal size USD 30k · per 3-month proof of concept. AWS Marketplace listing prices a 3-month proof of concept at USD 10,000 per month (USD 30,000 total), billed by GB of data and image streams ingested; production contract prices are custom and not public aws.amazon.com funding USD 80m, as of 2022-02-16 ('over USD 80m' after the USD 50m Series C; a March 2026 round from NVentures and Root Ventures has no disclosed amount) techcrunch.com Series C coverage reported 150% net revenue retention in 2021 across 30 customers on 50 programs, and the AWS Marketplace proof of concept is priced at USD 30,000 for 3 months; no production contract value or revenue is public. | - | Customer track record | reads sensors | 75 directory: cyberdb.co estimate 51-100 employees, midpoint used (https://www.cyberdb.co/ai-vendor/instrumental/) | instrumental.com instrumental.com techcrunch.com |
| Neurocle (뉴로클) | L2 | No-code auto-deep-learning vision inspection software (Neuro-T to train models from images, Neuro-R runtime, Neuro-T Engine for MLOps integration) used by manufacturers to classify and locate product defects. | any manufacturingsourceVision inspection used in semiconductor, battery, automotive, steel, food and beverage, pharma and medical devices. neuro-cle.com | CVwhyhiring 1 open tech role: CV 1 neurocle-career.com (two further open roles are sales and business development; all other listed roles marked closed) sources neurocle-career.com neuro-cle.com | detailNo revenue, funding amount or pricing found: the Unicorn Factory startup database shows investment and revenue fields only behind login, the own site discloses no licensing terms, and headcount is 37 (Aug 2026) against a claim of 400+ customers. sources unicornfactory.co.kr neuro-cle.com | LG Display (founder from)detailLG Display · founder from CEO and founder Lee Hong-seok is ex-LG Display (Edaily). edaily.co.kr | Customer track record | reads sensors | 41 THE VC profile (thevc.kr/neurocle): 41 employees as of 2026-09 | neuro-cle.com venturesquare.net thevc.kr |
| Polysense | L2L1 | Inline camera quality inspection for food lines whose AutoControl module adjusts machine parameters when defects drift. | food & beverage only | CV · controlwhyInline cameras inspect every product (Qualify) and AutoControl adjusts machine parameters when deviations appear, integrating with PLCs and SCADA per the homepage. hiring 2 open tech roles: CV 1, software 1 careers.polysense.ai sources careers.polysense.ai polysense.ai | -detailfunding USD 12.9m, as of 2026-07-08 (USD 10.7m seed led by Felix Capital plus USD 2.2m earlier) tech.eu Raised USD 12.9m in total, with the CEO describing a move from pilots to live deployments at large food manufacturers (Agristo, Darta, Poppies Bakeries) and a US sales hire open in Chicago, but no revenue, pricing or contract values are public. | - | Customer track record | writes to the machine | ? | polysense.ai freshplaza.com tech.eu |
| Ripik.ai | L2L3L1 | Computer vision AI agents that watch kilns, furnaces, conveyors, molten metal and raw material feeds in heavy industry to flag anomalies and support process optimisation. | process industriessourceMonitors heavy-industry processes at steel, aluminium and zinc producers, cement plants and power plants (Tata Steel, Hindalco, Hindustan Zinc). ripik.ai | CV · ML | -detailrevenue INR 235m, FY2024-25 inc42.com funding USD 500k, as of 2022-11 inc42.com Inc42 Datalabs (from statutory filings) shows FY25 revenue INR 23.5 crore (about USD 2.8m), up about 91% on FY24, with a loss after tax of INR 18.2 crore; disclosed equity is only about USD 0.5m, with an undisclosed pre-Series A from Accel and Venture Highway per indianstartuptimes. sources inc42.com indianstartuptimes.com | - | Customer track record | reads sensors | 92 Inc42 directory: 92 employees (https://inc42.com/company/ripikai/); LinkedIn 51-200 band | ripik.ai inc42.com inc42.com |
| Visionairy | L2 | End-to-end AI visual inspection platform (industrial PC, edge software, central analytics) using patented unsupervised models that learn from about 50 good-part images. | any manufacturingsourceIndustries menu lists electronics, automotive, food and beverage, recycling and waste treatment, and glass and optics, framed as one platform for any industry. visionairy.io | CV · edge / IIoTwhyHomepage describes patented unsupervised anomaly-detection vision models (CNRS / ENS Paris-Saclay) running on an industrial edge PC at 50 ms per part; no job listings could be read. sources visionairy.io | detailfunding EUR 1.2m, as of 2022 universite-paris-saclay.fr Homepage sells one bundled system ('Industrial PC, on-edge software, and centralized analytics. One system, one provider'), so hardware is part of each deal; no pricing published. sources visionairy.io | - | Customer track record | reads sensors | 15 press: https://www.universite-paris-saclay.fr/en/news/visionairy-artificial-intelligence-improve-quality (team of 15, July 2024); LinkedIn 11-50 range with 9 members listed | visionairy.io linkedin.com universite-paris-saclay.fr |
| 24 Vision System | L2 | Neural-network visual inspection that checks every product for defects and configuration errors on just-in-sequence automotive lines. | any manufacturingsourceHomepage headline targets the automotive industry, and its implementation list adds consumer products, medical devices, food and beverages, logistics and packaging, and agriculture. 24vs.io | CV | detailfunding CZK 37.5m, as of 2022 magazin.positiv.cz Site sells camera-plus-AI inspection with paid feasibility studies and 6-week prototyping, so each deal carries hardware and project work. sources 24vs.io | - | Early | reads sensors | ? | 24vs.io magazin.positiv.cz |
| Adacotech (株式会社アダコテック) | L2 | Non-deep-learning visual inspection AI (AdaInspector learns from good parts only, POKAMIRU assists manual inspection) that runs on CPU with few samples, from AIST image-analysis research. | any manufacturingsourceHomepage cases include an electronics assembler, a steel sheet maker and nonwovens, resin moulding and sintered parts inspection; the record names automotive parts, electronics and materials. adacotech.co.jp | CVwhyProducts (AdaInspector Cloud, POKAMIRU, Shiwaketter) are visual inspection and defect-image classification AI; the recruit page moved to a Notion site that did not render, so no hiring counts. sources adacotech.co.jp initial.inc | - | - | Early | reads sensors | 20 paiza.jp recruiter page: 20名 (2025/03現在), opened 2026-11-03 | adacotech.co.jp adacotech.co.jp paiza.jp |
| Deltaray | L2 | Inline 3D X-ray (3DAX) inspection of every part at production speed, with AI analysis of internal and external defects. | discrete, generalsourceimec press release names the automotive industry, the medical industry and the manufacturing industry as its markets for inline X-ray part inspection. imec-int.com | CVwhyimec press release: 3D accelerated X-ray imaging (3DAX), 100 times faster than CT, with scans analysed in 3D by AI. sources imec-int.com | detailfunding EUR 5.7m, as of 2022-01 imec-int.com Sold as fully automated X-ray inspection machines and 3DAX imaging kits for integration, so revenue is hardware-led. sources imec-int.com | - | Early | reads sensors | 25 LinkedIn 11-50 range | imec-int.com linkedin.com |
| Elementary | L2L3 | Sells AI-powered camera inspection stations and a QA platform that learn a product's quality standard and automatically flag defects on production lines. | any manufacturingsourceIndustries menu lists food and beverage, automotive, paper and pulp, CPG, industrial and construction, and electronics. elementaryml.com | CV · edge / IIoTwhyAbout page: machine vision system (edge hardware) plus cloud QualityOS platform; only openings found were aggregator listings (a 2025 senior Python software engineer), and the BambooHR board did not render. sources elementaryml.com builtin.com | USD 10k to 100kdetaildeal size USD 42k · per inspection station per year (USD 3,500 a month, third-party listing). Third-party directory robotomated.com lists the Elementary Label Inspector 'From $3,500/mo' (about USD 42k a year) and says pricing is on request from Elementary; Elementary's own blog confirms subscription ('Quality as a Service') but gives no figure. robotomated.com Sold as a subscription that bundles inspection hardware, updates and maintenance; a third-party listing puts a station at about USD 3,500 a month; Rockwell Automation made a strategic investment in 2024 (amount undisclosed). | Rockwell Automation (invested, partner) · Toyota (invested)detailRockwell Automation · invested Strategic investment announced May 2024; Toyota Ventures, an existing investor, also took part in the round. prnewswire.com Rockwell Automation · partner Rockwell's FactoryTalk Analytics VisionAI inspection product is built on partnership work with Elementary, which Rockwell's release calls 'a Rockwell investment'. rockwellautomation.com Toyota · invested Toyota AI Ventures invested through its first 'Call for Innovation'; it was an existing investor in the Series A and again in the 2024 round. pressroom.toyota.com | Early | reads sensors | 30 LinkedIn 11-50 band at linkedin.com/company/elementary-robotics (legacy slug, page shows current Elementary branding), midpoint used | elementaryml.com elementaryml.com linkedin.com |
| Frinks AI | L2 | Sells a no-code machine vision AI platform that automates inspection, compliance checks and shipment validation on production lines. | any manufacturingsourceServes automotive, consumer goods, cement, steel, building materials, chemicals and medical devices. india.entrepreneur.com | CVwhyPress release: 'foundational vision models' purpose-built for visual inspection, fine-tuned with few images, no-code, on 1,000+ production lines; careers page rendered no listings. sources prnewswire.com | -detailrevenue INR 51m, FY2024-25 inc42.com funding USD 6.2m, as of 2025-05 prnewswire.com Inc42 Datalabs shows FY25 revenue INR 5.1 crore (about USD 0.6m), roughly double FY24's INR 2.5 crore, with a loss after tax of INR 3.2 crore. sources inc42.com | - | Early | reads sensors | 30 LinkedIn 11-50 band at in.linkedin.com/company/frinks, midpoint used | india.entrepreneur.com in.linkedin.com |
| Kitov.ai | L2 | CAD-based automated visual inspection planning that combines AI with classical 3D vision to find cosmetic and assembly defects on production lines. | discrete, generalsourceSite lists defense, aerospace, automotive, medical devices, electronics and other markets. kitov.ai | CV | - | - | Early | reads sensors | 26 LinkedIn 11-50 range, 26 members listed | kitov.ai kitov.ai linkedin.com |
| Maddox AI | L2 | Machine-vision quality control software with collaborative annotation and model training, claiming over 80% fewer false rejects than classic vision. | any manufacturingsourceIndustries menu lists automotive, electrical engineering, electronics, consumer goods, medtech and pharma, industry and infrastructure, and food and beverage. maddox.ai | CV · MLwhyHomepage sells deep-learning defect detection that cuts false rejects; job ads on cyber-valley.de ask for self-supervised learning, domain adaptation and segmentation research. sources maddox.ai | -detailHomepage offers three delivery options (self-service software, retrofit onto existing inspection systems, end-to-end with hardware) and a 'pay only when your goals are achieved' risk-free trial; no prices published. sources maddox.ai | - | Early | reads sensors | - | maddox.ai linkedin.com |
| NeuroBot (矩视智能) | L2 | Browser-based low-code machine vision platform (600+ algorithms, 2D/3D/video) for defect detection, OCR, measurement and localisation that engineers train and run without writing code. | any manufacturingsourceMachine vision platform used in 3C electronics, automotive, pharma, food packaging and semiconductors. nb-ai.com | CVwhyOwn site: browser-based low-code deep-learning vision platform for OCR, segmentation-based defect detection, localisation and 3D point clouds; no job listings found. sources nb-ai.com | detail36Kr Japan (March 2020) says the cloud platform is free and revenue comes from licences for local deployment, with revenue of several million yuan and 30+ customers including Bayer; later rounds to Series A (March 2022) were 'tens of millions RMB', amounts undisclosed. sources 36kr.jp pitchhub.36kr.com | - | Early | reads sensors | ? | nb-ai.com |
| RESONIKS | L2 | QCFlex: acoustic-fingerprint end-of-line testing cell that gives pass/fail on surface and internal defects in seconds. | discrete, generalsourceHomepage industries section lists aerospace and defence, energy and utilities (wind blades, pipe systems), automotive and industrial equipment. resoniks.com | MLwhyHomepage: QCFlex resonates a part and an ML model trained on 100+ reference parts classifies its acoustic fingerprint, on premise; careers page showed no listings in the fetched content. sources resoniks.com resoniks.com | detailfunding EUR 2.6m, as of 2024 eit.europa.eu QCFlex is a CE-certified test cell 'delivered on one pallet', so each sale carries hardware; a further seed close (SHIFT Invest, Kvanted) was announced 2026-05-20 with no amount stated. | - | Early | reads sensors | 25 LinkedIn 11-50 range | resoniks.com linkedin.com |
| sentin | L2 | AI image analysis (sentin EXPLORER) for non-destructive testing: weld defects on X-ray and RGB images, wall thickness, surface inspection. | any manufacturingsourceSite says inspections are needed in many sectors from heavy industry to e-mobility, and its use cases page names oil and gas, manufacturing and infrastructure. sentin.ai | CVwhysentin EXPLORER analyses X-ray and RGB images for weld defects and wall thickness; of 6 open technical roles, 3 are software (desktop, backend/cloud, full stack), 1 data science/computer vision, 2 ML or AI application. hiring 6 open tech roles: ML 2, CV 1, software 3 sentin.ai | -detailBesides the EXPLORER software, the site sells NDT infrastructure setup and custom full-stack development services. sources sentin.ai | Bosch Rexroth (founder from) | Early | reads sensors | 6 LinkedIn 2-10 range; startbase 1-10 | sentin.ai startbase.com linkedin.com |
| SixSense | L2L3 | Sells an AI platform for semiconductor fabs and OSATs that classifies AOI defect images, predicts failures and links defects to tool and recipe data for root cause and yield. | semiconductors only | CV · MLwhyAbout page: AI defect detection and classification for semiconductor fabs plus learning from production signals for yield; of 3 open technical roles, 2 are backend/frontend software and 1 an AI implementation engineer. hiring 3 open tech roles: ML 1, software 2 careers.sixsense.ai sources sixsense.ai careers.sixsense.ai | -detailfunding USD 12m, as of 2025-07 techcrunch.com | Infineon (partner)detailInfineon · partner Infineon Singapore's AI initiative (more than USD 27m over three years) partners with SixSense to develop new AI solutions; Infineon's EVP of backend operations cited the collaboration on yield and cycle time. semi.org | Early | reads sensors | 35 sixsense.ai/about recruiter testimonial says team of 35 (undated); LinkedIn 11-50 band at linkedin.com/company/sixsenseai | sixsense.ai sixsense.ai techcrunch.com |
| Smartex | L2 | Camera hardware and computer vision fitted to knitting machines that detect fabric defects as they are produced. | textiles & apparel onlysourceRecord and tech.eu describe camera systems fitted to knitting machines for textile and knitting mills. tech.eu | CV · edge / IIoTwhyOf 6 technical openings on its Odoo job board, 2 are software (backend, intern), 2 hardware or IoT (VP of Hardware & Operations, IoT key-rotation thesis), 1 data engineer and 1 ML image-segmentation thesis, which fits camera hardware on knitting machines plus vision models. hiring 6 open tech roles: ML 1, data eng 1, software 2, embedded / edge 2 odoo.smartex.ai sources odoo.smartex.ai smartex.ai | detailHiring a VP of Hardware & Operations and IoT engineering shows each deployment ships devices onto customer machines. sources odoo.smartex.ai tech.eu | - | Early | reads sensors | 80 LinkedIn 51-200 range, 80 members listed | smartex.ai linkedin.com tech.eu |
