Nvidia CEO Jensen Huang has big plans for the AI data center industry, aiming to revolutionize the way companies approach artificial intelligence with its new Spectrum-X and Quantum-X silicon photonics networking systems.
The company’s systems will connect multiple Nvidia graphics chips together, enabling data centers to deploy up to 1 million GPU clusters. This is a massive jump from current largest AI supercomputer, xAI’s Colossus, which uses 100,000 Nvidia Hopper GPUs.
Huang sees the potential for creating “giant AI factories” that can connect multiple data centers, forming massive facilities. He believes this will be key to driving revenue opportunities and staying competitive in the industry.
“Nvidia AI enterprise is all about unlocking value from artificial intelligence,” Huang explained. The company’s advancements, including its new chips and networking systems, will help customers like Microsoft and Google operate more powerful AI systems.
In addition to its data center plans, Nvidia also showcased smaller AI-centered computers, including its mini desktop-sized DGX Spark and full-size desktop-sized DGX Station. These computers are designed for AI developers and researchers and can be paired with the company’s other AI platforms.
Huang estimates that enterprise IT represents a half-a-trillion-dollar market opportunity. The company aims to capitalize on this by providing an approach to enterprise IT that revolves around AI.
“Nvidia is about unlocking value from artificial intelligence,” Huang said. “It’s a new world, an industrial revolution.”
Source: https://finance.yahoo.com/news/nvidia-ceo-jensen-huang-says-well-soon-see-1-million-gpu-data-centers-visible-from-space-123801629.html