Fujitsu and NVIDIA aim to build a new AI platform that connects from the cloud to robots through the joint development of AI semiconductors and autonomous agents.
🟧 Fujitsu and NVIDIA strengthen AI alliance to integrate CPU and GPU to realize “physical AI”
Fujitsu has announced a comprehensive partnership with NVIDIA in the field of AI. The two companies will jointly promote the development of “physical AI” such as AI semiconductors, autonomous agents, and robots, and accelerate the implementation of an AI society from both hardware and software aspects.
- Joint development of AI semiconductors: Fujitsu’s MONAKA Arm-based CPU and NVIDIA’s GPU are connected at high speed using NVLink Fusion technology. Optimize data transfer between CPUs – GPUs to maximize AI computing performance.
- Development of autonomous agents: Utilizing NVIDIA’s software development platforms “NeMo” and “CUDA”, we will build an environment where Fujitsu’s AI applications can be executed and learned autonomously.
- Demonstration of physical AI: In collaboration with Yaskawa Electric’s industrial robots, we will verify the mechanism by which AI makes decisions and acts in the real world. We will consider expanding into manufacturing, logistics, autonomous driving, etc.
This initiative is the first step in evolving AI from “intelligence in the cloud” to “intelligence that moves in the field.”
🟧The battle for leadership in the AI ecosystem and Japan’s re-challenge
In recent years, the development of AI has become the focus of competition, not only on models and data, but also on the optimization of the computing infrastructure itself. NVIDIA is already partnering with cloud providers in the United States and Taiwan, but the collaboration with Fujitsu is a strategic move with a view to building Japan’s own AI infrastructure.
Fujitsu has the high-performance CPU technology cultivated in the Fugaku supercomputer and is expanding its application for AI. NVIDIA, on the other hand, is the world leader in GPU and AI software stacks. The collaboration between the two companies could be an opportunity for Japan to regain its presence in the AI infrastructure race led by the United States and China.
Competitors such as AMD’s Instinct and Intel’s Gaudi 3 are also developing accelerators for AI, but NVLink Fusion’s approach to tightly coupled CPU and GPU is considered superior in terms of performance. In addition, in the field of physical AI such as robots and autonomous driving, the Fujitsu × Yaskawa Electric Alliance is expected to be an area where Japan companies can leverage their strengths.
🟧Summary
This partnership is positioned as a fusion of “domestic computing infrastructure + world-class GPU” in the AI era.
The two companies are not only completing AI in the cloud, but also aiming to bridge it to the real world such as robots, autonomous driving, and industrial automation.
Fujitsu’s move to evolve “Fugaku’s DNA” into AI infrastructure is an attempt to redefine Japan’s technological capabilities. With NVLink Fusion as the axis, domestic CPUs may once again be in the spotlight on the world stage.
