30 Years After Sega's $5M Lifeline, 22 Japanese Firms Join Nvidia's Cosmos: The National-Scale Battle for Physical AI
Sources: TechCrunch / 21st Century Business Herald / Tencent News URL: https://techcrunch.com/2026/07/19/what-to-watch-for-after-jensen-huangs-japan-visit Date: 2026-07-19
What Happened
During a two-day Tokyo trip on July 15-16, 2026, Jensen Huang closed three major deals:
1. 22 Japanese companies joined the Cosmos alliance — Fanuc, Yaskawa, Kawasaki Heavy Industries, Fujitsu, Hitachi, NEC, Sony, SoftBank, Kubota, Omron, Komatsu, Mitsubishi Heavy Industries, Mitsui & Co., Mitsubishi Corporation, Honda Research Institute, Tokyo Electron, AIRoA, and more. 2. Cosmos 3 Edge launched — a lightweight world model that runs on Jetson Thor chips, enabling inference on the robot itself without pushing all data to the cloud. 3. Noetra, Japan's national AI infrastructure — led by SoftBank, Sony, NEC, and Honda with 44 Japanese companies participating. Tokyo committed ¥1 trillion (~$6.2 billion) over 5 years, building a "Vera Rubin AI factory" with 13,750 Vera CPUs + 27,500 Rubin GPUs + 140 MW, operational June 2028.
Deep Dive
Huang's Three Lines in Tokyo
On July 16, at METI's "Physical AI launch ceremony" — with Trade Minister Ryosei Akazawa on stage and PM Sanae Takaichi appearing via video — Huang's core messages:
> "Japan cannot outsource its national intelligence. Japan must own, improve, secure, and deploy AI that is Japan's own."
> "The next industrial revolution will be made in Japan."
> "Japan's exceptional mechatronics is ready to fuse with physical AI."
Three concepts at once: sovereign AI (countries building their own compute), physical AI (embodied intelligence in robots, cars, factories), and made-in-Japan as the execution base.
Noetra's Three-Phase Roadmap
- FY2026 (from this April): reasoning models with strong Japanese-language capability, successor to SoftBank's "Sorashi" LLM.
- FY2028: full multimodality (text / image / video / audio), synchronized with the Vera Rubin AI factory.
- FY2030: "Real-world Native AI" — foundation models designed for robot embodiment, opened to developers in phases.
- Fujitsu + Fanuc: factory physical AI, real-time production planning optimization.
- Fujitsu + Yaskawa: retail and logistics automation integrating sales and inventory data.
- Fujitsu + Kawasaki Heavy Industries: healthcare — surgical assistance, nursing robots, hospital transport using Nvidia's Holoscan IGX, Isaac for Healthcare, Isaac GR00T, and Cosmos.
- 10 million AI robots deployed across 18 sectors
- $65 billion in public + private investment (~¥9.6 trillion)
- 30% global AI robotics market share by 2040
- Market size ~¥20 trillion ($133 billion)
- Hard dependence on Nvidia hardware — Noetra trains on the Vera Rubin AI factory and infers on Jetson Thor; vendor lock-in is extreme and a single point of failure if export controls escalate.
- Real feasibility of the "third option" — Japan's own LLMs (Sorashi, Sarashina) lag the first tier noticeably on AA-Index; the 30% market-share path is long.
- Physical AI PMF unproven — public data on real deployment costs vs. human labor in factories, logistics, and healthcare remains thin.
- Human-robot relations in the cobot era — 10 million robots across 18 sectors will reshape Japan's labor structure; a topic Huang's visit did not touch.