| Braincube | L3L1 | Industrial AI platform for real-time process optimisation that finds best operating conditions from plant data and guides operators toward them. | process industriessourceRecord describes process manufacturing customers with a paper case (Norske Skog) and mining; homepage names only manufacturing and the paper case. braincube.com | ML · data engwhySite and about page describe 'Industrial AI' with 'proprietary AI algorithms' and a 'product-clone model' for real-time process optimisation on an Industrial IoT platform. | -detailfunding EUR 96m, as of 2023 braincube.com About page lists EUR 12m Series A (2018) and EUR 84m Series B (2023) and claims 250+ companies in 30 countries; no revenue or pricing published. sources braincube.com | Siemens (invested) | Customer track record | reads sensors | 125 LinkedIn 51-200 range (midpoint used) | braincube.com braincube.com linkedin.com |
| ENLYZE | L3LXL2 | Manufacturing data platform with retrofit connectivity and a unified namespace, delivering automated OEE, parameter-level traceability and root-cause analysis. | process industriessourceHomepage branch section lists plastic packaging, paper packaging, chemicals, metals, building materials, and food and life sciences. enlyze.com | data eng · edge / IIoT · analyticswhyProduct is retrofit edge connectivity, a unified namespace and automated OEE and downtime KPIs per homepage and pricing page; the one open role is a Senior Backend Engineer (Python). hiring 1 open tech role: software 1 enlyze.com | USD 10k to 100kdetaildeal size EUR 23.4k · per year, Starter, one site. Published price list: Starter from EUR 1,950 per month (one site, one edge device included, about EUR 23k a year); Optimize (up to 2 sites) and Innovate (unlimited sites) are custom; add-ons EUR 500 to 850 a month; services from EUR 1,500 a day. enlyze.com Public monthly subscription priced by tier of sites with edge devices included (extra devices EUR 25 a month), plus optional services at EUR 1,500 a day; entry ACV is about EUR 23k. sources enlyze.com enlyze.com | - | Customer track record | reads sensors | - | enlyze.com linkedin.com |
| Ethon | L3L2L1 | A factory analytics platform that unifies production data in a process knowledge graph and runs causal root-cause analysis, setpoint recommendations and camera inspection, delivered by forward-deployed engineers. | any manufacturingsourceBy Industry menu lists assembly, batch processes, consumer packaged goods, electronics, and metals and materials. ethon.com | causal · ontology · MLwhyPlatform page: 'proprietary reasoning model based on Causal AI' over unified, contextualised production data, plus a 'foundation model for manufacturing'; both open roles (Forward Deployed Engineer, Deployment Strategist) are deployment roles. hiring 1 open tech role: software 1 ethon.com | -detailImplementation includes 'a forward-deployed team of manufacturing, IT/OT, and AI experts' and both current openings are deployment roles, so services weigh on each deal; disclosed equity is USD 6.8m seed (2023) plus CHF 15m Series A led by Index Ventures (2024). | - | Customer track record | reads sensors | 50 LinkedIn 51-200 range (linkedin.com/company/ethonai); 'about 50 people' per a Siemens page cited in our 2026-10-08 report | ethon.com ethon.com ethon.com |
| Fabriq | L3L5 | Connected workforce platform that digitises Lean rituals (tier meetings, SQCDP, Gemba walks, audits, problem solving) and uses AI to turn field reports into standardised operational knowledge. | any manufacturingsourceIndustry menu lists aerospace and defense, food and beverage, automotive, luxury goods and cosmetics, pharmaceutical and medical, and logistics. fabriq.tech | analytics · RAGwhyHomepage: digital Lean workflows and SQCDP KPI tracking, with AI 'Similar Tickets' and operational knowledge search on the Enterprise tier; no job listings could be read. sources fabriq.tech fabriq.tech | detailPricing page says 'Priced per site, billed annually' with tiers by plant count (1, 2 to 10, 10+), Enterprise aimed at multi-region networks; SaaS with no installation; site claims 1,000+ sites and 80,000+ users; no amounts published. sources fabriq.tech fabriq.tech | - | Customer track record | documents only | 148 LinkedIn 51-200 range, 148 members listed | fabriq.tech linkedin.com |
| Factbird | L3L5LX | Plug-and-play edge sensors and PLC links feeding OEE and downtime analytics, plus connected-operations workflows and skills modules. | any manufacturingsourceIndustries menu lists food and beverage, consumer packaged goods, packaging and containers, industrial equipment, pharmaceutical, construction and building, and automotive. factbird.com | edge / IIoT · analyticswhyHomepage: pre-configured IIoT devices and sensors (Factbird DUO, VIEW, energy monitor) feeding OEE and downtime analytics, with an AI assistant and AI visual counter on top; careers page listed no positions in fetched content. sources factbird.com factbird.com | USD 10k to 100kdetaildeal size USD 16.8k · per year, one site with one line. Published price list: platform fee USD 1,000 per site per month plus USD 399 per line per month for Production Insights or Connected Operations, Knowledge Excellence USD 500 per site per month; one site with one line is about USD 17k a year, ten lines about USD 60k. factbird.com Transparent per-site plus per-line monthly pricing with sensors bundled; USD 16m raised Nov 2023 from Expedition Growth Capital and EIFO. | - | Customer track record | reads sensors | - | factbird.com iotforall.com |
| FourJaw Manufacturing Analytics | L3L0 | Plug-and-play MachineLink hardware and software that turn live machine activity plus operator and job data into OEE, downtime and capacity-loss analytics. | any manufacturingsourceIndustries menu lists aerospace, automotive, metal products, food production, medical devices, and building and construction. fourjaw.com | edge / IIoT · analyticswhyProduct is clip-on MachineLink power sensors feeding OEE and downtime dashboards; the homepage makes no ML or AI claim, and the careers page lists no open roles. hiring 0 open tech roles fourjaw.com sources fourjaw.com fourjaw.com | under USD 10kdetaildeal size GBP 4.3k · per year, 5 machines, Standard billed annually. Public price list: Standard GBP 90 per machine per month (GBP 72 billed annually), Pro GBP 180 (GBP 144 annually), minimum 5 machines, GBP 200 one-off MachineLink device on Standard, volume discounts up to 40% from 10 machines. A 5-machine Standard site is about GBP 4,300 per year. fourjaw.com funding almost GBP 4m, as of 2023-10-10 uktech.news Low-ticket self-serve SaaS: public per-machine prices of GBP 72 to 180 a month with self-installed hardware, so a typical customer is worth a few thousand to low tens of thousands of GBP a year; last raise was GBP 1.8m from Mercia/NPIF in October 2023 with about GBP 4m raised in total. sources fourjaw.com uktech.news | - | Customer track record | reads sensors | 29 press: own about page https://fourjaw.com/about states 29 employees; LinkedIn 11-50 range with 35 members listed | fourjaw.com fourjaw.com linkedin.com |
| Guidewheel | L3L0 | Sells a FactoryOps platform that clips power sensors onto any machine to track real-time utilisation and downtime and alerts teams to unusual machine patterns. | any manufacturingsourceServes plastics and packaging, food and beverage, metals, automotive, building materials and consumer goods (Hershey, US Steel, Cargill). guidewheel.com | edge / IIoT · ML · analyticswhyhiring 3 open tech roles: software 3 guidewheel.com | USD 10k to 100kdetaildeal size USD 25k · per year, stated floor. Pricing page: prices start at USD 25k annually; final price depends on number of machines and sites, contract length and hardware option. guidewheel.com Mid-market SaaS with a stated USD 25k per year entry price scaling with machines and sites, plus clip-on sensor hardware; USD 31m Series B in August 2024, no revenue or customer count disclosed beyond 'hundreds of manufacturers'. | - | Customer track record | reads sensors | 125 LinkedIn 51-200 band, midpoint used | guidewheel.com guidewheel.com linkedin.com |
| Raven.ai | L3 | Sells OEE and downtime software that pulls machine data to auto-tag downtime and prompts operators on HMIs to label the cause, giving supervisors a contextualised production timeline. | any manufacturingsourceServes discrete manufacturing, food and beverage and life sciences (Coca-Cola, Michelin, Sanofi, Stanley Black & Decker). raven.ai | edge / IIoT · analytics | -detailPricing page shows Core, Action and Premium tiers plus enterprise plans sold as a monthly OpEx subscription with no prices or unit shown; no revenue figure found. sources raven.ai | - | Customer track record | reads sensors | 125 LinkedIn 51-200 band at ca.linkedin.com/company/ravenai, midpoint used | raven.ai raven.ai ca.linkedin.com |
| TEEPTRAK | L3 | PerfTrak, QualTrak and PaceTrak boxes with plug-and-play sensors that capture stops, micro-stops, cycle times and scrap for real-time OEE, with operators logging stop causes. Machine learning is offered as a complementary solution through partner Jemba.ai, per its about page. | any manufacturingsourceIndustries menu lists automotive and mobility, aerospace and defense, food and beverage, pharma and cosmetics, electronics, and building materials, glass and mineral. teeptrak.com | edge / IIoT · analyticswhyPlug-and-play sensor boxes capture stops, cycle times and scrap for OEE dashboards; the about page says machine learning is offered as a complementary solution through partner Jemba.ai, and the only open tech role is a Senior Ruby on Rails developer. hiring 1 open tech role: software 1 teeptrak.com | detailPer-machine subscription bundling hardware, software, updates and support with volume discounts (own pricing article, no figures); raised EUR 5m in November 2024 from SHIFT Invest with XAnge and EDF while stated to be profitable, revenue multiplied by 6 between 2020 and 2024. sources teeptrak.com consultor.fr | Renault (founder from) | Customer track record | reads sensors | 37 LinkedIn 11-50 range, 37 members listed | teeptrak.com linkedin.com teeptrak.com |
| Arch Systems | L3L4 | Sells a factory intelligence platform that pulls machine, MES, ERP and SCADA data from electronics lines and uses AI agents to turn it into prescriptive actions for operators and managers. | electronics assembly, automotive, aerospace & defence | agents · ML · data engwhyOwn site and news centre the product on agentic AI (e.g. an award-winning Quality Defect Investigation Agent) on top of an 'Arch Data Twin' factory data foundation, with ML anomaly detection and root-cause analysis per Electronic Design; hiring gives no signal (careers page lists 2 roles but both ATS links show closed). hiring 0 open tech roles archsys.io | -detailEnterprise sale into large EMS players (Jabil, Flex, HARMAN, Syrma SGS named on its news page) with a USD 30m Vistara Growth investment dated September 2026 on its own news page. sources archsys.io trysignalbase.com | - | Early | reads sensors | 85 directory: builtin.com 85 total employees (https://builtin.com/company/arch-systems-inc/offices); LinkedIn 51-200 band | archsys.io archsys.io linkedin.com |
| MIPU Predictive Hub | L3L0 | Industrial AI platform for machine health, process intelligence and energy optimisation with digital twins and real-time energy monitoring for plants. Filings show the operating business is Mipu Machine Care S.r.l.; Mipu Predictive Hub is the holding. | any manufacturingsourceUse cases page groups by sector: manufacturing, chemical, pulp and paper, food, and pharmaceutical. mipu.eu | ML · edge / IIoT · analyticswhyMachine Health product page says sensors capture vibration, temperature and electrical data 24/7 and AI identifies and explains anomalies on a monitoring dashboard; the careers page lists no open roles, so ranking rests on product pages only. hiring 0 open tech roles mipu.eu | detailrevenue EUR 6.5m, 2020 forbes.it Forbes Italia put group revenue at EUR 6.5m in 2020 across predictive maintenance services, energy data science consulting and custom development with 160 active Italian clients; the filed entities are small and consulting-coded (holding Mipu Predictive Hub EUR 302k revenue and 1 employee in 2024, Mipu Machine Care EUR 2.81m revenue and 18 staff in 2024). | - | Early | reads sensors | 87 LinkedIn 51-200 range, 87 members listed | mipu.eu linkedin.com aziende.it |
| Shuyi Link (数益工联) | L3 | Digital lean factory software for discrete manufacturers that captures shop-floor and machine data to drive value-stream, efficiency and quality improvement. | discrete, generalsourceDigital lean factory software for discrete manufacturers, with auto parts and machinery named. pitchhub.36kr.com | edge / IIoT · data eng · analyticswhyshuyilink.com describes an edge data layer of machine IoT collection plus shop-floor terminals feeding 18 standard modules (real OEE, MTBF/MTTR, andon Pareto, traceability), with APS, process mining and quality big data listed only as add-on 'smart decision' items. sources shuyilink.com pitchhub.36kr.com | detailDelivery is project-shaped: on-site survey by digital-factory consultants and IoT engineers, a standard path of one week survey and three months implementation including network, servers and shop-floor devices, and PitchHub calls it an integrated hardware-plus-software solution; disclosed funding is Angel RMB 10m+ (2020), Pre-A RMB 10m+ (2021) and Series A RMB 100m (2022), so at least RMB 120m in total, with no revenue or price found. sources shuyilink.com pitchhub.36kr.com | - | Early | reads sensors | ? | pitchhub.36kr.com shuyilink.com |
| Smart Craft (株式会社Smart Craft) | L3L5L4 | Cloud MES for small and mid-size factories: tablet work instructions, actuals and defect capture, progress and OEE-style monitoring, quality records, inventory, IoT equipment data, and a natural-language AI assistant over the collected data. | discrete, generalsourceCloud MES for SME factories in automotive parts, electrical and electronic components, metalworking, industrial machinery and moulded products. smartcraft.jp | data eng · analyticswhyThe product is a cloud MES that records instructions, actuals and defects from tablets and aggregates them into progress and analysis dashboards, with a natural-language chat that visualises the collected data and IoT equipment collection as add-ons; the only tech role among 6 HERP openings is a Rails/React software engineer whose posting lists IoT auto-collection as future work and mentions no ML. hiring 1 open tech role: software 1 herp.careers | detailNo price is published; HERP states JPY 470m raised in total at a JPY 1.2bn valuation with 11 staff, and the OKI case shows a cloud SaaS rollout that still needed three months of on-site education by Smart Craft, starting in one inspection process and spreading across three business units. | Omron (invested) · Keyence (founder from)detailOmron · invested Omron Ventures signed an investment agreement with Smart Craft on 2026-06-11 (pre-Series A extension, amount undisclosed). omron.com Keyence · founder from Founder and CEO Fumiya Ukibe joined Keyence after university and did consultative sales of FA and IoT products to manufacturers, then Accenture. startupclass.co.jp | Early | reads MES / ERP | - | smartcraft.jp herp.careers thebridge.jp |
| Archlet | L4 | AI-native eSourcing with scenario optimisation and Spark agents for RFQs, auctions and bid analysis. | any manufacturingsourceSays it is built for every industry; industries menu lists consumer goods, food and beverage, healthcare and pharma, retail, logistics, manufacturing, automotive, oil and gas, chemicals and technology. archlet.io | OR · agents · analyticswhyProduct is scenario optimisation of supplier awards under capacity, incumbency, volume, regional-split and rebate constraints plus four LLM agents (setup, auction, scenario, summary) per the product pages; the solver itself is not disclosed and the 3 open roles include no ML or OR engineers. hiring 1 open tech role: other 1 archlet.jobs.personio.com (3 open roles: Enterprise Pre-Sales Engineer (US remote), Solution Consultant (Zurich), Security & Compliance Lead (Engineering, Spain); only the last is counted as technical) | -detailbuyer procurement Only the USD 10m Series A (30 Sep 2021, led by HV Capital) is confirmed on an opened page; the company claimed more than USD 20bn of spend optimised since 2019 and savings of up to 15%, but no price, contract value or revenue is public. sources startupticker.ch archlet.io | - | Customer track record | documents only | - | archlet.io linkedin.com |
