Nvidia and SoftBank have announced a groundbreaking partnership aimed at significantly enhancing Japan’s AI capabilities through two primary projects: the construction of an AI supercomputer and the deployment of a cutting-edge 5G AI-RAN network. This ambitious initiative, revealed by Nvidia CEO Jensen Huang at the AI Summit Japan, underscores the companies’ commitment to technological innovation and boosting computational power. The new AI supercomputer will be built on Nvidia’s Grace Blackwell platform, marking it as the most powerful AI system in Japan. It will utilize Nvidia’s DGX B200 systems, AI Enterprise software, and Quantum-2 InfiniBand networking, all elements critical to its high-performance design.
SoftBank, leveraging Nvidia’s DGX SuperPod, aims to propel its own AI research and development efforts. The DGX SuperPod is also expected to serve universities, research institutions, and businesses across Japan. This access to state-of-the-art computing infrastructure will empower various sectors to undertake complex AI projects, fostering a new era of innovation. Furthermore, an additional supercomputer featuring Nvidia’s GB200 NVL72 system is planned. This system, noteworthy for its liquid-cooled GB200 GPUs and high computational capacity, represents another significant step toward bolstering Japan’s AI infrastructure. Together, these developments highlight a mutually beneficial collaboration between Nvidia and SoftBank, promising considerable advancements in AI research and applications.
Development of the AI Supercomputer
Building the AI supercomputer on Nvidia’s Grace Blackwell platform signifies a remarkable leap in the realm of artificial intelligence for Japan. This system is designed to accommodate the most demanding AI workloads, including large-scale machine learning tasks and complex data analytics. The platform’s versatility and superior performance capabilities ensure it will play a pivotal role in driving AI innovation in various sectors. Nvidia’s DGX B200 systems, which form the core of this supercomputer, are purpose-built for AI applications, offering unmatched processing power and efficiency.
The integration of AI Enterprise software further enhances the system’s capabilities, providing a comprehensive suite of tools for AI development and deployment. Quantum-2 InfiniBand networking ensures high-speed data transfer and low latency, crucial for maintaining the system’s performance and efficiency. These technologies combined create an AI infrastructure that will empower researchers and developers to tackle some of the most challenging problems in AI. SoftBank, utilizing the DGX SuperPod, will not only advance its own AI initiatives but also democratize access to cutting-edge computing resources.
The additional supercomputer featuring Nvidia’s GB200 NVL72 system promises to complement the primary AI infrastructure. Its design, featuring liquid-cooled GB200 GPUs, is tailored for high computational capacity and energy efficiency. This system is expected to support a wide range of AI applications, from academic research to industrial innovation. By providing such advanced capabilities, Nvidia and SoftBank are positioning Japan as a leader in AI-driven technologies. The collaboration aims to create a robust AI ecosystem, facilitating unprecedented advancements in fields like natural language processing, computer vision, and autonomous systems.
Launch of the 5G AI-RAN Network
Nvidia and SoftBank have announced a landmark collaboration to significantly enhance Japan’s AI capabilities through two major initiatives: the creation of an AI supercomputer and the rollout of an advanced 5G AI-RAN network. Unveiled by Nvidia CEO Jensen Huang at the AI Summit Japan, this ambitious project highlights the companies’ dedication to technological innovation and computational power. The AI supercomputer will be based on Nvidia’s Grace Blackwell platform, marking it as Japan’s most powerful AI system. It will incorporate Nvidia’s DGX B200 systems, AI Enterprise software, and Quantum-2 InfiniBand networking, all critical to its high performance.
By using Nvidia’s DGX SuperPod, SoftBank aims to elevate its AI research and development activities. The DGX SuperPod will also be available to universities, research institutions, and businesses throughout Japan. This state-of-the-art computing infrastructure will enable various sectors to tackle complex AI projects, ushering in a new era of innovation. Additionally, another supercomputer featuring Nvidia’s GB200 NVL72 system, known for its liquid-cooled GB200 GPUs and high computational abilities, is planned. This collaboration between Nvidia and SoftBank promises notable advancements in AI research and applications in Japan.
