Jensen Huang, CEO of Nvidia, recently visited Tokyo on July 15 and 16, engaging with key players in Japan’s industrial and semiconductor sectors. This trip followed a presentation in Taiwan and a prior visit to South Korea. Huang secured significant agreements that encompass Japan’s tech landscape, including initiatives for a national AI factory, partnerships with top robotics companies, and collaborations with suppliers of chip materials essential for Nvidia’s next-generation AI technologies. Huang emphasized his vision that the future of AI lies in the automation of factories, robotics, and machinery, with a strong focus on Japan leading this transformation.
Three decades ago, a $5 million investment from Sega played a crucial role in Nvidia’s survival; now, both Nvidia and Japan’s industrial leaders are looking to collaborate again, aiming to usher in a new era of physical AI through these key initiatives:
Noetra — Japan’s initiative for sovereign AI. Japan seeks to develop its own AI for its factories and robots, avoiding reliance on American or Chinese developments. The government has united approximately 44 domestic companies, including SoftBank, Sony, NEC, and Honda, to create indigenous AI solutions aimed at robotics, vehicles, and industrial automation. Tokyo plans to invest up to 1 trillion yen (around $6.2 billion) over the next five years to promote homegrown “physical AI,” with foundational models designed to operate machinery. Nvidia is set to construct a large-scale AI facility known as the Vera Rubin AI factory, slated to become operational in 2028, equipped with a robust array of chips to meet the growing demand for data processing. Noetra will manage the construction of the facility, which will unfold in three phases: a reasoning model focused on Japanese-language skills starting in 2026; an omni-modal version supporting various media formats by 2028; and a “Real-world Native AI” tailored for robotics by 2030, with phased access for external developers.
The robotics coalition — Japan’s leading industries align with Cosmos. Recognizing the potential of Nvidia’s technology, many of Japan’s foremost robotics and manufacturing firms are coming together. Companies such as Fanuc, Yaskawa, Kawasaki Heavy, Fujitsu, Hitachi, NEC, and SoftBank have announced plans to utilize Nvidia’s Cosmos models, an initiative launched earlier this year to collaborate with global AI research centers. During his visit, Huang introduced Cosmos 3 Edge, a version designed to operate with Nvidia’s Jetson Thor chips located directly in machinery. Various companies, including Honda R&D and Omron, are already experimenting with shared control systems and developing new tools. Huang remarked, “The next frontier of AI is in the physical world, presenting Japan a once-in-a-generation chance to redefine manufacturing.”
Toyota — merging automotive technology with physical AI. Toyota has integrated Nvidia chips throughout its technology stack and has committed its future vehicle developments to Nvidia’s Drive platform, starting in January 2025. The new initiatives extend Nvidia’s influence into Toyota’s manufacturing processes, utilizing simulations for optimizing production lines and developing software for vehicle operations, including advanced driver assistance systems. Unlike competitors like Waymo and Tesla, Toyota is taking a more gradual approach, still requiring driver input.
Huang’s visit effectively positioned physical AI at the forefront of Japan’s industrial strategy, with Tokyo prepared to invest significantly to match its ambitions. Confronted with a shrinking workforce, Japan aims to implement 10 million AI-driven robots across 18 industries by 2040, supported by a combined public and private investment of $65 billion into physical AI.
Looking ahead, Japan’s AI Robotics Strategy, revealed in March, aspires to capture over 30% of the global AI robotics market by 2040, a sector valued by Tokyo at approximately ¥20 trillion (around $133 billion). The Ministry of Economy, Trade, and Industry (METI) funds domestic model development for the management of robotic systems, with Noetra’s Nvidia-supported factory set to develop complex models for training.
Beneath the industrial rationale lies a strategic aim for sovereignty. As the U.S. and China advance in extensive AI capabilities, Japan seeks to harness its own data and computing power, reducing dependence on external infrastructure. On July 16, Huang appeared alongside Japan’s trade minister at the government’s launch event for physical AI, with Prime Minister Sanae Takaichi participating virtually. The Takaichi administration has prioritized AI and semiconductor development as core elements of a growth plan aiming for ¥370 trillion ($2.3 trillion) in public and private investments by 2040. Noetra’s factory, touted by Nvidia as “the world’s first national AI infrastructure,” stands as a significant undertaking. Currently, Japan’s push for autonomy, particularly in semiconductor reliance, hinges on American technology.
In just two days, Huang engaged with almost every major figure in Japan’s technology sector, including top executives from Toyota, Fanuc, Yaskawa, Fujitsu, and Kawasaki, over shared meals and discussions in local dining establishments.
This visit mirrors the strategy employed during his recent trips, from the keynote in Taiwan to a significant GPU partnership in Seoul last fall. Now, with a focus on Japan, the emphasis remains on robotics, supply chain enhancement, and the underlying technology that supports these developments.