| Arkestro | L4 | Predictive procurement platform that uses machine learning, game theory and behavioural science to simulate sourcing events, suggest prices and automate supplier negotiations. | any manufacturingsourceProcurement customers span oil and gas, manufacturing, automotive, life sciences, logistics and CPG (Koch, Chevron, Dover). arkestro.com | ML · ORwhyProduct suggests ML-predicted prices per line item from baselines and quote history and models weighted award scenarios across suppliers; the press release describes AI, behavioural science and game theory, and the only open engineering-side data role is a Lead AI/ML Engineer. hiring 2 open tech roles: ML 1, software 1 job-boards.greenhouse.io (6 open roles: Lead AI/ML Engineer, Senior SDET (QA), plus 4 non-technical (customer outcomes, professional services, BDR, solutions consultant)) | -detailbuyer procurement USD 36m strategic round in May 2025 led by Altira Group and Aramco Ventures (with NEA, KDT, Activant); the company claims 18.8% average savings and ROI in 45 days and announces multi-year enterprise agreements (Chevron, Valvoline, Nissan Americas), but no price, contract value or revenue is public. | Saudi Aramco (invested)detailSaudi Aramco · invested Aramco Ventures co-led the USD 36m strategic round of May 2025 with Altira Group. pulse2.com | Customer track record | reads MES / ERP | 125 LinkedIn 51-200 band (about 140 profiles listed); Welcome to the Jungle says 101-200, bouncewatch says 102; 125 is the band midpoint | arkestro.com arkestro.com linkedin.com |
| Circonomit | L4 | A web app in which a mid-sized manufacturer models its production decisions as connected blocks, simulates scenarios and optimises order sequence, capacity and stock with MILP and CP-SAT solvers. | discrete, generalsourceRecord names German Mittelstand in mechanical engineering, electronics, medical devices and plant engineering. circonomit.com | OR · simulation · data engwhyOf 3 open technical roles on Workable, one is an Applied Operations Research Engineer owning the MILP/CP solver engine (CP-SAT and Gurobi named), one a Senior Data Platform Engineer and one a Forward Deployed Engineer building optimisation or simulation models on messy ERP and Excel data. hiring 3 open tech roles: OR / optimisation 1, data eng 1, software 1 apply.workable.com (6 distinct roles (each listed twice, Cologne and Berlin, on the Workable widget API): Applied OR Engineer, Senior Data Platform Engineer, Forward Deployed Engineer (counted as software), Founding Product Manager, Delivery Lead, Junior Sales.) | detailfunding EUR 2.8m, as of 2025-07 circonomit.com | - | Customer track record | reads MES / ERP | ? LinkedIn 2-10 range (de.linkedin.com/company/circonomit); a June 2024 article said six employees; seven openings were live in Sept 2026; no current integer found | circonomit.com circonomit.com circonomit.com |
| Cofactr | L4 | Electronics supply-chain platform combining AI component intelligence, procurement execution across many suppliers and distributors, and ITAR-registered warehousing/kitting for hardware builders. | electronics assembly, aerospace & defence | data eng · agentswhyThe core is a component knowledge graph aggregating parts, specs and multi-distributor offers (Component Cloud API at graph.cofactr.com), plus Factor.io, acquired Feb 2025, whose AI reads supplier emails and attachments against the ERP; no technical roles are open, so hiring gives no signal. hiring 0 open tech roles jobs.lever.co. | detailbuyer procurement Pricing page: free to start, then a 7.5% service fee on managed orders with volume discounts, plus annual platform subscriptions and monthly storage, receiving, kitting and shipping charges; it earns like a procurement and logistics service on parts spend, not like pure software (Series A USD 17.2M Dec 2024). | - | Customer track record | reads MES / ERP | 125 LinkedIn 51-200 range (midpoint used) | cofactr.com linkedin.com |
| Enmovil | L4 | AI supply chain planning and visibility platform (CADDIE) that connects to ERP, TMS, WMS and carriers for demand forecasting, inventory planning, dispatch and load optimisation and disruption tracing across suppliers, plants and customers. | any manufacturingsourceServes automotive, FMCG, chemicals, cement, oil and gas and pharma (Maruti Suzuki, Hero MotoCorp, HPCL, Nestle). enmovil.ai | OR · ML · data engwhyProduct page lists dispatch, 3D load building, route and freight optimisation plus demand forecasting, delivered as a layer over ERP, TMS and WMS via prebuilt connectors; no algorithm detail published and the careers page lists no open roles, so the ranking rests on product pages only. hiring 0 open tech roles enmovil.ai (Page states 'No open roles right now.') sources enmovil.ai enmovil.ai | -detailfunding USD 9.5m, as of 2025-08 (sum of USD 6M Series A and USD 3.48M earlier funding, both stated in the Entrackr piece) entrackr.com sources entrackr.com enmovil.ai | - | Customer track record | reads MES / ERP | 183 directory: Inc42 (https://inc42.com/company/enmovil/) 183 employees; Built In says 150; LinkedIn page not located | enmovil.ai entrackr.com builtin.com |
| InterAgent (塔台智能網絡) | L4 | SARA, a cloud AI production-scheduling (APS) system for made-to-order SME factories that takes orders, inventory and capacity from Excel, ERP (SAP, Oracle, Digiwin) or MES and generates and re-generates schedules, delivery-date answers and shortage and bottleneck warnings in minutes. | any manufacturingsourceAPS for made-to-order SMEs in electronics, SMT, metal machining, auto parts, pharma, bicycles, printing and food packaging. interagent.io | OR · simulationwhyThe scheduling page shows the SARA engine placing each operation under due-date, machine capacity, tooling and material constraints and rescheduling up to 1,000 jobs per minute with a fast production simulation, plus finite-capacity material calculation; the founder's listed expertise is production scheduling optimisation. | under USD 10kdetaildeal size TWD 144k · per year, Starter plan, one plant. Public price list (updated 2026-10-29): Starter from NT$12,000/month (single plant), Professional from NT$30,000/month billed annually (single plant), Enterprise from NT$42,000/month (multi-plant), plus NT$1,000/month per extra manager seat, NT$500 per extra user, compute add-ons, and NT$300,000 one-off per ERP/MES integration. Starter is NT$144,000 a year. interagent.io Self-published list prices put a typical single-plant customer at NT$144k to NT$504k per year in subscription, with paid extras for ERP integration (NT$300k per system), consulting at NT$3,000/hour and a NT$100k government AI-tool subsidy that lowers entry cost; cloud by default, with private-cloud, on-prem and perpetual licence quoted per project for Enterprise. sources interagent.io interagent.io | - | Customer track record | reads MES / ERP | ? | interagent.io interagent.io meet.bnext.com.tw |
| Keelvar | L4 | Agentic sourcing software that runs procurement events end to end, from intake to award, with optimisation across direct materials, MRO, packaging and logistics. | any manufacturingsourceBy-industry menu lists automotive, food and beverage, hi-tech, manufacturing, pharmaceutical and retail. keelvar.com | OR · agentswhyCore product is Sourcing Optimizer (expressive bidding with conditional discounts and bundles, constraint-based award scenarios) plus Autonomous Sourcing agents and a GenAI orchestrator (Kai) that run RFx workflows from intake to award, per the product pages; hiring gives no signal (1 open role, non-technical). hiring 0 open tech roles keelvar.jobs.personio.com (1 open role: Senior Manager, Professional Services (US remote)) | -detailfunding USD 43m, as of 2022-05-16 keelvar.com buyer procurement Raised USD 43m in total by the May 2022 Series B, when it reported ARR growth of 113% in 2021 and 200% headcount growth; in 2021 it said it managed over EUR 80bn of annual customer spend; no revenue figure or price is public on an opened page. sources keelvar.com irishexaminer.com | - | Customer track record | reads MES / ERP | 120 LinkedIn 51-200 range, 120 members listed | keelvar.com keelvar.com linkedin.com |
| Kodiak Hub | L4 | AI-assisted supplier relationship management: onboarding, performance, risk and audit data across supplier tiers for procurement teams. | any manufacturingsourceIndustries menu lists manufacturing, food and beverage, energy and infrastructure, process industry, services, and retail and wholesale; claims suppliers across 20+ industries. kodiakhub.com | data eng · analytics · agentswhyProduct is a supplier data cockpit: onboarding workflows, scorecards and risk and compliance screening (sanctions, ESG, financial, media) with AI-assisted automation, per the homepage and the Nov 2024 funding note on building AI-driven data analytics tools; no open roles to weigh. hiring 0 open tech roles kodiakhub.com (careers page states there are no current openings) | -detailbuyer procurement Two disclosed rounds, SEK 55m (Apr 2023, led by BLQ Invest) and EUR 6m (Nov 2024, Oxx), with customer count rising from over 50 to over 80. | - | Customer track record | documents only | 15 company about page says 15 employees; LinkedIn shows 51-200 range | kodiakhub.com kodiakhub.com linkedin.com |
| LingMao SCM (领猫SCM) | L4 | Apparel supply-chain SaaS that connects brands with their contract factories and fabric suppliers for planning, sourcing, production tracking and settlement across the network. | textiles & apparel only | data eng · analyticswhyThe product is workflow SaaS that links brands, contract factories and suppliers across planning, sampling, purchasing, production and settlement, claiming 1,000+ API scenarios with 40+ OMS/ERP/POS/MES/RFID vendors, with a separate BI 'control tower' (领谋DI) for KPI dashboards and drill-down. | USD 10k to 100kdetaildeal size CNY 78k · per year, package A (20 users). Product page price cards: SaaS package A CNY 78,000/year (20 users), B CNY 98,000 (50 users), C CNY 128,000 (100 users), X CNY 158,000 (unlimited users), each with a first-year implementation fee from CNY 20,000 (X: 30,000) and unlimited external supplier accounts; self-hosted KA package on quote. cnunify.com Public list prices of CNY 78k to 158k per year plus a small one-off implementation fee, 100% renewal and 3x SaaS revenue growth reported by 36Kr in 2021 with ~300 customers, and 1,100+ brands claimed on the current site point to a pure-SaaS, mid-ticket business; funding is Series A ~RMB 100m (2021) and A+ USD 30m (2022) with no combined total published. | - | Customer track record | reads MES / ERP | 120 press: 36Kr Aug 2021 article says team grew from ~50 to 120 (https://www.36kr.com/p/1361234622989704); no newer figure found | 36kr.com pitchhub.36kr.com cnunify.com |
| Oplit | L4 | Production planning and scheduling platform (Oplit Core and the AI-first Oplit Kosmos with optimisation and reinforcement learning) fed by native ERP and MES connectors. | any manufacturingsourceIndustries section lists aerospace and defense, watchmaking and luxury, electronics, medical, food and beverage, automotive, industrial parts, pharma and cosmetics. oplit.com | OR · data engwhyHomepage describes hybrid optimisation engines (with reinforcement learning) scheduling work centres against OTD, WIP and margin, fed by native ERP and MES connectors; no reachable job listings, so this rests on product pages only. sources oplit.com | detail | - | Customer track record | reads MES / ERP | 38 LinkedIn 11-50 range, 38 members listed | oplit.com linkedin.com |
| Pactum | L4 | Autonomous procurement agents that negotiate deals with suppliers (including a single-source negotiation agent) and place purchase orders on behalf of large enterprises. | - | agents · MLwhyAutonomous negotiation agents built on ML, NLP and game-theoretic decision algorithms per Pactum's own explainer, with IEEE Spectrum describing a hybrid ML plus LLM system that generates contract options on price, delivery and billing terms; the careers board had no open roles. hiring 0 open tech roles careers.pactum.com (Teamtailor board states no open positions) | detailfunding USD 100m, as of 2025-06-09 (press release says over USD 100m after the USD 54m Series C) pactum.com buyer procurement Pricing page says one annual subscription priced by the customer's revenue band (under USD 1bn up to over USD 20bn) and the agents chosen, explicitly with no value-share or percentage of savings, 12-month minimum, implementation included, typical customers above USD 250m revenue; the June 2025 Series C (USD 54m, Insight Partners) took funding over USD 100m and reported ARR up 2.5x in a year, with no absolute revenue disclosed. | Tyson Foods (invested) | Customer track record | reads MES / ERP | 125 LinkedIn 51-200 range (midpoint used) | pactum.com linkedin.com |
| paretos | L4 | Socrates decision-intelligence system: demand forecasting, stock optimisation and production scheduling recommendations on ERP, CRM and BI data for the Mittelstand. | any manufacturingsourceSite lists only two branches, retail and manufacturing, plus other branches; no manufacturing sub-industry is named as a focus. paretos.com | ML · stochastic opt · agentswhyProduct is demand forecasting plus inventory, replenishment and production-schedule optimisation via 'Expert Engines for prediction and optimization', with an agentic LLM layer that explains recommendations; 3 of 3 open tech roles are applied-AI roles (data scientist, forward deployed engineer, staff AI backend engineer). hiring 3 open tech roles: ML 1, LLM / AI 1, solutions / deployment 1 paretos-gmbh.jobs.personio.de | detailbuyer supply chain head competes with SAP IBP, OMP, Kinaxis Website says pricing is a fixed annual license fee plus a one-time onboarding package, with extensions billed by project effort and cost depending on use-case scope; no prices, revenue or ARR published, and the Series A article only gives the CEO's claim of over 100% first-year ROI. sources paretos.com startbase.de | Mercedes-Benz (founder from)detailMercedes-Benz · founder from Co-founder and CEO Thorsten Heilig was COO of moovel Group, Daimler's mobility subsidiary (later folded into the BMW and Daimler REACH NOW venture), before founding paretos in 2020. paretos.com businessinsider.com | Customer track record | reads MES / ERP | - | paretos.com startbase.de |
| Tacto | L4 | Procurement AI for industrial mid-caps: supplier, spend, sourcing and compliance intelligence with RFQ, should-cost and negotiation agents over ERP data and external signals. | any manufacturingsourceIndustries menu lists machinery and plant engineering, medtech, chemicals and pharma, food, electronics, automotive, packaging and paper, and consumer goods. tacto.ai | agents · data eng · analyticswhyTacto hires only generalist 'Builders' (5 of 20 open roles on its Ashby board are Product & Technology build roles, none ML or OR specialists), and its Forward Deployed Engineer posting describes building data pipelines, agents and custom logic on customers' ERP spend, supplier and part data. hiring 5 open tech roles: software 4, solutions / deployment 1 jobs.ashbyhq.com | -detailbuyer procurement Job postings state more than 300 industrial companies run on Tacto, and the Series B article says the platform manages several billion euros of procurement volume. sources jobs.ashbyhq.com munich-startup.de | - | Customer track record | reads MES / ERP | 137 press: https://getlatka.com/companies/en.tacto.ai (137 employees, Sep 2025, citing LinkedIn) | tacto.ai tacto.ai getlatka.com |
| ThroughPut Inc | L4 | AI supply chain analytics software for demand sensing, capacity planning, inventory rebalancing and logistics planning on top of a client's existing ERP and operational data. | any manufacturingsourceLists heavy industry, cement, automotive, food and beverage, retail, defense, ports and rail customers. throughput.world | ML · OR · analyticswhyThe only data role on the careers page (Supply Chain Data Scientist) asks for ML forecasting with scikit-learn and statsmodels plus linear programming with Gurobi and Google OR-Tools, and Lokad's review describes the product as statistical-control bottleneck analytics that ranks operational problems. hiring 3 open tech roles: ML 1, software 2 throughput.world | over USD 1mdetaildeal size USD 1.2m · US Air Force SBIR Phase II contract (government R&D award). US Air Force (AFWERX) SBIR Direct-to-Phase II contract of USD 1,248,627, May 2024; a government R&D award, not a commercial ACV vktr.com Public money signals are US Air Force SBIR awards (USD 1.25m Direct-to-Phase II in May 2024, a second Phase III contract in Dec 2025 with value undisclosed) and a USD 6m angel round in 2022; Lokad notes the commercial subscription terms are not public and that a services-led Catalyst program runs the analysis for some customers. | - | Customer track record | reads MES / ERP | 28 directory: Latka (https://getlatka.com/companies/throughput) 28 employees as of Oct 2024 | throughput.world throughput.world getlatka.com |
