NVIDIA (CEO Jensen Huang), a leader in AI computing technology, is supporting Meta's data centers with its H100 GPUs to solve Meta's key AI problems.

NVIDIA announced on the 20th that it supports NVIDIA Hopper on Meta's next-generation AI platform, Grand Teton.

At the 2022 OCP Global Summit, an open computing project conference, Alexis Bjorlin, Vice President of Meta Infrastructure Hardware, said, “The Grand Teton system provides more memory, network bandwidth, and computing capacity compared to the previous generation Zion EX platform.”

AI models are widely used on Facebook for services such as news feeds, content recommendations, and hate speech identification.

In a preparatory speech for the conference, Bjorn said, “I am very pleased to introduce the new family at this summit,” and thanked NVIDIA for their deep collaboration on the design of Grand Teton and OCP’s continued support.

Named after a 13,000-foot mountain in one of Wyoming's two national parks, Grand Teton uses NVIDIA H100 Tensor Core GPUs to train AI models that require more computing power as they grow rapidly in size and capability, and Execute.

The NVIDIA Hopper architecture, which is the basis of the H100, includes a Transformer engine to accelerate these neural network tasks. It is called a foundational model because it can handle a set of applications that extend from natural language processing to healthcare, robotics, and more.

The NVIDIA H100 is designed for performance and energy efficiency. When connected to NVIDIA Networking across thousands of servers in a hyperscale data center, H100 accelerated servers can be 300 times more energy efficient than CPU-only servers.

“NVIDIA Hopper GPUs are built to solve the world’s hard problems while adding scalability and lowering costs, while delivering accelerated computing with greater energy efficiency and enhanced performance. By sharing the H100-based Grand Teton platform, Meta will soon enable system builders worldwide to access open designs for hyperscale data center computing infrastructure to empower AI across industries,” said Ian Buck, Vice President of Hyperscale and High Performance Computing at NVIDIA.