Kyocera (6971.T), Kobe Steel (5406.T), and roughly 20 other Japanese manufacturers will form a corporate consortium to commercialize “physical AI” — artificial intelligence that operates robots on manufacturing floors — it was learned on the 29th. The group will train Japanese-made AI, conceived by a software developer in Kanazawa, on each company’s manufacturing data and embark on joint development of machining robots that autonomously perform cutting and processing tasks. The initiative launches in October with an anticipated four-year project period.
The deployment of physical AI, which the Japanese government has identified as one of Japan’s “winning strategies,” is now moving into full-scale, private-sector-driven action. At the core of the consortium is the “TTMC,” an AI-equipped automated machine tool developed by Alm (Kanazawa). Using this Japanese-made AI as its foundation, participating companies will pool their manufacturing-floor expertise and processing data, aiming to realize robots that can automatically handle a wide variety of cutting and machining tasks without relying on skilled craftsmanship.
The consortium also includes Nidec Machine Tool, a subsidiary of Nidec (formerly Nidec Corporation, 6594.T), and a Sojitz (2768.T)-affiliated company. Additionally, the Japanese arm of U.S.-based Microsoft is in discussions to join in a supporting role for Alm. Other Japanese manufacturers are expected to join as well, with the final structure projected to reach approximately 20 companies.
Government Strategy and Investment Scale
Physical AI is positioned as one of the core elements within the 17 strategic technology fields formulated by the Japanese government. The government has set a target of investing more than ¥370 trillion (approximately $2.3 trillion) in combined public and private funding by fiscal 2040, of which ¥10.5 trillion is earmarked for the physical AI field.
As the technological hegemony race with the United States and China intensifies, the key question is whether Japan can establish a distinctive position in manufacturing-sector AI. A defining feature of this consortium is its emphasis on Japanese-made technology. By avoiding dependence on foreign-made AI and consolidating on-site data from Japanese manufacturers, the group aims to develop AI optimized for Japan’s manufacturing industry.
Challenges to Commercialization
Commercializing physical AI requires not only improving software performance but also the ability to respond to unpredictable situations that arise in actual factories. In cutting and machining, the AI must assess subtle variations in raw materials and tool wear conditions in real time, and transplanting the tacit knowledge of skilled workers into AI represents a significant technical hurdle.
The consortium’s strength lies in its ability to share manufacturing data across multiple companies, enabling the AI to learn from a diverse range of machining conditions that would be difficult for any single company to accumulate on its own. On the other hand, adjustments will be needed going forward regarding how much data competing companies can disclose to one another and how intellectual property rights will be allocated.
Whether the four-year project period can produce practical results will serve as a litmus test for the feasibility of the government’s investment targets.