| akeno | L4 | Autonomous production planning and scheduling for process manufacturers, reading and writing back to ERP and MES. | process industriessourceIndustries page lists chemicals, food and beverage, metals, pharma, mining, oil and gas, and building materials. akeno.ai | OR · data engwhyThe Senior Solutions Engineer posting describes production planning and scheduling software built by an 'AI & Optimization' team, with solution engineers writing operations-research code and leading ERP and MES data integrations at customer sites; the other open tech roles are a Python backend and a full stack engineer. hiring 3 open tech roles: software 2, solutions / deployment 1 apply.workable.com | -detailsources apply.workable.com thesaasnews.com | - | Early | reads MES / ERP | 25 LinkedIn 11-50 range | akeno.ai thesaasnews.com linkedin.com |
| ALGO ARTIS (株式会社ALGO ARTIS) | L4 | Optimisation-algorithm company (DeNA spin-off) selling Optium Plant production-planning optimisation and Planium, a packaged production scheduler for the chemical industry, plus ship and fuel planning for energy. | process industriessourceProduction planning optimisation for chemicals, steel and other process manufacturing, plus thermal power and shipping. algo-artis.com | ORwhyThe algorithm engineer posting says the company tackles scheduling and other optimisation problems with heuristic optimisation in C# and Python, parallelised in the cloud, and its staff place highly in AtCoder Heuristic Contests; products are ship, production and fuel planning optimisers (Optium, Planium). hiring 3 open tech roles: OR / optimisation 1, software 1, solutions / deployment 1 herp.careers | detailfunding JPY 2.5bn, as of 2026-06-09 algo-artis.com Cumulative equity funding is JPY 2.514bn (JPY 3.244bn including debt) after a JPY 1.538bn Series B led by Salesforce Ventures in June 2026; revenue sits behind INITIAL's paywall, and the engineering posting describes each optimisation project running from initial validation through MVP to full build and operation, so delivery is project-heavy. | Sony (invested) · Yokogawa (invested)detailSony · invested Sony Innovation Fund joined the JPY 690m round announced September 2023. thebridge.jp Yokogawa · invested Yokogawa Digital (横河デジタル, Yokogawa group) joined the JPY 690m round announced September 2023. thebridge.jp | Early | reads MES / ERP | - | algo-artis.com initial.inc herp.careers |
| Aolin Technology (傲林科技) | L4L3 | Enterprise-level digital twin that simulates a whole plant's operations to recommend procurement timing and volumes, inventory levels and production plans, aimed at steel and other process industries. | process industriessourcePlant-wide digital twin aimed at steel and other process industries, with steel mills, chemicals and coal named. tmtpost.com | simulation · MLwhyTMTPost and 36Kr PitchHub describe an enterprise-level operational digital twin that simulates supply-production-sales decisions (经营决策的模拟仿真) with AI models for key metrics; no hiring page or tech blog was found. sources pitchhub.36kr.com tmtpost.com | detailOnly funding is public (Pre-A about RMB 100m in Oct 2021, with steel group Rizhao Steel as investor); TMTPost describes project and integration-style engagements; no revenue or contract figures found. sources pitchhub.36kr.com tmtpost.com | Rizhao Steel (invested)detailRizhao Steel · invested Rizhao Steel Group invested in the Pre-A round of nearly RMB 100m in October 2021, alongside 熠辉嘉泰. pitchhub.36kr.com | Early | reads MES / ERP | ? | pitchhub.36kr.com tmtpost.com |
| ChAI | L4 | AI raw-material price forecasting and market intelligence, plus price-risk insurance, for manufacturers' procurement teams. | - | ML · data engwhyProduct is commodity price forecasting from thousands of alternative datasets (satellite imagery of harvests, shipping movements, market positions, weather) per London Loves Tech and Reinsurance News; no careers page found, so no hiring evidence. | detailfunding EUR 4.4m, as of 2022-11-23 funding.tech.eu buyer procurement Small raise (about EUR 4.4m over four rounds, last in 2022); two revenue lines, a forecasting platform (ChAI Insight) and an FCA-regulated parametric price-risk insurance product (ChAI Protect) backed by tier-1 underwriters, so part of revenue is insurance premium rather than software. | - | Early | documents only | 30 LinkedIn 11-50 range, 30 members listed | chaipredict.com linkedin.com |
| Didero | L4 | AI agents on top of a manufacturer's ERP that read supplier emails, POs and packing lists and update orders, confirmations and exceptions automatically. | any manufacturingsourceSite names manufacturers and distributors in industrials, chemicals, packaging, food and beverage, and automotive. didero.ai | agents · RAG · data engwhyTechCrunch describes an agentic layer on existing ERPs that reads emails, WeChat, POs and packing lists and executes updates; on its Ashby board the only engineering roles are forward deployed engineers (2 distinct titles across 7 location postings). hiring 2 open tech roles: software 2 jobs.ashbyhq.com | -detailfunding USD 30m, as of 2026-02-12 (Series A only; earlier seed amount not on an opened page) techcrunch.com buyer procurement USD 30m Series A (Chemistry, Headline, M12) with dozens of customers but no revenue or pricing disclosed; of 25 Ashby postings, 14 are forward deployed engineers or AI deployment strategists. sources techcrunch.com jobs.ashbyhq.com | - | Early | reads MES / ERP | 125 LinkedIn 51-200 range (midpoint used) | didero.ai linkedin.com techcrunch.com |
| Flexciton | L4 | Mathematical-optimisation planning and scheduling for semiconductor fabs (Flex Planner, Flex Local, Flex Global) that autonomously schedules lots to tools. | semiconductors onlysourceSite describes autonomous planning and scheduling for semiconductor manufacturing and wafer fabs. flexciton.com | ORwhyFortune and eeNews Europe describe the core as mixed integer linear programming with decomposition, solving fab schedules in under 15 minutes; the careers page currently lists only 2 roles, both customer implementation (no engineering roles open). hiring 0 open tech roles careers.flexciton.com | -detailfunding GBP 21m, as of 2021-10-22 eenewseurope.com buyer operations exec GBP 15m Series A in 2021 (GBP 21m total then); CEO quoted 7 to 10 percent fab efficiency gains worth USD 3m to 5m per month for a large fab. sources fortune.com eenewseurope.com | Intel (partner) · Seagate (partner)detailIntel · partner Partnership announced 25 September 2025 combining Intel Automated Factory Solutions (Factory Pathfinder, Operations Recon) with Flexciton's autonomous planning and scheduling as one end-to-end fab software offering. flexciton.com Seagate · partner Joint Flexciton and Seagate production planning project for Seagate's wafer fab, co-funded by Innovate UK (announced 2025); goes beyond a customer relationship as a co-development project. flexciton.com | Early | reads MES / ERP | 36 LinkedIn 11-50 range, 36 members listed | flexciton.com flexciton.com linkedin.com |
| Flowlity | L4 | AI supply chain planning (demand forecasting, inventory optimisation, supply and S&OP planning, collaborative planning with suppliers) built to be resilient to uncertainty. | any manufacturingsourceIndustries menu lists retail and e-commerce, manufacturing, spare parts management and wholesale; no manufacturing sub-industry is named. flowlity.com | ML · stochastic optwhyHomepage describes ML probabilistic demand forecasts with confidence ranges feeding dynamic safety stock and multi-echelon inventory optimisation under demand and lead-time variability; the careers page lists 3 roles, none engineering. hiring 0 open tech roles flowlity.com sources flowlity.com flowlity.com | USD 10k to 100kdetaildeal size USD 10k to 100k · per year, "tens of thousands" stated for mid-market. Homepage states SaaS pricing starting in the tens of thousands per year for mid-market companies, varying by SKU count, complexity, modules and horizon flowlity.com funding EUR 5m, as of 2022-03-11 funding.tech.eu buyer supply chain head Mid-market SaaS priced from tens of thousands per year, go-live under 2 months versus 12 to 24 for legacy APS; only EUR 5m raised (2022) and no newer round. sources flowlity.com funding.tech.eu | - | Early | reads MES / ERP | 28 LinkedIn 11-50 range, 28 members listed | flowlity.com linkedin.com |
| Fulcrum | L4L3 | machinery & metalworking only | OR · agents | -detailfunding USD 39.9m, as of 2023-08 (Series A) simplify.jobs buyer operations exec competes with ProShop, ECI JobBOSS2 / E2, Global Shop Solutions USD 39.9m raised (Bessemer, Battery); positioned as a migration target from legacy job-shop ERPs and claims 80 percent faster launch with an in-house launch team; no pricing or revenue published. | - | Early | reads MES / ERP | 75 LinkedIn 51-200 band; FundedIQ says 51-100; 75 is the midpoint of the narrower band | fulcrumpro.com linkedin.com vcnewsdaily.com | |
| Impactive AI (임팩티브AI) | L4 | Deepflow, a SaaS demand-forecasting and inventory tool for manufacturers that predicts 6-12 month product demand and raw-material prices, flags stock-out and overstock items, and proposes order quantities for purchasing teams. | any manufacturingsourceIndustry pages list F&B, retail, steel and construction, manufacturing and secondary batteries. impactive-ai.com | MLwhyHomepage says Deepflow picks the best of 224 forecasting models by competition over customer data joined with 50,000+ external series, with 70+ AI patents; the one open engineering role on its GreetingHR board is an ML engineer. hiring 1 open tech role: ML 1 impactive-ai.career.greetinghr.com | -detailbuyer supply chain head Sells demand forecasting (Forecast) and raw-material price forecasting (Materials) with a free PoC and a 2-week demo; also hiring an AI Transformation project manager for SCM. | - | Early | reads MES / ERP | 36 THE VC profile (thevc.kr/impactiveai): 36 employees as of 2026-09 | impactive-ai.com thevc.kr hellot.net |
| Kavida.ai | L4 | AI procurement agents ('Agent PO') that chase order confirmations, changes and documents with a manufacturer's suppliers after the PO is placed. Acquired by QAD Redzone, per its homepage. | discrete, general | agents · data eng · RAGwhyHomepage describes procurement agents automating 100+ supplier follow-up tasks (RFQs, invoice-PO matching, document validation) on SAP, Oracle and NetSuite integrations, with a natural-language copilot for data queries; no hiring page found after the QAD Redzone acquisition. sources kavida.ai | -detailbuyer procurement Acquired by QAD Redzone (terms not disclosed on the homepage), so standalone money signals no longer apply. sources kavida.ai | - | Early | reads MES / ERP | 30 LinkedIn 11-50 range (midpoint used) | kavida.ai linkedin.com |
| KREATIZE | L4 | Networked procurement platform that pools capacity from 70+ specialist supplier factories and lets industrial OEMs order mechanical parts through their own ERP, with KREATIZE matching, producing and QC-ing across the network. | machinery & metalworking onlysourceSupplies mechanical parts through its partner factory network to German industrial machinery and equipment manufacturers. kreatize.com | ML · data engwhyHomepage names FORM.1, an AI decision model scoring parts on complexity, quality risk, material, feasibility and availability, and 'AI-based capacity networking' to route ERP-placed orders to 70+ partner shops; no open roles listed (careers points to LinkedIn). sources kreatize.com kreatize.com | detailbuyer procurement Not a software vendor: KREATIZE sells finished parts at fixed guaranteed prices through its partner network and earns the margin, so revenue is gross part sales with manufacturing-broker margins; FY2024 accounts are filed (Northdata) but figures were paywalled. sources kreatize.com northdata.de | - | Early | reads MES / ERP | 125 LinkedIn 51-200 range (midpoint used) | kreatize.com linkedin.com |
| LevaData | L4 | Direct-materials sourcing platform that benchmarks component prices across buyers and distributors, scores supply risk from lead-time and lifecycle data, and assists supplier negotiations. | electronics assembly onlysourceSite serves electronics and high-tech OEM direct-material procurement (Cisco, Honeywell). levadata.com | data eng · analytics · agentswhyHomepage and about page describe ingesting ERP/PLM data, normalising it by manufacturer part number and benchmarking against prices other customers and distributors paid, with an agent (Zeta) that drafts quote requests and tracks negotiations; careers page only redirected to a form. | -detailbuyer procurement Value claim is about 12 percent spend savings in year one on direct materials for customers such as Cisco and Honeywell. sources levadata.com | Cisco (founder from)detailCisco · founder from Founder Rajesh Kalidindi founded LevaData in 2013 after years as Director of Supply Chain Operations at Cisco, managing multibillion-dollar component spend. levadata.com | Early | reads MES / ERP | 125 LinkedIn 51-200 range (midpoint used) | levadata.com linkedin.com |
| Lytica | L4 | Electronic-component spend analytics and risk intelligence built from anonymised cross-buyer transaction data, with agentic AI (Neo, Rik, Lisa) for negotiation and supplier strategy. | electronics assembly only | data eng · analytics · agentswhyHomepage says the core is billions of market signals from anonymised, aggregated buyer transaction data (100M+ parts, USD 550B+ spend analysed) with agents Neo, Rik and Lisa on top; the careers page lists no specific roles, only general submissions. hiring 0 open tech roles lytica.com sources lytica.com lytica.com | -detailbuyer procurement Enterprise customer list (Intel, Jabil, Flex, Medtronic, Safran and 40+ others) but no pricing, revenue or funding on opened pages. sources lytica.com | Nortel (founder from)detailNortel · founder from Founder Ken Bradley was Nortel's Chief Procurement Officer 1996 to 2002 (over USD 18bn material spend), then CEO of Coresim, before founding Lytica in 2005; Nortel was a telecom equipment maker, liquidated 2009 onward. kenjbradley.com | Early | reads MES / ERP | 125 LinkedIn 51-200 range (midpoint used) | lytica.com linkedin.com |
| Nulogy | L4L3 | Software that gives brand manufacturers real-time capacity, inventory and production visibility across their external co-packers and contract manufacturers, feeding constraint-aware plans into tools like Kinaxis Maestro. | food & beverage, packagingsourceSite targets CPG and food and beverage brands and their co-packers and contract manufacturers. nulogy.com | data eng · analyticswhySupplier Collaboration page describes real-time shared capacity, inventory and production status with co-packers and a certified feed into Kinaxis Maestro, with no specific AI method described; of 4 open roles, only a systems-integration consultant is technical. hiring 1 open tech role: data eng 1 nulogy.applytojob.com | detailbuyer supply chain head Cloud SaaS in tiered plans (Lean, Black Belt, Enterprise) priced by sites, machines and users, live in days; no revenue or funding figures on opened pages. sources nulogy.com nulogy.com | - | Early | reads MES / ERP | 125 LinkedIn 51-200 range (midpoint used) | nulogy.com nulogy.com linkedin.com |
| Partsimony | L4 | Manufacturing intelligence layer for hardware supply chains that links engineering decisions to a company's network of contract manufacturers, flagging supplier problems early. | discrete, generalsourceHomepage names hardware companies and processes: additive, machining and fabrication, injection molding, electronics, castings, assemblies. partsimony.com | data eng · analyticswhyHomepage describes a supplier-network platform giving 'data-driven recommendations from our ever-improving algorithm' for design, order logistics and workflow across contract manufacturers, without naming a method; no careers page found. sources partsimony.com | -detailsources partsimony.com | GE (founder from)detailGE · founder from Co-founder and CEO Richard Mokuolu worked at GE across manufacturing and supply chain management in aviation, oil and gas, locomotives and renewable energy (GE Aviation leadership program, Corporate Audit Staff, GE Renewable Energy, about 2013 to 2017). refashiond.com | Early | documents only | 6 LinkedIn 2-10 range (midpoint used) | partsimony.com linkedin.com |
| Skydisc (株式会社スカイディスク) | L4 | AI SaaS production scheduler 最適ワークス that turns order data into optimal machine and staff allocation plans for factories. | any manufacturingsourceHomepage says it supports DX for domestic manufacturing in general, with cases at Toyota Motor Kyushu, Miki Pulley and a chemical company. skydisc.jp | ORwhySaiteki Works (最適ワークス) is an AI SaaS production scheduler that turns orders into equipment and staff allocation plans. hiring 3 open tech roles: software 3 herp.careers | -detailARR JPY 100m, 2023 thebridge.jp funding JPY 2.5bn, as of 2023-05 (includes debt) thebridge.jp Saiteki Works hit JPY 100m ARR within 8 months of launch; the company also runs a DX Solution Studio for custom development, so revenue mixes SaaS and custom projects; JPY 2.5bn raised cumulatively including debt. | - | Early | reads MES / ERP | - | skydisc.jp herp.careers thebridge.jp |
| SourceDay | L4 | Supplier collaboration platform that syncs purchase-order dates, prices and quantities between a buyer's ERP and its suppliers, with AI handling confirmations and changes. | discrete, generalsourceRecord describes discrete manufacturers and distributors on mid-market ERPs, with no branch named. sourceday.com | data eng · agentswhyCore is certified two-way integration with mid-market ERPs (Infor, Epicor, NetSuite, Acumatica; 500+ integrations claimed) syncing PO changes with suppliers, with AI agents and intelligent document processing layered on for chasing acknowledgements and reading supplier emails; the 2 open roles on Built In are a sales rep and a technical product manager, so hiring gives no technical signal. hiring 0 open tech roles builtin.com sources sourceday.com builtin.com | USD 100k to 1mdetaildeal size USD 190k · per year, average of 3 deals (Vendr). Vendr transaction data: annual price from USD 21,000 to about USD 530,000, average contract value USD 190,319 (based on 3 completed deals) vendr.com buyer procurement Vendr reports annual contracts of USD 21k to 530k with an average around USD 190k, from only 3 deals, for an ERP-integrated SaaS; the homepage claims 123K suppliers activated and USD 6B direct spend per quarter flowing through it. sources vendr.com sourceday.com | - | Early | reads MES / ERP | 106 press: https://builtinaustin.com/company/sourceday-inc ('106 Total Employees') | sourceday.com builtinaustin.com |
| SupplyET | L4 | AI capacity and scenario planning across a multi-site production network, flagging bottlenecks before they occur. | any manufacturingsourceHamburg startup profile says the solution was designed from the start to apply across all industrial sectors; its one named customer is Bosch Rexroth. startupcity.hamburg | OR · simulationwhyPress describes the software as built on graphical modelling, model transformation and mathematical optimization, calculating capacity scenarios across a production network in seconds with visual simulation; no careers page or website was found to check hiring. | -detailbuyer operations exec Only a six-figure 2024 round plus BMBF grants is public. | Bosch Rexroth (founder from, partner)detailBosch Rexroth · founder from Founder Dr.-Ing. Christian Hochmuth joined Bosch Rexroth as a PhD student and built the production-network planning prototype there (live 2014); SupplyET takes that software concept further. startupcity.hamburg Bosch Rexroth · partner Bosch Rexroth was pilot partner for developing the next-generation software in its production network; global rollout across its 40+ sites underway since early 2026. startupcity.hamburg | Early | reads MES / ERP | 4 press: https://startupcity.hamburg/news-events/stories/supplyet-uses-ai-to-future-proof-production-networks (four core team members) | startupcity.hamburg |
| SyncFab | L4 | Procurement platform that connects aerospace, space and defense OEM buyers with a vetted network of compliant US suppliers, managing quotes, orders and compliance across companies. | aerospace & defence onlysourceSite connects aerospace, space and defense OEM buyers with US machining and fabrication suppliers. syncfab.com | data engwhyThe homepage describes cloud automation of supplier verification, quoting, order tracking and compliance (AS9100, ITAR, NIST 800-171) and data-driven buyer-to-supplier matching, with no AI or optimisation technique named; no careers page was found. sources syncfab.com | -detailbuyer procurement CB Insights lists one undisclosed seed round (2020, Starburst Aerospace Accelerator) and no revenue. sources cbinsights.com | Lockheed Martin (partner)detailLockheed Martin · partner April 2021 strategic deal giving Lockheed Martin access to SyncFab's supplier intelligence platform to match OEMs with Swissmem member suppliers across Switzerland. globenewswire.com | Early | documents only | 30 LinkedIn 11-50 range (midpoint used) | syncfab.com linkedin.com |
| Yuanmu Intelligence (圆木智能) | L4 | 今天排, an LLM-based scheduling agent for job shops that reads orders, routings, machine calendars, subcontracting and live progress and regenerates the production plan in seconds when orders or machines change. | machinery & metalworking onlysourceScheduling agent for job shops; customers are precision machining shops serving aerospace, semiconductor and medical device makers. jintianpai.com | agents · ORwhyThe product page says an AI agent and an optimisation algorithm jointly run each scheduling job, the agent translating goals and trade-offs and the algorithm handling global constraints and multi-objective balancing; press names LLM partnerships with Baidu and Zhipu. No job listings found. sources jintianpai.com news.pedaily.cn | -detailfunding , as of 2026-02 (tens of millions, exact figure not disclosed) news.pedaily.cn Only an angel round of tens of millions of RMB is public. sources news.pedaily.cn | - | Early | reads MES / ERP | ? | jintianpai.com news.pedaily.cn |
| Zentio | L4L5L3 | Agentic manufacturing OS: specialist agents for scheduling, staffing and master data over ERP and MES data, with an assistant for the shop floor. | any manufacturingsourceIndustries menu lists automotive, aerospace and defense, machinery, electronics, food and beverage, and consumer goods. zentio.ai | agents · ORwhyProduct page describes specialist agents (scheduling, scenarios, staffing, master data) over an 'Agentic Twin' that simulate sequences and replan in under 5 minutes. hiring 1 open tech role: software 1 join.com | -detailfunding EUR 1.4m, as of 2025-12 invest-in.berlin Pre-seed only (EUR 1.4m per the existing record) with pilot customers; no pricing or revenue public. sources invest-in.berlin | Ford (founder from) | Early | reads MES / ERP | ? | zentio.ai invest-in.berlin htgf.de |
| Augmented Industries | L5 | Flow Tool: AI that turns manuals, videos and technician knowledge into step-by-step work instructions and troubleshooting guides. | discrete, generalsourcetech.eu article says it serves industrial manufacturers and service providers, naming pneumatics, electronics and robotics, wind turbine maintenance and milling machines. tech.eu | RAG · workflow softwarewhyNo careers page could be opened (augmented-industries.com is a different company; flow-tool.com did not render). sources tech.eu | - | - | Customer track record | documents only | 25 LinkedIn 11-50 range | tech.eu linkedin.com |
| Datch | L5L3 | Sells an AI Diagnostic Agent and mobile EAM front end that lets maintenance technicians and operators troubleshoot equipment by voice or chat using work orders, manuals and schematics. | any manufacturingsourceCustomers are mainly energy utilities (Southern Company, Con Edison, Powercor) plus manufacturers (ABB, ATI). datch.io | RAG · agentswhyThe about page describes agentic AI diagnostic agents and voice-visual assistants for asset troubleshooting, and TechCrunch describes voice parsing of industrial commands on phones linked to enterprise systems; the 3 open roles (AE, sales engineer, SDR) are all go-to-market, so zero technical hires. hiring 0 open tech roles datch.io sources datch.io techcrunch.com | -detailfunding USD 15m, as of 2022-07 techcrunch.com buyer operations exec USD 15m raised by July 2022 with a plan to grow from 20 to 55-60 staff, yet LinkedIn showed 33 employees in Nov 2026 and current hiring is sales only; no pricing or revenue public. sources techcrunch.com datch.io | Siemens (founder from) · Boeing (invested)detailSiemens · founder from Co-founders Aric Thorn and Ben Purcell are former Siemens engineers (TechCrunch, Series A coverage 2022). techcrunch.com | Customer track record | reads MES / ERP | 33 LinkedIn 11-50 band, 'View all 33 employees' | datch.io datch.io linkedin.com |
| Dirac | L5L4 | Sells BuildOS, an AI build platform that turns CAD models into animated work instructions, manufacturing BOMs and bills of process executed at the station. | discrete, generalsourceCustomers are discrete assemblers in aerospace and defense, heavy equipment and energy hardware (Anduril, Blue Origin, Grimme). diracinc.com | 3D geometrywhyhiring 1 open tech role: solutions / deployment 1 job-boards.greenhouse.io | detailfunding USD 10.7m, as of 2025-08 alleywatch.com buyer engineering SaaS licensed per seat, site or assembly line (AlleyWatch), with a Siemens Teamcenter partnership and Founders Fund and Coatue backing a USD 10.7m round; the company benchmarks its market against CAD software, but no deal sizes or revenue are public. sources alleywatch.com aviationweek.com | Siemens (partner) · Northrop Grumman (founder from)detailSiemens · partner Strategic partnership announced Aug 2025 with the USD 10.7m seed: BuildOS is being integrated with Siemens Teamcenter PLM to reach Siemens' manufacturing customers. prnewswire.com Northrop Grumman · founder from Founder and CEO Filip Aronshtein worked in hardware engineering at Northrop Grumman (Jun 2020 to Sep 2021) and cites that time as the motivation for Dirac. alleywatch.com | Customer track record | documents only | 39 LinkedIn 11-50 band, 'View all 39 employees' | diracinc.com diracinc.com linkedin.com |
| Glartek | L5 | AI-native, mobile-first EHSQ and connected-worker platform with a copilot agent, voice input, risk assessment, root-cause analysis and procedure authoring for frontline teams. | any manufacturingsourceHomepage Industries section lists manufacturing, energy and utilities, chemicals, paper and pulp, pharmaceutical, and food and beverage. glartek.com | agents · workflow software | -detailbuyer operations exec A EUR 2m round in July 2025 (Portugal Ventures, August One) is the only figure found; pricing is quote-only and revenue undisclosed. sources prweb.com glartek.com | - | Customer track record | documents only | 33 LinkedIn 11-50 range, 33 members listed | glartek.com linkedin.com |
| Intoware (WorkfloPlus) | L5 | WorkfloPlus digital work instructions and inspections on mobile, desktop and smart glasses, with FloGenAI drafting instructions from a process description. | any manufacturingsourceHomepage lists construction and highways, logistics, machinery and equipment manufacturing, MRO, oil and gas, pharmaceutical, rail and utilities. intoware.com | workflow softwarewhysources intoware.com | - | - | Customer track record | documents only | 16 LinkedIn 11-50 range, 16 members listed | intoware.com linkedin.com |
| Operations1 | L5 | Connected-worker SaaS that replaces paper on the shop floor with digital work instructions, checklists, audits and training, now with a RAG assistant that answers workers from uploaded manuals with page citations. | any manufacturingsourceHomepage FAQ names manufacturing companies in mechanical engineering, plant engineering, vehicle construction and the food industry. operations1.com | workflow software · RAGwhyhiring 2 open tech roles: LLM / AI 1, software 1 career.operations1.com | -detailbuyer operations exec Pricing is demo-only; named customers include Daimler Truck, Trumpf, Liebherr and Thyssenkrupp, and the existing record notes a USD 12.5m Series A (2022) and a 2026 Fortino Capital investment of undisclosed size, but no revenue or deal size was found. sources operations1.com | - | Customer track record | reads MES / ERP | 54 LinkedIn 51-200 range (linkedin.com/company/operations1); careers page says '50+' | operations1.com operations1.com operations1.com |
| Skillnote (株式会社スキルノート) | L5 | Skills and training management system for manufacturing that maps worker skills to stations for placement and know-how transfer, now with an AI agent. | any manufacturingsourceHomepage calls the product specialised in manufacturing (製造業特化) without naming a main branch; cases and standards touch aerospace, pharma, food and water treatment. skillnote.jp | workflow software · RAGwhyThe career site lists engineer hiring as paused ('募集再開次第、お知らせいたします'). hiring 0 open tech roles career.skillnote.jp sources skillnote.jp career.skillnote.jp | detailcompetes with SAP Price page shows an initial implementation fee plus a basic plan quoted per person per month (no amounts), and the company also sells skill-management consulting; a PR Times release quoted on the career site says it partners with SAP SuccessFactors and served 120 companies and 100,000+ users as of Dec 2021. sources skillnote.jp career.skillnote.jp | Fujitsu (invested, partner) · Kawasaki Heavy Industries (partner) · Shin-Etsu Chemical (founder from)detailFujitsu · invested Fujitsu Ventures fund took part in the June 2023 third-party allotment (about JPY 410m round, Carbide Ventures lead). corp.skillnote.jp Fujitsu · partner Capital and business alliance with Fujitsu announced June 2023 to integrate Skillnote skill data with Fujitsu's manufacturing systems. corp.skillnote.jp Kawasaki Heavy Industries · partner Four-company collaboration (Skillnote, Fujitsu, Kawasaki Heavy Industries, SAP Japan) announced Nov 2022 to build a manufacturing DX platform service. corp.skillnote.jp Shin-Etsu Chemical · founder from Founder and CEO Takashi Yamakawa joined Shin-Etsu Chemical after university, two years in a plant then technical sales and business development in the electronic materials division. storium.jp | Customer track record | documents only | 125 LinkedIn 51-200 range (midpoint used) | skillnote.jp corp.skillnote.jp linkedin.com |
| Squint | L5L3 | Sells an industrial AI platform that captures tribal knowledge and manuals into AR/mobile procedures, answers operator questions, and runs industrial agents such as video-based time-and-motion analysis. | any manufacturingsourceIndustries menu lists automotive, equipment, food and beverage, CPG, oil and gas, power and utilities and heavy industrials. squint.ai | RAG · agents · CVwhyPlatform turns tribal knowledge and manuals into AR/mobile procedures with AI workflows and is pushing 'agentic manufacturing' (Series B release); the existing record adds a video time-and-motion agent. The Ashby board lists 7 roles, only 1 technical (Deployment Engineer), the rest sales, IT, exec and an open application. hiring 1 open tech role: solutions / deployment 1 jobs.ashbyhq.com | detailfunding USD 59m, as of 2025-08 prnewswire.com buyer operations exec Raised over USD 59m; the Series B release cites a Fortune 50 customer expanding to 70+ more sites and a Fortune 500 rollout to 10,000 field technicians. sources prnewswire.com | - | Customer track record | documents only | 79 LinkedIn 51-200 band, 'View all 79 employees' | squint.ai squint.ai prnewswire.com |
