Japan’s industrial giants just launched a big bet on “Physical AI” — and they’re not trying to build the next ChatGPT.
SoftBank, NEC, Honda and Sony Group have formed a new company, roughly translated as Japan AI Foundation Model Development, with a single, focused mission: build a trillion-parameter AI model designed to control machines — robot arms, cars, and factory systems — rather than to chat. SoftBank and NEC will lead model development; Honda will be the primary autonomous-driving customer; Sony contributes robotics and gaming hardware expertise; and Tokyo’s respected AI shop Preferred Networks is also on board.
This isn’t a small startup play. The new entity plans to hire about 100 AI engineers and will be led by a SoftBank executive as president. Investors extend well beyond tech: Nippon Steel, Kobe Steel, MUFG Bank, Sumitomo Mitsui Banking, and Mizuho Bank are all listed as backers, signaling an industrial-scale effort that blends private capital, corporate demand, and national priorities.
Crucially, Japan is aiming to keep the work and data domestic. For years Japanese firms have relied on U.S. cloud infrastructure — sending valuable industrial data overseas and deepening dependence on foreign AI stacks. This project aims to train models on Japanese data, run in Japan, and feed Japanese manufacturing and mobility — a strategic pivot that contrasts with SoftBank’s recent global AI plays. (SoftBank notably led a $40 billion investment round into OpenAI in 2025.)
Public funding is queued up too. NEDO, Japan’s national R&D agency, has earmarked roughly ¥1 trillion (about $6.28 billion) in AI support over five years beginning fiscal 2026. The newly formed company is expected to apply for that program and is widely seen as a near-certain recipient. NEDO opened its proposal window in late March, so the timeline is already in motion, and local reports put practical deployment targets for Physical AI around 2030.
Physical AI is heating up worldwide. Tesla is building robots, OpenAI is backing robotics startups, and China has major state-backed plans in the space. Even crypto-linked players have started to show interest: stablecoin issuer Tether invested in humanoid-robotics startup Generative Bionics earlier this year, reflecting growing crossover between digital-asset capital and real-world robotics.
What this means for crypto audiences: the rise of machine-focused foundation models could create new demand for secure data marketplaces, on-device compute resources, and provenance systems — areas where blockchain and tokenized incentives could play a role. Institutional interest from banks, steelmakers and national R&D funds may also open new deal flow for tokenized infrastructure financing or digital-asset-backed investment vehicles. Keep an eye on how Japan’s push for domestic AI stacks could shift data sovereignty debates and create bidding opportunities for infrastructure and middleware that tie physical systems to digital ledgers.
Bottom line: Japan’s cluster of legacy industrial powerhouses and tech firms are assembling to build trillion-parameter models with a clear use case — not to be your virtual assistant, but to run the real world. The implications for manufacturing, mobility and the intersection of AI with crypto-enabled infrastructure are only starting to emerge.
Read more AI-generated news on: undefined/news
