Nvidia unveils Japan AI factory and new Thor robot modules

Nvidia ↘ 0.03% on Tuesday unveiled two new robotics computing modules and announced a partnership with Japan to build what it called the world's first national AI infrastructure for physical AI, extending the chipmaker's reach from cloud data centers into factories, hospitals and autonomous systems.

New Thor Modules for Mainstream Robotics

The Jetson T3000 and T2000, announced July 14, expand Nvidia's Thor architecture into smaller, more power-efficient form factors aimed at mass-market robotics and edge AI. The T3000 pairs a Blackwell GPU delivering 865 FP4 teraflops with an eight-core Arm Neoverse CPU, 32GB of LPDDR5X memory and 25 GbE connectivity — roughly half the size and power of the existing T5000 module. The T2000, an entry-level option at 400 teraflops with 16GB of memory, targets autonomous mobile robots, industrial manipulators and visual AI agents. Both modules are scheduled for availability in the first quarter of 2027, though T3000 emulation mode will arrive later this month via JetPack 7.2.1.blogs.nvidia+2

With the additions, Nvidia now offers edge AI hardware spanning 70 TOPS to 2,000 teraflops, a range the company says covers virtually any edge workload.blogs.nvidia

FRONTia: A National AI Factory for Physical AI

In Tokyo on July 16, Nvidia announced it is working with Noetra Corp. to build an AI factory equipped with 13,750 Vera CPUs and 27,500 Rubin GPUs, delivering 140 megawatts of data center capacity on the company's DSX platform. The facility will underpin Japan's FRONTia Project — formally titled "Development of Multimodal Foundation Models with a View to AI Robotics and Physical AI" — launched by Japan's Ministry of Economy, Trade and Industry.stocktitan

"Japan invented modern manufacturing. Now, it is building the AI factories that will power the next industrial revolution," Jensen Huang said at the event. Japan's Minister of Economy, Trade and Industry Ryosei Akazawa said the project would leverage the country's manufacturing expertise to "build highly reliable multimodal foundation models and contribute to solving global social challenges".stocktitan

Pretrained model weights developed by Noetra will be shared broadly with domestic developers alongside Nvidia software including Nemotron, Cosmos, Isaac GR00T and NeMo. Japan's AI Robotics Strategy, released in March, targets more than 30% of the estimated $133 billion global AI robotics market by 2040.stocktitan

Bridging Cloud and Edge

The twin announcements reflect Nvidia's strategy of linking massive cloud-scale training infrastructure with compact edge modules that can deploy resulting models in physical environments. Huang's Tokyo visit — his first extended engagement in Japan after criticism over a perceived "Japan passing" earlier this year — underscored the company's commitment to the country's AI ambitions.kucoin+1