| Testify | L5 | Digital checklists, inspections and audit documentation for plants, with an MCP server so LLM assistants can query and report on the data. | any manufacturingsourceBranches menu lists machinery and vehicle building, railway and infrastructure, chemical industry, food industry, aerospace and defence, and other branches; site says Testify works in every branch. testify.io | workflow softwarewhysources testify.io testify.io | under USD 10kdetaildeal size EUR 4.8k · per year, Starter plan (plus EUR 3,360 setup). List prices EUR 397/month (Starter, unlimited users, 1,000 checks), EUR 1,595/month (Standard, 50 users, 10,000 checks), EUR 3,749/month (Advanced, 100 users, 100,000 checks), annual billing, plus a one-time EUR 3,360 setup fee; Standard and Advanced are about EUR 19k to 45k a year, Starter about EUR 4.8k testify.io buyer quality head Public list prices put a customer at roughly EUR 4.8k to 45k a year plus a EUR 3,360 setup fee, a small-ticket SaaS motion. sources testify.io | - | Customer track record | documents only | 25 LinkedIn 11-50 range | testify.io testify.io linkedin.com |
| Augmentir | L5L3 | Sells an AI-native connected worker platform: digital work instructions, skills tracking, an industrial GenAI assistant (Augie) and no-code AI agents for frontline manufacturing operations. | any manufacturingsourceLists automotive, building materials, chemicals, consumer goods, food and beverage, pharma, mining and oil and gas. augmentir.ai | RAG · agents · MLwhyHomepage leads with Augie, an industrial GenAI suite that generates procedures from Excel, Word, PDFs, images or video, an Agent Studio for custom industrial AI agents, and 'Smart AI' over worker execution data (15M+ time-and-motion studies analysed). sources augmentir.ai augmentir.ai | -detailPricing is request-only ('scales from one line to global operations'); no revenue or deal size found. sources augmentir.ai | Fortive (partner)detailFortive · partner 2024 partnership agreement to integrate Augmentir with Fluke Reliability's AI-powered enterprise asset management solution for maintenance teams (Frost & Sullivan award write-up). frost.com | Early | reads MES / ERP | 38 LinkedIn 11-50 band, 'View all 38 employees' | augmentir.ai augmentir.ai augmentir.ai |
| Azumuta | L5L3 | Connected-worker platform combining digital work instructions, audits, training matrices and quality checks, with AI document conversion. | any manufacturingsourceIndustries menu lists pharmaceuticals, food and beverage, high tech, aerospace and defense, chemicals, automotive, metallurgical, machinery and electronics. azumuta.com | workflow software · RAGwhyhiring 2 open tech roles: software 2 azumuta.recruitee.com sources azumuta.com azumuta.recruitee.com | detailfunding EUR 8m, as of 2025-11 tools.prnewswire.com buyer operations exec Pricing page is quote-only but says plans are SaaS, monthly or annual, no perpetual licence and no hardware needed, with optional paid Customer Success packs and a 2-month joint pilot; it claims 125,000+ operators.The Nov 2025 release reports an EUR 8m Series A (Keen Venture Partners lead) and nearly 100 manufacturer customers. sources azumuta.com tools.prnewswire.com | - | Early | documents only | 25 LinkedIn 11-50 range | azumuta.com tools.prnewswire.com linkedin.com |
| Edmund AI | L5L3 | Industrial debugging assistant whose agents read PLC programs, robot programs, ePLAN schematics, SQL/MES data and manuals so maintenance teams find line-stop root causes via chat. | discrete, generalsourceHomepage names automotive Tier-1/Tier-2 suppliers and packaging first, then metalworking, plastics and polymers, engineering and tooling, industrial automation, heavy industry and consumer goods. edmundai.com | agents · RAG · ontologywhySite lists separate agents that read PLC logic across vendors, robot programs, ePLAN schematics, SQL/MES data and manuals, feeding a graph and answering with page-level citations (e.g. tracing a servo enable in TIA Portal SCL); the careers page shows 1 open role, a Senior Acquisition Engineer in sales, so no technical hiring now (team is 3 founders and 15 engineers). hiring 0 open tech roles edmundai.com sources edmundai.com | -detailfunding EUR 500k, as of 2025-02 en.ain.ua buyer plant manager Only a EUR 500k pre-seed (Feb 2025, LightHouse Ventures lead) is public; the site sells via a free root-cause analysis offer and has no pricing. sources edmundai.com en.ain.ua | Festo (founder from) | Early | reads MES / ERP | 18 LinkedIn 11-50 range, 18 members listed | edmundai.com linkedin.com en.ain.ua |
| HiScene (亮风台) | L5 | AR glasses and software that overlay step-by-step work instructions, equipment information and remote expert guidance on the shop floor, using computer vision. | any manufacturingsourceCompany profile names smart manufacturing, public safety and interactive entertainment; manufacturing branch not specified (cases: Gree Electric, AECC aero engines). hiscene.com | CVwhyCompany profile describes a 'spatial AI computing platform' behind AR glasses and remote-guidance software; the recruitment link (hiscene.zhiye.com) returns a not-found page, so no hiring data. sources hiscene.com | detailCompany timeline lists rounds including B (RMB 100m), B+ (RMB 120m) and C+ (RMB 270m, Sept 2021), at least RMB 490m across those three; it sells AR glasses with software, so hardware margins apply. No revenue found. sources hiscene.com | Huawei (partner) · OPPO (partner) · China Baowu (partner)detailHuawei · partner Company timeline: signed an important cooperation agreement with Huawei for comprehensive strategic cooperation in AR (January, listed after the Sept 2021 C+ round); also launched a 5G AR product solution jointly with Huawei and carriers. hiscene.com OPPO · partner Company timeline: worked with OPPO on its first UGC AR product, O-Video (early years after the 2012 founding). hiscene.com China Baowu · partner HiScene's company profile names Baosight (宝信), the industrial software arm of China Baowu Steel, among the industry giants for which it is an AR platform partner; Baowu Steel is also listed as a deployment. hiscene.com | Early | documents only | 125 LinkedIn 51-200 range (midpoint used) | hiscene.com hiscene.com linkedin.com |
| Kaminashi (株式会社カミナシ) | L5 | Frontline-worker SaaS for factories: digital checklists and production records, maintenance, and training with AI-generated manuals and translation. | any manufacturingsourceRecord says food manufacturing leads (site claims 10% of listed food companies), alongside machinery and materials, process manufacturing, hotels and logistics. kaminashi.jp | workflow software · RAGwhyhiring 14 open tech roles: LLM / AI 1, software 12, embedded / edge 1 herp.careers sources kaminashi.jp herp.careers | detailbuyer plant manager Pricing is an initial fee plus a monthly subscription scaled by number of user accounts (amounts not public); it has 50+ business staff and 17 open sales roles, and no ARR or funding figure was found. sources kaminashi.jp herp.careers | - | Early | documents only | 125 LinkedIn 51-200 range (midpoint used) | kaminashi.jp linkedin.com |
| Leucine | L5L3 | Sells an AI platform for pharma manufacturing: AI-powered batch execution and electronic batch records, digital logbooks, QMS with an AI deviation investigator, and lab execution. | pharma & life sciences only | agents · RAG · data engwhyProduct converts paper SOPs into digital batch-record workflows with an LLM framework and does AI root-cause analysis of deviations (TechCrunch 2023); open tech roles are integration, forward-deployed and MES engineers, no ML research roles. hiring 3 open tech roles: data eng 1, software 2 apply.workable.com sources apply.workable.com techcrunch.com | detailbuyer quality head competes with Siemens, Rockwell, MasterControl Cloud SaaS with tiered pricing that rises with volume and 30+ customers running 300+ facilities at its USD 7m Series A in 2023; 3 of 9 open roles in 2026 are MES consultants or delivery architects. sources techcrunch.com apply.workable.com | Ecolab (invested)detailEcolab · invested Ecolab, described as a strategic investor, led the USD 7m Series A in October 2023. techcrunch.com | Early | reads MES / ERP | 116 LinkedIn 51-200 band, 'View all 116 employees' | leucine.io linkedin.com |
| LIGHTz (株式会社LIGHTz) | L5 | Structures the tacit know-how of veteran engineers into 'Brain Models' and sells blooplinter, a knowledge platform with 3D-CAD similarity search, automated requirement checks and past-issue matching for design and maintenance teams. | any manufacturingsourceKnowledge platform for large manufacturers in design, production engineering and maintenance; no industry focus named on its site. lightz-inc.com | ontology · RAG · agentswhyCore is the BrainModel, a tree-structured model of veteran engineers' judgment criteria, which the Hanchika AI product combines with RAG search and task agents (IPROS product page); the HERP job board lists no engineering roles, only one dev-department project manager next to consulting and sales. hiring 0 open tech roles herp.careers | detailfunding JPY 1.8bn, as of 2024-11 thebridge.jp buyer engineering No public price (IPROS listing says pricing is negotiated and delivery combines consulting with software); the job board has a consulting department with 2 open roles and no open engineering roles. sources pr.mono.ipros.com herp.careers | Omron (invested)detailOmron · invested Omron Ventures invested in LIGHTz, announced 2025-05-29. omron.com | Early | documents only | - | lightz-inc.com lightz-inc.com thebridge.jp |
| Sourcemap | L5 | any manufacturingsourceRecord describes Global 1000 brands in apparel, food, automotive and electronics. sourcemap.com | data eng · analyticswhyProduct is n-tier supplier data collection and mapping across a network of over two million registered businesses, with AI-powered alerts for watchlist screening and disruptions (Series B release 2023); the only open role is a senior Node/TypeScript frontend engineer, no ML roles. hiring 1 open tech role: software 1 sourcemap-inc.breezy.hr | -detailUSD 20m Series B in June 2023 with over 250 brands as customers and 100+ added in the prior year (company release); no price, revenue or total-raised figure on an opened page (a ZoomInfo revenue estimate of USD 17.4m on getlatka was not used because it is an estimate). sources sourcemap.com | - | Early | documents only | 125 LinkedIn 51-200 range (midpoint used) | sourcemap.com sourcemap.com linkedin.com | |
| Team Solution (팀솔루션) | L5 | DXE digital-twin modules for heavy industry, including DXE Manual AI (launched Oct 2026) which turns PDF equipment manuals and 2D drawings into step-aligned 3D work manuals with animations, editable by natural-language prompts, cutting manual creation from about a month to a week. | discrete, general | 3D geometry · simulation · RAG | - | - | Early | documents only | - | timsolution.io newswire.co.kr dataeconomy.co.kr |
| Kinaxis (Maestro) | L6L4 | Concurrent supply chain planning platform (Maestro) with scenario modelling and OEM-to-tiered-supplier collaboration. | any manufacturingsourceIndustries menu lists aerospace and defense, automotive, chemical, consumer products, high-tech, industrial, life sciences and logistics. kinaxis.com | OR · MLwhyOf 23 distinct open technical roles, 5 are operations research or solver roles (incl. Head of Operations Research and a production-scheduling algorithm developer) and 4 are ML roles (incl. a time-series and demand-forecasting applied scientist), plus 3 LLM/agentic AI roles. hiring 23 open tech roles: ML 4, OR / optimisation 5, data eng 2, LLM / AI 3, software 9 careers-kinaxis.icims.com | over USD 1mdetaildeal size more than USD 1m · per year, average contract value of each of its 20+ largest deals in 2025. Q4 2025 release: more than 20 deals above $1 million in average annual software contract value, and over 100 software deals above $1 million in total contract value kinaxis.com revenue USD 548m, 2025 kinaxis.com buyer supply chain head FY2025 revenue USD 548.0m (SaaS 362.4m, professional services 155.4m, about 28%), ARR USD 433m up 20%, gross margin 65% and adjusted EBITDA margin 25%; listed on TSX so no venture funding total applies. sources kinaxis.com | Nortel (founder from)detailNortel · founder from Co-founder Duncan Klett held various positions at Bell-Northern Research, Nortel's R&D arm, before co-founding Kinaxis in 1984 (the three co-founders were also former Mitel engineers; Mitel left out as below the size bar). investors.kinaxis.com | Established | reads MES / ERP | ? | kinaxis.com kinaxis.com kinaxis.com |
| o9 Solutions | L6L4 | Integrated planning platform (Digital Brain) with supplier collaboration, Tier 2/3 mapping and what-if analysis on sourcing and supplier failure. | any manufacturingsourceIndustries page lists automotive, consumer products, energy and resources, high tech, industrial manufacturing, life sciences, retail, telecom and wholesale distribution. o9solutions.com | ontology · MLwhyPlatform page says it is built on the Enterprise Knowledge Graph as a digital twin of the enterprise with AI/ML models predicting demand; of 15 open technical roles on its Workday board, 10 are DevOps or software architects and 3 are data science or MLOps, with no OR roles, while most of the 75 postings are delivery and consulting roles. hiring 15 open tech roles: ML 3, data eng 2, software 10 o9solutions.wd5.myworkdayjobs.com | -detailbuyer supply chain head Private: raised USD 116m in August 2023 at a USD 3.7bn valuation (up from 2.7bn in January 2022) with ARR growth of 55% year on year in Q2 2023; FY2025 release reports 28 consecutive quarters of ARR growth but no dollar figure; roughly half of its 75 open postings are delivery, functional-consultant and solution-architect roles. | Texas Instruments (founder from)detailTexas Instruments · founder from Co-founder and chairman Sanjiv Sidhu worked at Texas Instruments' AI laboratory in Dallas, building simulation software for TI production planning, before founding i2 Technologies and later o9. en.wikipedia.org | Established | reads MES / ERP | 2,500 Dallas Innovates article 2023-11-16: 2,500 employees, 17 offices (https://dallasinnovates.com/how-dallas-o9-solutions-built-a-multi-billion-dollar-digital-brain/) | o9solutions.com dallasinnovates.com |
| Resilinc | L6 | Multi-tier supplier mapping and risk platform that simulates disruption impact across suppliers, sites and parts and quantifies revenue at risk. | any manufacturingsourceSolutions by industry: aerospace and defense, healthcare and life sciences, automotive and industrial, high tech and semiconductor. resilinc.ai | data eng · agents · RAGwhyThe core asset is a multi-tier supplier and part-site map plus event monitoring over 100 million+ sources (homepage), now sold as agentic AI and EventWatchAI; the 5 open technical roles on Lever are a data engineer and 4 software or SRE roles, with no ML or OR openings. hiring 5 open tech roles: data eng 1, software 4 jobs.lever.co | over USD 1mdetaildeal size USD 1.2m · US HHS contract, 2020 to 2025. USAspending: HHS award 75A50120C00188 'Medical supply chain risk management data', USD 1,203,879.69 over about 4.8 years (roughly USD 250k a year); a DoD task order FA714623C0012 was USD 688,500 for 2023-05 to 2024-07 usaspending.gov Vista Equity Partners made an undisclosed growth investment in May 2023 from its Endeavor Fund, which targets companies with at least USD 10m recurring revenue; federal awards show roughly USD 250k to 690k a year per government customer. | Cisco (founder from) · Flex (founder from)detailCisco · founder from Founder Bindiya Vakil worked in sourcing and procurement roles at Cisco before founding Resilinc (her own words in a video interview). youtube.com Flex · founder from Founder Bindiya Vakil worked in sourcing and procurement roles at Flextronics before founding Resilinc. youtube.com | Established | reads MES / ERP | 344 Built In company profile: 344 total employees (https://www.builtinchicago.org/company/resilinc) | resilinc.ai builtinchicago.org |
| SupplyOn | L6L4 | Multi-enterprise supply chain network for manufacturers whose Capacity Assessment maps supplier workcenters and simulates demand vs capacity scenarios to find bottlenecks; also Catena-X DCM. | any manufacturingsourceAbout page lists aerospace and defense, automotive, electronics, medtech and pharma, plant engineering and railway. supplyon.com | data eng · analyticswhySupplyOn is a multi-enterprise collaboration network (200+ buy-side customers, 140k business partners) whose value is integrating buyer and supplier data across purchasing, quality and supply chain processes; its 16 open roles include 4 technical ones, all DevOps, platform or database roles, and none in ML or optimisation. hiring 4 open tech roles: software 4 jobs.supplyon.com | -detailbuyer procurement Bain Capital agreed in July 2026 to buy SupplyOn from Aumovio, Bosch, Schaeffler and ZF on undisclosed terms; no revenue figure was found on an opened page (Northdata financials are behind a paywall). | Bosch (JV) · Continental (JV) · Schaeffler (JV) · ZF (JV)detailBosch · JV Shareholder of SupplyOn until Bain Capital agreed to buy it in July 2026. supplyon.com Continental · JV Continental AG was a shareholder; the seller in the 2026 Bain deal is AUMOVIO, Continental's automotive spin-off. supplyon.com Schaeffler · JV Shareholder of SupplyOn until Bain Capital agreed to buy it in July 2026. supplyon.com ZF · JV Shareholder of SupplyOn until Bain Capital agreed to buy it in July 2026. supplyon.com | Established | reads MES / ERP | 500 SupplyOn about page: '500+ employees worldwide' (https://www.supplyon.com/en/about_us/) | supplyon.com supplyon.com |
| Optilogic (Cosmic Frog) | L6 | Supply chain design platform that models the network from supplier to customer and runs optimisation and simulation what-ifs. | any manufacturingsourcePage names consumer products, manufacturing, high-tech, industrial, life sciences, automotive, logistics, energy oil and gas, and retail. optilogic.com | OR · simulationwhyCosmic Frog builds network models from supplier to customer and runs optimisation (asset footprint, transportation lanes and modes) plus simulation with real-world variability and per-scenario risk ratings, per its product page; the Rippling careers board had no open roles, so hiring gives no signal. hiring 0 open tech roles ats.rippling.com | detailbuyer supply chain head Raised a USD 40m Series B in April 2025 led by NewRoad Capital Partners with MK Capital and Moore Strategic Ventures; product page offers a free account, but no paid pricing, revenue or total funding was found on an opened page. sources vcaonline.com optilogic.com | - | Customer track record | reads MES / ERP | ? | optilogic.com |
| Cosmo Tech | L6L4 | Prescriptive simulation digital twins of whole supply chains and asset networks that let planners test scenarios and pick the best production and investment plan. | any manufacturingsourceIndustries menu lists automotive, consumer packaged goods, aerospace and defense, manufacturing, energy and utilities, and mobility. cosmotech.com | simulation · stochastic optwhyThe product builds 'Prescriptive Simulation Twins' of supply-chain networks and runs what-if scenarios to find 'the most robust plan' under uncertainty, without needing historical records. | -detailThe about page states EUR 32m raised to date and 70+ employees after 16 years. sources cosmotech.com | - | Early | reads MES / ERP | 70 press: own about page https://cosmotech.com/about-us/ states '+70 Employees'; LinkedIn 51-200 range with 120 members listed | cosmotech.com cosmotech.com cosmotech.com |
| Resilire (株式会社Resilire) | L6L4 | Supply-chain risk management cloud that maps tier-2 and deeper suppliers, sites and parts as a tree, detects disasters and geopolitical events in real time, and uses AI to assess how macro changes hit the network. | any manufacturing | data eng · simulation · agentswhyCore is collecting and updating tier-2 and tier-3 supplier data plus impact and PSI scenario simulation and an AI agent layer launched 2025-11 (LNEWS); open tech roles are a Go/React product engineer and an SRE, no ML or OR roles. hiring 2 open tech roles: software 2 green-japan.com | detailfunding JPY 3bn, as of 2026-08-19 thebridge.jp buyer procurement JPY 1.73bn Series B first close in 2026-08 (cumulative about JPY 3.03bn, including debt per the article) from DNX, Salesforce Ventures and Sumitomo's CVC. sources thebridge.jp | - | Early | documents only | - | lnews.jp thebridge.jp resilire.jp |
| Barbara | LX | Industrial edge platform that lets OT teams deploy, run and monitor containerised apps and AI models on plant edge devices, with industrial connectors and MLOps. | any manufacturingsourceHomepage names digital grid, smart water and smart manufacturing as its sectors, plus other Industry 4.0 uses; no manufacturing branch is specified. barbara.tech | edge / IIoTwhyThe product is an edge orchestration platform that deploys and manages containerised apps and AI models on distributed plant and grid devices (funding coverage describes it as an edge AI orchestration platform), so device fleet management, not model building, is the core; the careers page lists no open positions. hiring 0 open tech roles barbara.tech | detailSmall funding so far: a EUR 4.5m round reported 2026-04-20 for expansion into Europe, the Middle East and the US, on top of earlier public grants (Horizon 2020, CDTI, Hazitek) listed on its site; no price list, revenue or total-funding figure was found on an opened page. sources kurrant.com barbara.tech | Saudi Aramco (invested)detailSaudi Aramco · invested Aramco Ventures led the EUR 4.5m strategic round announced 2026-02-18. barbara.tech | Customer track record | reads sensors | 86 LinkedIn 11-50 range (Spanish page: 'De 11 a 50 empleados') but 86 members listed; band and member count disagree | barbara.tech es.linkedin.com |
| Cybus | LX | Cybus Connectware, an on-premise factory data hub that connects OT protocols to IT systems and distributes standardised plant data in real time. | any manufacturingsourceRecord lists automotive, machinery, metalworking, food and agricultural equipment customers (Porsche, CLAAS, Liebherr, Krone, Miele). cybus.io | edge / IIoT · data eng · streamingwhyConnectware is an on-premise factory data hub that connects OT protocols to IT systems and distributes plant data in real time; the careers page shows no open technical roles, only a general application, so the product pages are the only evidence. hiring 0 open tech roles cybus.io | detailConnectware is a term license with a stated expiry and no public price; Cybus raised a EUR 8m Series A led by PwC Holdings Germany in March 2022, when it had 30 staff and expected to triple annual sales, and TeamViewer later joined as an investor, but no total funding or revenue figure was found on an opened page. | - | Customer track record | reads sensors | ? LinkedIn 51-200 range (https://www.linkedin.com/company/cybus) | cybus.io linkedin.com |
| GoodLinker (谷林運算) | LXL3 | LESI edge gateways, tower-light sensors and screen-data extractors that retrofit legacy machines, feeding a cloud command-centre dashboard (and since 2024 the GenAIoT analytics platform on AWS) that shows utilisation, downtime and alarms for SME factories. | any manufacturingsourceSME factories in food, metal processing, rubber and plastic moulding, foundries, electroplating, PCB and SMT. goodlinker.io | edge / IIoT · analyticswhyThe core is retrofit data capture (LESI EDGE gateways, tower-light sensors, screen data extractors) feeding LESI CLOUD dashboards for equipment status, output, alarms and trends; AI appears only as an 'AI Data Layer' and monthly AI reports, and no job listings were reachable. sources goodlinker.io goodlinker.io | detailSells AIoT data-collection hardware plus a subscription cloud war-room on AWS that customers can also pay for through telecom monthly fees; the NT$15m seed in November 2022 valued the company above NT$100m, a small business by the evidence found. sources goodlinker.io goodlinker.io | Foxconn (founder from)detailFoxconn · founder from Founder Ethan Feng (馮輝譯) was a senior engineer at Hon Hai Precision 2014 to 2018 and worked on Nokia and iPhone smart-factory projects. might.com.tw | Customer track record | reads sensors | - | goodlinker.io goodlinker.io digitimes.com.tw |
| HighByte | LX | HighByte Intelligence Hub, edge-native no-code industrial DataOps software that models, contextualises and delivers OT data to IT and AI systems. | any manufacturingsourceHomepage targets manufacturers broadly, naming customers such as Toyota, ConocoPhillips, National Grid and meal-kit maker Gousto. highbyte.com | data eng · ontology · edge / IIoTwhyIntelligence Hub models and contextualises OT data from PLCs, historians and MQTT/OPC sources into reusable data models and pipelines (pricing page lists unlimited data models and pipelines, UNS client, MQTT broker); the only open technical role is one software developer. hiring 1 open tech role: software 1 highbyte.com | USD 10k to 100kdetaildeal size USD 18.5k · per year, one factory. Public price list: Professional USD 18,500/year for one factory, Factory Starter Pack USD 50,000/year for 3 factories, Data Center Starter Pack USD 65,000/year, Enterprise custom; AWS Marketplace lists USD 18,500 per 12-month contract per site. highbyte.com funding USD 16.5m, as of 2024-04-23 highbyte.com Pure software at a published USD 18,500 per site per year up to custom enterprise deals; the Series A release says ARR tripled for the third year running in 2023 with 60+ industrial customers in 18 countries, on USD 16.5m total equity. | Siemens (partner)detailSiemens · partner Siemens announced a partnership at Realize LIVE Americas 2026; HighByte Intelligence Hub runs natively on Siemens Industrial Edge and is sold on the Industrial Edge Marketplace. chemanager-online.com | Customer track record | reads sensors | ? LinkedIn 11-50 range (https://www.linkedin.com/company/highbyte) | highbyte.com linkedin.com |
| i-flow | LX | Industrial unified namespace (hub, broker, edge) with 200+ connectors from S7/OPC UA to SAP and MES, plus an AI agent for one-click data modelling. | any manufacturingsourceHomepage says it is proven across industries from automotive and mining to pharma and data centres, and lists process industry, chemicals, food, pharma, machinery and construction. i-flow.io | edge / IIoT · data eng · streamingwhyThe product is an industrial unified namespace (hub, broker, edge) with 200+ connectors from S7/OPC UA to SAP and MES, and the site cites 750 million data operations a day; the only open engineering role is one senior full-stack developer. hiring 1 open tech role: software 1 i-flow.io | Bosch Rexroth (founder from) · Mercedes-Benz (founder from)detailBosch Rexroth · founder from Co-founder and CEO Daniel Goldeband lists Bosch Rexroth and Bosch in his work history. Mercedes-Benz · founder from Co-founder and CEO Daniel Goldeband lists Mercedes-Benz in his work history. | Customer track record | reads sensors | ? | i-flow.io i-flow.io jobs.techstars.com | |
| Litmus | LX | Industrial edge data platform (Litmus Edge, Edge Manager, Unify UNS, Data Catalog, MCP Server) that connects plant assets, standardises OT data and deploys analytics and AI models at the edge. | any manufacturingsourceHomepage says it is trusted by manufacturers scaling industrial data and AI, naming customers across sectors (Philips, JLR, SLB, Parker, Nature Fresh Farms) without one focus. litmus.io | edge / IIoT · data eng · streamingwhyOf 15 open engineering roles on its Ashby board, 8 are Golang/frontend/SRE software roles, 2 are Yocto embedded Linux roles and 4 are QA, with no ML or data-science roles, which fits a product that connects plant devices and processes OT data at the edge rather than one built on in-house modelling. hiring 15 open tech roles: data eng 1, software 12, embedded / edge 2 jobs.ashbyhq.com | detailfunding USD 42.6m, as of 2022-09-13 litmus.io Litmus Edge Manager is sold on AWS Marketplace at USD 7.50 per tag per year with a 1,000-tag minimum per instance (so at least USD 7,500 per instance per year); total equity was USD 42.6m after the 2022 Series B led by Belden, and a further undisclosed round from Insight Partners and Munich Re Ventures followed in 2025. | Panasonic (partner)detailPanasonic · partner Litmus Automation Japan and Panasonic Solution Technologies announced a strategic alliance for edge-to-cloud industrial IoT on 2025-09-18. litmus.io | Customer track record | reads sensors | 100 LinkedIn 51-200 range; litmus.io page metadata states '100 employees' | litmus.io linkedin.com |
| United Manufacturing Hub | LXL3 | Open-source, ISA-95 unified-namespace data platform that connects PLCs, sensors and ERP into a real-time hub for KPIs, condition monitoring and AI agents. | any manufacturingsourceBy-industry menu lists automotive, food and beverage, consumer and luxury goods, basic materials and utilities, and industrial products. umh.app | data eng · edge / IIoT · streaming | detailOpen-core model: UMH Core is free and open source, while the Enterprise Platform is one annual platform fee (first site included) plus a per-site annual fee scaled by site size (Satellite, Site, Campus), with prices quote-only; it raised a EUR 5m seed led by KOMPAS VC in January 2026. | - | Customer track record | reads sensors | 40 website says 11-50, LinkedIn shows 51-200 range; 40 is a midpoint guess between the two | umh.app tech.eu linkedin.com |
| Cirrus Link Solutions | LX | MQTT and Sparkplug infrastructure for plants: MQTT modules for Ignition (Engine, Distributor, Recorder, Injector), Chariot MQTT server and cloud IoT bridges. | any manufacturingsourceIndustries menu lists oil and gas, manufacturing, water, food and beverage, logistics, power and energy, data centers, pharma and automotive. cirrus-link.com | edge / IIoT · streamingwhyThe product line is MQTT/Sparkplug infrastructure (Ignition MQTT Engine, Transmission, Distributor, Recorder, cloud Injector modules and the Chariot MQTT broker), i.e. device connectivity and real-time message transport; no careers page or job listings were found. | under USD 10kdetaildeal size USD 1.6k · one-off, cheapest module per Ignition gateway. Inductive Automation list prices for Cirrus Link modules: MQTT Engine USD 2,150, MQTT Transmission USD 1,550, MQTT Distributor USD 3,350, MQTT Recorder USD 10,950, AWS/Azure/GCP Injector USD 4,175 each (list marked as an estimate, volume discounts on request). inductiveautomation.com Cirrus Link sells low-ticket software modules through the Inductive Automation Ignition price list (USD 1,550 to 10,950 per module per gateway; the Enterprise Integration suite bundling MQTT Engine and Transmission with Kafka and MongoDB connectors is USD 4,100), so revenue rides on Ignition's install base; no funding or revenue figures were found. | Koch (founder from) · Amoco (founder from)detailKoch · founder from Founder Arlen Nipper started his career as a SCADA engineer for the pipeline and transportation divisions of Koch Industries (Koch Oil). cirrus-link.com Amoco · founder from Founder Arlen Nipper began with embedded computers as a co-op student at Amoco pipeline in 1978; the about page lists Amoco among his leadership roles. cirrus-link.com | Early | reads sensors | ? LinkedIn 2-10 range (https://www.linkedin.com/company/cirrus-link-solutions) | cirrus-link.com linkedin.com |
| craftworks | LXL2L0 | navio MLOps platform to deploy and monitor ML models in plants, plus navio VISION anomaly inspection and predictive quality and maintenance projects. | any manufacturingsourceIndustries menu lists automotive, manufacturing, energy and utilities, mechanical engineering, material industries and pharmaceutical. craftworks.ai | ML · CVwhynavio is an MLOps platform to deploy and monitor ML models and navio VISION does real-time visual anomaly detection, sold alongside predictive quality and maintenance projects; the careers page has only a general 'Talent - Focus AI' pool entry, no specific opening. hiring 0 open tech roles craftworks.ai | detailsources craftworks.ai | - | Early | reads sensors | 25 LinkedIn 11-50 range | craftworks.ai linkedin.com |
| Factry | LXL3 | Factry Historian (plant time-series collection and analytics) and FactryOS production execution software for process plants. | process industriessourceTargets process plants: food and beverage, specialty chemicals, energy, textiles and heavy industry (Agristo, Lesaffre, AGC Glass Europe, Puratos). factry.io | data eng · analyticswhyProducts are a plant time-series historian (collectors, unlimited tags) and FactryOS production software; the jobs page lists 1 software engineer and 1 account executive, so evidence is mainly product and pricing pages. hiring 1 open tech role: software 1 factry.io | USD 10k to 100kdetaildeal size EUR 15k · per year, FactryOS from, one site. Pricing page: FactryOS 'Starting from EUR 15,000 / year', most single-site setups EUR 15,000 to 20,000 per year; Historian tiers have unlimited users, tags and collectors, prices not listed. factry.io Belgian filings (FY Jul-Jun) show no turnover but gross margin rising from EUR 0.16m (FY2020) to EUR 1.15m (FY2025), with a small net loss of EUR 48k in FY2025 and 13.8 FTE; no outside funding found; list price EUR 15-20k per site per year for FactryOS. sources factry.io busibee.be | - | Early | reads sensors | 59 LinkedIn 11-50 range with 59 members listed (https://www.linkedin.com/company/factry.io/) | factry.io factry.io linkedin.com |
