The market in numbers
The global industrial AI market was $43.6 billion in 2024. IoT Analytics forecasts growth to $153.9 billion by 2030, about 23% a year.
This is still a small share of the industry’s money: spending on industrial AI equals 0.1% of manufacturers’ revenue. Growth starts from a low base.
In Russia, Kama Flow ranks industrial AI first among the segments open to suppliers: $1.1 billion by 2030. Next come AI agents and process automation ($940 million) and medtech ($840 million), and all segments together exceed $6.8 billion.
For a plant manager, the Kama Flow estimate means one thing: industrial AI is seen as the market with the highest demand in Russia, so many suppliers will make offers. Choose between them by the metric of a pilot, not by a presentation.
Deals
In June 2026, London-based PhysicsX, which builds AI models for the engineering calculation of parts, from jet engines to semiconductors, raised $300 million at a valuation of about $2.4 billion. Singapore’s Temasek led the round, and Nvidia took part.
In September 2026, Japan’s CADDi, an AI platform for drawings and production data, raised $114 million at a valuation of $1.2 billion. The company says more than half of the 100 largest Japanese manufacturers use it.
Both deals concern engineering calculations and drawings, the work of design and process engineers. For a plant this is a signal: money in industrial AI goes not only to robots, but also to documents and calculations.
A separate wave of money goes into robots. According to Crunchbase, as of 22 June 2026 robotics startups had raised $18.8 billion since the start of the year, more than in all of 2025 ($15 billion) and in the peak year 2021 ($14.1 billion).
Robot makers are also going public. In the third quarter of 2026, Unitree Robotics raised $905 million in its IPO at a valuation of $9 billion.
All rounds with amounts and sources are in the deal tracker. Investment in robots is covered on the page about investment in robotics.
Russia
Actual venture money in Russian manufacturing is still small. According to Venture Guide, industrial technology received about $19 million in 12 deals in the first 9 months of 2026.
Robotics received $13 million in 9 deals over the same period. The data is preliminary, as of 4 October 2026.
AI and machine learning as a whole raised $32 million in 33 deals in the first 9 months of 2026, first place by both volume and number of deals. This data is also preliminary.
You cannot compare these sums with the $1.1 billion estimate directly: one number covers deals over 9 months, the other covers demand by 2030. The practical conclusion for a plant: money for AI today comes from its own budget, not from outside investors.
The overall picture for the country is on the page about AI investment in Russia.
Where it pays back
An example with a measured effect: the Siemens plant in Nanjing introduced more than 50 AI applications and digital twins. Compared with 2022, order lead time fell by 78%, and time to market became 33% shorter.
By 2024, productivity grew by 14%, and customer failures fell by 46%. The effect came from dozens of applications, not from one project.
Siemens measures the effect against a 2022 baseline. Do the same: measure order lead time, defects and downtime before the start, or in a year you will have nothing to compare the result with.
Data often limits the effect. According to a Rockwell Automation survey (May 2026, 1,560 respondents from 17 countries), AI already supports 34% of production operations, but only 43% of the collected data is used effectively.
For a plant this means that payback starts not with the model, but with the data on machines, defects and orders that is already collected, and the form it is stored in.
Examples of AI tasks in production are on the page about AI use in manufacturing.
What a company should do
If you run production, start with data, not with the choice of a supplier. According to the Rockwell Automation survey, only 43% of collected data is used effectively; the rest sits unused.
1. Describe the data you already collect: downtime, defects, order lead times, raw material use.
2. Choose one operation with a measurable metric: order lead time, defect rate, downtime.
3. Calculate the effect in money before the pilot and write the metric into the contract with the vendor.
4. After the pilot, scale one operation at a time: the Siemens effect in Nanjing came from more than 50 applications.
This order is set out in the AI solutions map: the plant processes where AI returns money, the solutions, and the price and effect of each. How to calculate the return before you buy is described on the page about calculating the effect of AI.
Where AI money goes outside manufacturing is shown on the page about AI investment trends.