Nvidia's Japan robotics blitz raises the bar for India's AI push

The News

Jensen Huang has turned Japan into the centrepiece of Nvidia's push beyond chatbots and into the factory floor. Over a two-day visit to Tokyo on 15 and 16 July 2026, the Nvidia chief executive stitched together agreements with the country's biggest industrial names, from Toyota to SoftBank, all pointed at what he calls physical AI: intelligence that runs robots, machines and production lines rather than search boxes.

The anchor of the trip is Noetra, Japan's sovereign AI programme. The government has committed 1 trillion yen, roughly 6.2 billion dollars, over five years, and a coalition of 44 domestic firms will build homegrown foundation models in three phases: a Japanese-language reasoning model in fiscal 2026, an omni-modal system handling text, images, video and audio by 2028, and a "Real-world Native AI" tuned for robots by 2030.

The hardware behind it is enormous. A data centre due to launch in 2028 will run 13,750 Vera CPUs and 27,500 Rubin GPUs, drawing 140 megawatts of power. On the robotics side, Fanuc, Yaskawa, Kawasaki Heavy, Fujitsu, Hitachi, NEC, Sony, SoftBank and Kubota agreed to build on Nvidia's Cosmos models, and Nvidia unveiled Cosmos 3 Edge running on its Jetson Thor chips. Toyota extended its use of the Nvidia Drive platform into manufacturing, simulation and vehicle software.

Japan's stated ambition is staggering: 10 million AI-equipped robots across 18 sectors by 2040, 65 billion dollars of physical-AI investment, and more than a 30 percent share of a global AI robotics market it values at 20 trillion yen, or about 133 billion dollars.

Why It Matters

This is a national-champion strategy aimed squarely at a problem money alone cannot fix: Japan's shrinking, ageing workforce. By fusing its manufacturing heritage with Nvidia's silicon, Tokyo is betting that robots become the labour force of the 2030s.

For Nvidia, the visit fits a pattern. The company has spent the past year signing sovereign-AI pacts across the Gulf, Europe and Asia, converting geopolitical anxiety about chip access into multi-year GPU commitments. The last time a single vendor sat this close to a nation's industrial policy, it was the semiconductor foundry deals of the 2010s, and those relationships defined a decade of supply chains.

The strategic tell is the shift from language to labour. Markets have spent three years obsessing over chatbots. Huang is signalling that the next wave is machines that see, move and act, and he wants Nvidia inside every one.

Indian Angle

For New Delhi, the Noetra blueprint is both a mirror and a warning. India runs its own sovereign push through the IndiaAI Mission, backed by roughly 10,000 crore rupees, and has subsidised tens of thousands of GPUs for startups. Yet the Indian effort has leaned heavily on language: Sarvam, Krutrim and others are racing to build Indic foundation models, while physical AI and robotics remain thinly funded.

Japan's plan exposes that gap. India has real manufacturing ambitions under the production-linked incentive schemes and Make in India, but very little AI capital is flowing into factory-floor automation. If Indian electronics and auto plants want to compete with robot-dense Japanese lines by 2030, MeitY and industrial policymakers may need a physical-AI track, not just a chatbot one.

There is also a commercial opening. Nvidia already partners with Reliance and Tata on Indian AI infrastructure. Indian robotics startups and integrators that master Cosmos and Jetson Thor now could become the export arm serving factories across South and South-East Asia, rather than importing finished automation.

FAQ

What is Noetra?

Noetra is Japan's sovereign AI programme. Tokyo has pledged 1 trillion yen, about 6.2 billion dollars, over five years, and 44 domestic firms will build homegrown foundation models in three phases through 2030, moving from Japanese-language reasoning to an omni-modal system and finally to AI designed to control real-world robots.

How does this compare with India's AI push?

India's IndiaAI Mission is backed by roughly 10,000 crore rupees and has subsidised GPUs for startups, but it is weighted towards language models such as those from Sarvam and Krutrim. Japan's programme is far larger and centred on robotics and manufacturing rather than chatbots.

What is physical AI?

Physical AI refers to models that perceive and act in the real world: controlling robots, factory machines, vehicles and production lines. It is the focus of Nvidia's Cosmos models and Jetson Thor chips, and Huang argues it is the next major frontier after text-based generative AI.

Where can I read the original announcement?

TechCrunch reported on the deals and commitments from Jensen Huang's Tokyo visit. The full coverage, including the roster of Japanese partners, is linked in the attribution below.

This story was reported by TechCrunch. Read the full original coverage at TechCrunch.