| Flow Software | LXL3 | Industrial information and DataOps platform: Flow Infohub (KPI calculations and governance), Timebase Historian (free time-series historian) and Timebase Atlas (knowledge modelling). | any manufacturingsourceHomepage pitches to manufacturers in general; named customers span beverages (AB InBev, Coca-Cola), aerospace (RTX) and waste management. flow-software.com | data eng · analyticswhyInfohub models, calculates and governs KPIs over existing plant data and Timebase is a time-series historian (pricing counts 'KPI Measures/Events'); no careers page or ATS was found, so this rests on product and pricing pages. | USD 10k to 100kdetaildeal size USD 11.4k · perpetual licence, Starter, one site. Pricing page, per site: Starter USD 11,390 perpetual; Unlimited USD 62,230 perpetual or USD 16,595 annual; Distributed USD 128,450 perpetual or USD 34,250 annual; Enterprise custom; AI Gateway USD 7,500 perpetual or 2,000 annual per instance. flow-software.com Public per-site list prices of roughly USD 11k to 128k perpetual (or 17k to 34k a year) for Infohub, with the Timebase historian given away free; no funding or revenue figures found (a 2026 Exa Capital acquisition of a 'Flow Software' concerns a different ANZ integration firm). | AB InBev (founder from) | Early | reads sensors | ? LinkedIn 11-50 range (https://www.linkedin.com/company/flow-software-pty-ltd) | flow-software.com linkedin.com |
| HiveMQ | LX | MQTT broker and industrial data platform (HiveMQ Broker, HiveMQ Edge, extensions) used as the backbone of unified-namespace architectures in factories. | any manufacturingsourceHomepage case studies cover automotive (Ford, Mercedes-Benz), pharma (Eli Lilly), consumer goods (Unilever) and oil and gas, pitched at manufacturers generally. hivemq.com | streaming · edge / IIoT · data engwhyCore product is an MQTT broker plus HiveMQ Edge and Pulse for real-time device messaging in unified-namespace setups; the Ashby board shows only 2 open roles (a data-streaming product manager and an account executive), so evidence is product pages, not hiring. hiring 0 open tech roles jobs.ashbyhq.com | under USD 10kdetaildeal size USD 3.6k · per year, self-managed Launch tier, one broker. Pricing page: self-managed Launch $299/month, Run $499/month for one broker deployment; HiveMQ Cloud Starter from $0.34/hour plus $0.80 per million messages; Enterprise priced by capacity on quote. hivemq.com funding EUR 49.3m, as of 2022-05 tech.eu Self-funded with 130+ production customers before its EUR 9.3m seed (2021), then a EUR 40m Series A led by Molten in 2022; public list prices are small (USD 299 to 499 per month per broker) with enterprise deals priced by capacity. | - | Early | reads sensors | ? LinkedIn 51-200 range (https://www.linkedin.com/company/hivemq-gmbh) | hivemq.com linkedin.com |
| Idein (Idein株式会社) | LXL0L2 | Actcast edge-AI platform that deploys, monitors and updates camera and sensor AI on fleets of Raspberry Pi-class devices on site, with factory apps for inspection, stock counting, worker behaviour analysis, anomaly detection and meter reading. | any manufacturingsourceBusiness pages list manufacturing, retail, automotive and city and infrastructure; manufacturing is one segment among several. idein.jp | edge / IIoT · CV · MLwhyActcast is a platform to deploy and manage AI apps on fleets of edge devices (over 15,000 registered by April 2022 per the company deck), and 5 of 7 open technical roles on HERP are machine learning roles (including radar and autonomous-driving ML), plus an IoT platform engineer and a forward deployed engineer. hiring 7 open tech roles: ML 5, software 1, embedded / edge 1 herp.careers | -detailfunding JPY 4.3bn, as of 2025-10 speakerdeck.com Company deck states cumulative fundraising of about JPY 4.3bn, with Canon Marketing Japan investing in July 2025; no revenue or Actcast prices were readable (the pricing page is a client-side app that rendered no prices). sources speakerdeck.com actcast.io | Aisin (invested) · Canon (invested)detailAisin · invested Capital and business alliance with Aisin Seiki in March 2019; Aisin Seiki also listed among investors in the JPY 2bn round of October 2020. idein.jp Canon · invested Canon Marketing Japan MIRAI Fund (Canon MJ group CVC) invested in July 2025, amount undisclosed. corporate.jp.canon | Early | reads sensors | 55 company profile deck on speakerdeck: 55名 (2025年10月現在), and HERP profile 55名, both opened 2026-11-03 | idein.jp idein.jp speakerdeck.com |
| IOTech Systems | LXL0 | Edge data platform (Edge Central, commercial EdgeX Foundry) that connects industrial devices and makes their data ready for edge AI. | any manufacturingsourceHomepage names building automation, renewable energy and manufacturing as its sectors. iotechsys.com | edge / IIoT · data engwhyProducts (Edge Central, Edge Connect, Edge Manager) are device connectivity and an edge data platform built on EdgeX Foundry; the BambooHR board lists 0 open roles. hiring 0 open tech roles iotechsys.bamboohr.com | -detailfunding USD 10m, as of 2019-11 netimesmagazine.co.uk Disclosed funding is a USD 2.5m seed (2018) and a USD 7.5m Series A led by Dell Technologies Capital (Nov 2019). | Dell (invested)detailDell · invested Dell Technologies Capital led the USD 7.5m Series A in November 2019. pulse2.com | Early | reads sensors | 125 LinkedIn 51-200 range (midpoint used) | iotechsys.com linkedin.com |
| N3uron | LX | Modular industrial DataOps and edge platform (protocol connectors, MQTT/Sparkplug, historian and web visualisation) linking plant systems, SCADA and cloud. | any manufacturingsourceRecord lists energy (solar, wind, hydro), manufacturing, utilities and water, with customers Repsol, Veolia, Eli Lilly and Endesa. n3uron.com | edge / IIoT · data engwhyDocs organise the product as data acquisition (protocol connector) modules, data delivery modules (including SparkPipe, an MQTT Sparkplug bridge to cloud) and edge computing and visualisation modules managed by a Fleet Manager. sources docs.n3uron.com | - | - | Early | reads sensors | ? LinkedIn 11-50 range (https://www.linkedin.com/company/n3uron) | n3uron.com linkedin.com |
| NeuCloud (寄云科技) | LXL0L5 | NeuSeer industrial data intelligence platform: IIoT connectivity, time-series store and analytics, edge computing, low-code apps and the Neu.ai industrial agent built on an enterprise ontology. | any manufacturingsourceIndustrial data platform with energy, oil and gas, semiconductors, chemicals and rail named as sectors. neucloud.cn | data eng · ML · edge / IIoTwhyNeuSeer suite is an IIoT platform with edge-cloud collaboration, an industrial data management layer and time-series and ML modelling modules (forecasting, root cause, condition monitoring), and the 3 open technical roles are a big-data engineer, a data algorithm engineer and a backend engineer (of 6 openings, 3 in sales and project management). hiring 3 open tech roles: ML 1, data eng 1, software 1 neucloud.cn sources neucloud.cn neucloud.cn | - | CRRC (invested) · CETC (invested)detailCRRC · invested CRRC Capital (中车资本) invested in the 2022 Series C+ round of over RMB 100m. startup.aliyun.com CETC · invested CETC Investment (电科投资) invested in the January 2021 Series C (about RMB 100m) and the July 2022 Series C+. pitchhub.36kr.com | Early | reads sensors | - | neucloud.cn |
| Prosys OPC | LX | OPC UA software vendor: OPC UA Forge gateway, SDKs for Java, C/C++, .NET and Delphi, Simulation Server and Monitor. | - | edge / IIoT · data engwhyProducts are OPC UA SDKs (Java, C/C++, .NET), simulation and monitoring tools, and OPC UA Forge, an IT/OT integration gateway with protocol connectors, MQTT/REST publishing and history; no careers page was found (site map and about page show none). | detailrevenue EUR 1.4m, 2025 finder.fi Finnish filings show flat revenue around EUR 1.4-1.5m (2023 to 2025) with 17 staff and an operating loss of EUR 216k in 2025; SDKs are sold per developer with no runtime fees and Forge per installation or per site, prices on request. | - | Early | reads sensors | ? LinkedIn 11-50 range (https://www.linkedin.com/company/prosys-opc-ltd) | prosysopc.com linkedin.com |
| Quartic.ai | LXL1L3 | Sells an industrial DataOps and decision-intelligence platform (iLuminator, eXponence) with ML applications for process optimisation, batch multivariate analysis, reliability and PAT in process manufacturing. | process industriessourceTargets process manufacturing in life sciences, chemicals, CPG and food and beverage. quartic.ai | ML · data eng · analyticswhyProcess Optimizer builds 'sample-efficient' surrogate ML models from setpoints and outcomes for constrained multi-KPI setpoint recommendations, on top of iLuminator streaming DataOps (MQTT, Kafka, Flink) and Batch MVDA; the careers page says 'not hiring at the moment'. hiring 0 open tech roles quartic.ai | -detailAbout page names ZS, Celesta Capital and Michael Marks as investors but no amounts, and no pricing or revenue is published; customers named include Novo Nordisk and Sanofi. sources quartic.ai | Siemens (founder from) · Emerson (founder from)detailSiemens · founder from Founder and CEO Rajiv Anand started his career at Siemens. celesta.vc Emerson · founder from Founder and CEO Rajiv Anand held engineering, management and leadership positions with Emerson in process automation. celesta.vc | Early | reads sensors | 70 LinkedIn 51-200 band, 'View all 70 employees' | quartic.ai quartic.ai linkedin.com |
| Rhize (formerly Libre) | LX | Rhize Manufacturing Data Hub, a real-time event-driven ISA-95 data platform that gives plants a standardised, application-ready data model (headless, GraphQL). | any manufacturingsourceLists life sciences, food and beverage/CPG, serialized discrete and steel (Moderna, BlueScope, Cartier). rhize.com | ontology · data eng · streaming | -detailNo pricing, funding or revenue published; the company page says Rhize grew out of the consultancy Spruik Technologies and its earlier headless MES Libre. sources rhize.com | - | Early | reads MES / ERP | ? LinkedIn 11-50 range (https://www.linkedin.com/company/rhizemdh/) | rhize.com linkedin.com |
| Ulala Lab (울랄라랩) | LXL3L0 | WimFactory, a low-cost IoT smart-factory platform for SME plants: WICON gateways and add-on sensors collect equipment status data, a cloud dashboard shows real-time machine and line status, and ML models raise malfunction and emergency alerts. | any manufacturingsourceulalalab.com | edge / IIoT · analyticswhyWimFactory is described as a real-time production data collection and analysis IoT platform for smart factories, with an energy-data facility management add-on; the company website was blocked by the local network filter and THE VC hides job and revenue data, so no hiring data. | -detailOnly disclosed round found is KRW 2bn in April 2018 (Magna Investment and one other VC); THE VC lists 4 rounds, a 2020 Series A and 6 public-sector (B2G) contracts but hides amounts behind a paid plan. sources venturesquare.net thevc.kr | - | Early | reads sensors | 18 THE VC profile (thevc.kr/ulalalab): 18 employees as of 2026-09 | thevc.kr venturesquare.net venturesquare.net |
| ZEDEDA | LX | Edge orchestration platform to deploy, update and secure AI models and applications on distributed edge hardware, including factory floors. | any manufacturing | edge / IIoTwhyProduct is an edge orchestration platform that deploys, updates and secures apps and AI models on fleets of edge nodes; the Rippling board shows 2 roles, 1 a senior core-platform software engineer. hiring 1 open tech role: software 1 ats.rippling.com | USD 10k to 100kdetaildeal size USD 30k · per 12-month Starter Kit, up to 30 nodes. AWS Marketplace listing: ZEDEDA Starter Kit USD 30,000 per 12-month contract for up to 30 nodes, standard support included. aws.amazon.com funding USD 127m, as of 2024-02 futuriom.com USD 72m Series C (Feb 2024, led by Smith Point) took total funding to USD 127m; the company reported ARR up 250% and nodes under management up over 300% in the prior year, with list pricing around USD 1,000 per node per year on the AWS starter kit. sources aws.amazon.com futuriom.com | Rockwell Automation (invested) · Emerson (invested) · Samsung (invested) · Porsche (invested) · Chevron (invested) · Juniper (invested) · Saudi Aramco (invested)detailRockwell Automation · invested Existing investor at the USD 26m Series B (July 2022); returning investor in the USD 72m Series C (2024). zededa.com Emerson · invested Emerson Ventures named among returning investors in the USD 72m Series C (2024). futurumgroup.com Samsung · invested Samsung Next joined as a new investor in the USD 26m Series B (July 2022). zededa.com Porsche · invested Porsche Ventures joined in the Series B (July 2022) and returned in the Series C (2024). zededa.com Chevron · invested Chevron Technology Ventures joined in the Series B (July 2022) and returned in the Series C (2024). zededa.com Juniper · invested Existing investor at the Series B (July 2022); returning in the Series C (2024). zededa.com | Early | reads sensors | 125 LinkedIn 51-200 range (midpoint used) | zededa.com zededa.com linkedin.com |
| Zerynth | LXL3 | Plug-and-play industrial IoT and AI platform that connects any machine in under three hours and feeds OEE, energy and predictive analytics. | any manufacturingsourceRecord describes customers as manufacturing companies and machine builders with no branch named. linkedin.com | edge / IIoT · analyticswhyDocs list an IoT Core for device management and data integration, industrial apps for machine and energy monitoring, a no-code configurator and an 'Industrial Copilot', and the company GitHub holds firmware board-support and sensor libraries. sources docs.zerynth.com github.com | - | - | Early | reads sensors | 45 LinkedIn 11-50 range, 45 members listed | linkedin.com |
Technical core: the one to three competencies the product stands on, ranked, from open roles, engineering posts and product pages; open "why" for the reasoning, the role counts and the sources. Money signals: deal-size band and what it is per, then delivery model and pricing unit; open "detail" for revenue, funding, buyer, named competitors and sources. Manufacturer ties: large manufacturers the company is tied to and how; open "detail" for each tie and its source. A tie is one of six kinds with a large manufacturer or industrial group (more than 10,000 employees or more than USD 5bn revenue) or its venture arm or subsidiary: it invested, a joint venture, the company is a spin-off, a founder worked there, it acquired the company, or a formal partnership with a joint announcement or a co-built product. Customers alone do not count. Left out: utilities and energy suppliers, trading houses, mining and oilfield services, research institutes, fabless chip and outsourced-assembly designers, multi-company funds, and partner claims that rest only on a logo, a marketplace or ecosystem listing, or the startup's own partner page. Industry focus: what the company says it serves, on its own site, customer list or case studies. One industry (for example only cement), 2 to 3 named related industries, process industries broadly, discrete manufacturing broadly, or any manufacturing. A single named customer does not make a company single-industry. Open "source" for the sentence and page it rests on. A dash means nothing was found in a public source. Amounts in other currencies are converted to US dollars only to place them in a band, at approximately 1 EUR = 1.15 USD, 1 GBP = 1.33 USD, 1 CHF = 1.25 USD, 1 AUD = 0.65 USD, 1 CAD = 0.72 USD, 1 CNY = 0.14 USD, 1 TWD = 0.032 USD, 1 JPY = 0.0066 USD, 1 KRW = 0.00071 USD, 1 INR = 0.0113 USD, 1 CZK = 0.047 USD, 1 NOK = 0.098 USD, 1 SEK = 0.105 USD; bands are rough.
method
Companies were found by web search, and each row rests on the sources linked in its last column. The eight layers run from the sensor on a machine out to the suppliers and customers around a factory, and each company sits on the layers of what it sells. A dash means nothing was found in a public source.