- TechCrunch: https://techcrunch.com/2026/07/19/what-to-watch-for-after-jensen-huangs-japan-visit
- 21st Century Business Herald: https://new.qq.com/rain/a/20260716A0ALVK00
- Tencent News (deep commentary): https://news.qq.com/rain/a/20260717A0B9L200
- Nvidia official: https://nvidianews.nvidia.com/news/japans-robotics-and-manufacturing-leaders-build-on-nvidia-cosmos-to-advance-physical-ai-frontier
Key point: Noetra aims to train foundation models that "robots can run," not chatbots. Hirotsune Tanba of SoftBank's LLM effort serves as Noetra president — calling it "a true third option that Japan and other countries can choose."
Cosmos 3 Edge + Jetson Thor: The AI Brain Inside the Robot
Jetson Thor is Nvidia's robot-specific chip announced in 2024, in mass production since 2025. Cosmos 3 Edge is a world-model variant tailored for Thor — meaning robots can handle most inference on-board, removing latency and bandwidth bottlenecks of cloud-based approaches.
The Fujitsu × Fanuc × Yaskawa × Kawasaki chain:
"Getting half of Japanese industry to sign the same robot underlying protocol" may matter more than the Cosmos models themselves.
Toyota: Drive Platform + Factory Simulation + Conservative ADAS
Toyota's cooperation with Nvidia spans three layers:
| Layer | Content | |---|---| | Vehicles | Next-gen models to use Nvidia Drive platform (pledged at CES 2025) | | Manufacturing | Production lines designed with Nvidia simulation | | Software | Drive OS for vehicle software + road traffic interpretation |
But Toyota's ADAS strategy is "advanced driver assistance, driver still required" — not the driverless approach of Waymo or Tesla. This signals, clearly to English-language observers, that Japan's automotive AI path is more conservative than China's and the US's.
The 2040 National Target: 10 Million AI Robots
Japan's AI robot strategy (published March 2026) targets:
The Takaichi cabinet lists AI and semiconductors as core growth pillars, with ¥370 trillion ($2.3 trillion) in public-private investment across 17 strategic fields through 2040. Noetra's Vera Rubin AI factory is the first substantive piece of this national strategy to land.
The 30-Year-Old Story
In 1995-96, Nvidia bet wrong on 3D graphics; the Sega-partnered NV2 chip failed and the company neared bankruptcy. Sega CEO Shoichiro Irimajiri provided $5 million in emergency funding, keeping Nvidia alive until the RIVA 128 turned things around in 1997.
On July 15, Huang returned to Akihabara — standing on the former site of a Sega game arcade — reuniting with current Sega CEO Haruki Satomi and Irimajiri. Thirty years ago, $5 million saved an American GPU company; today that company is plowing back into Japan's entire industrial robotics chain. This is not nostalgia marketing; it is a textbook case of supply-chain diplomacy.
Why It Matters
1. A national AI strategy landing via one American company — sovereign AI has been a slogan for years; Noetra is the first plan with concrete compute, companies, and a timetable. 2. A de facto standard for the robot's on-board AI brain — Cosmos 3 Edge + Jetson Thor + 22 Japanese backers suggest Nvidia's "Android moment" in physical AI may already have happened. 3. Physical AI is the next battlefield after LLMs — embodied intelligence collectively started in H2 2026 (MiniCPM-Robot, Riemann-1.0, Poko Robotics, Nvidia Cosmos). 4. Japan-made vs. Chinese embodied AI — Japan holds the industrial robot base (Fanuc, Yaskawa, 40 years deep); China holds software and data (world models, VLA, data governance). Nvidia is betting on both sides as a supplier. 5. $65B + ¥370T capital pools — physical AI's funding scale in H2 2026 far exceeds language models; any embodied/physical AI startup's addressable market just grew.
Risks and Open Questions
---
One-Line Summary
22 Japanese companies + Cosmos 3 Edge + ¥1 trillion Noetra + the 2028 Vera Rubin AI factory — this is Nvidia's key campaign pushing "physical AI" from concept to national infrastructure, and Japan's biggest bet in the sovereign AI era. The window for a Japan-made vs. Chinese embodied AI showdown is officially open.
Source links: