NVIDIA CEO Jensen Huang announced plans to collaborate with Japanese research institutions and companies to establish an ecosystem for artificial intelligence technology in Japan. This initiative aims to meet the country's demand for AI powered by GPUs (Graphics Processing Units).
During a meeting with Japan's Minister of Economy, Trade, and Industry, Yasutoshi Nishimura, Huang expressed NVIDIA's commitment to setting up an AI research lab, investing in local startups, and educating the public on AI usage. "We will establish an AI factory network in Japan, enabling the country to process societal data and create intelligence for society and industry," Huang stated.
NVIDIA's GPUs, known for their parallel computing capabilities to process large amounts of data, have become the preferred tool for training large AI models. In earlier discussions with Japanese Prime Minister Fumio Kishida, Huang acknowledged the high demand for NVIDIA GPUs and pledged to prioritize Japan's needs for these processors. He also expressed his desire to collaborate with Japanese companies focused on building large data centers and developing language models.
Japan is leveraging its expertise in material science and chip manufacturing tools, offering billions of dollars in subsidies to increase domestic production of cutting-edge semiconductors and regain its technological leadership. However, so far, Japan has largely missed out on the most significant profit shares related to generative AI. Huang believes Japan can establish its own AI ecosystem.
"Japan has all the technological expertise and industrial capabilities to create its own AI," Huang told Nishimura. He added that Japan's decades of expertise in mechatronics, manufacturing, and robotics could revolutionize Japanese robotics technology when combined with generative AI.
Less than two weeks before Huang's visit to Japan, the country passed an additional budget, including about 2 trillion yen specifically for chip investments. Part of this funding is expected to support chip manufacturers like TSMC and Rapidus, a chip foundry company planning to manufacture cutting-edge chips in Hokkaido.
"The semiconductor industry that Japan is now starting to develop and nurture will be able to produce GPUs," said Huang.
NVIDIA provides chips and servers to Japanese companies like SoftBank Group, NEC Corporation, and Nippon Telegraph and Telephone Corporation. These companies are competing to build large Japanese language models and other AI infrastructures unique to Japan.
"To develop our own generative AI in Japan, we need NVIDIA's GPUs. Without collaboration, we cannot achieve this," said Nishimura. "I hope to see Japan lead the world in innovation by deepening its partnership with NVIDIA."