AI & MACHINE LEARNING PHYSICAL AI

Japan Pairs With NVIDIA to Build Huge 140MW Physical AI Site

TM
Techmediaglobal
| 5 min read
140MW
DATA CENTRE CAPACITY
¥380bn
GOVERNMENT-BACKED (US$2.3BN)
27,500
RUBIN GPUS
30%
2040 MARKET SHARE GOAL

Japan has partnered with NVIDIA and a consortium of industrial leaders to build what could become the country's first national infrastructure dedicated to physical AI. Backed by a ¥380bn (US$2.3bn) government grant, the project centers on an AI factory delivering 140MW of data centre capacity, built around NVIDIA's DSX platform and 27,500 Rubin GPUs.

Building the AI Factory

NVIDIA will partner with Japanese AI consortium Noetra to build the NVIDIA Vera Rubin AI factory, comprising 13,750 Vera CPUs and 27,500 Rubin GPUs. Noetra's private-sector backers include Honda and SoftBank.

The project carries backing from Japan's Ministry of Economy, Trade and Industry (METI) as part of its FRONTia Project, which targets AI agents, digital twins, and industrial robotics development across manufacturing, logistics, healthcare, and telecommunications.

Noetra will architect the facility using NVIDIA Vera Rubin NVL72 racks on the DSX platform, connected and scaled with NVIDIA Spectrum-X Ethernet networking, laying the groundwork for open multimodal foundation models to power AI agents, digital twins, and robotics.

"Japan invented modern manufacturing. Now, it is building the AI factories that will power the next industrial revolution."

— Jensen Huang, Founder and CEO, NVIDIA

A National Robotics Push

Under Japan's AI Robotics Strategy, released in March 2025, the country aims to secure more than 30% of the global AI robotics market by 2040, an opportunity estimated at US$133bn. Japan's industry ministry also aims to introduce about 10 million AI-equipped robots across 18 sectors by 2040 in response to the country's growing labour shortage.

Japan currently produces 45% of the global supply of industrial robots, making it the world's largest robot manufacturer. However, installation figures paint a different competitive picture: China installed roughly 295,000 industrial robots in 2024, compared with 44,500 in Japan, the world's second-largest market.

Voices Behind the Project

Hironobu Tamba, CEO of Noetra, notes that bringing physical AI into the real world requires enormous computing, data, and foundational technologies that no single company can solve alone.

Japan's Minister of Economy, Trade and Industry, Ryosei Akazawa, describes the FRONTia Project as the core of the country's physical AI ecosystem, built by combining Japan's onsite expertise and manufacturing infrastructure with global partners such as NVIDIA to develop highly reliable multimodal foundation models.

Changing Market Dynamics

China's domestic manufacturing capacity is expanding fast. According to the IFR's World Robotics 2025 Report, Chinese manufacturers outsold foreign manufacturers in China's own market for the first time in 2024, with domestic share climbing to 57%, up from 47% in 2023.

Japan remains a major robotics exporter to China, with more than a third of Japanese robotics and automation exports destined for the Chinese market. The new AI infrastructure could help Japan defend its manufacturing position as competitive dynamics continue to shift.

Key Takeaways

  • Japan and NVIDIA are building a 140MW physical AI factory with the Noetra consortium, backed by a ¥380bn (US$2.3bn) government grant.
  • The site will run on 13,750 Vera CPUs and 27,500 Rubin GPUs using NVIDIA's DSX platform and Spectrum-X networking.
  • The project is part of Japan's FRONTia Project under METI, targeting AI agents, digital twins, and industrial robotics.
  • Japan aims to capture 30%+ of the global AI robotics market by 2040, an estimated $133bn opportunity, and deploy 10 million AI-equipped robots.
  • Japan produces 45% of the world's industrial robots but trails China in annual installations (44,500 vs. roughly 295,000 in 2024).
  • China's domestic robot manufacturers now hold 57% share of their home market, up from 47% in 2023, intensifying competitive pressure.
Tags: NVIDIA Physical AI Japan Robotics Data Centres Rubin GPUs FRONTia Project AI Infrastructure