NVIDIA unveiled the GB200 Grace Blackwell Superchip and the B200 Tensor Core GPUs that are structured to provide the highest AI performance and real-time LLM interference for models.
The Blackwell Superchip is powered by 208 billion transistors along with the Blackwell-architecture GPUs that are made of custom-built 4NP TSMC process with two-reticle limited GPU.
NVIDIA Claims Revolutionary AI With Blackwell Platforms
According to NVIDIA, Blackwell is based on six revolutionary technologies. The platform can support doubled compute and model sizes with the new 4-bit floating point AI inference capabilities.
In addition, the Blackwell architecture is designed to offer AI-based preventative maintenance that can diagnose and forecast reliability issues. This means the systems can run uninterrupted for up to months, reducing operating costs.
"Generative AI is the defining technology of our time. Blackwell is the engine to power this new industrial revolution. Working with the most dynamic companies in the world, we will realize the promise of AI for every industry," said Jensen Huang, founder and CEO of NVIDIA.
Tech Companies Support NVIDIA Blackwell
In a press release, several companies are expected to adopt Blackwell on their systems. This includes Amazon Web Services, Dell Technologies, Google, Meta, OpenAI, xAI, Microsoft, and more.
"By bringing the GB200 Grace Blackwell processor to our datacenters globally, we are building on our long-standing history of optimizing NVIDIA GPUs for our cloud, as we make the promise of AI real for organizations everywhere," said Satya Nadella, executive chairman and CEO of Microsoft.
NVIDIA also promised that Blackwell is equipped with advanced confidential computing capabilities, protecting customer data without any compromise to performance. Moreover, a dedicated decompression engine supports the latest formats as the goal is to increase the GPU acceleration in data processing.
Blackwell-based products will be available to NVIDIA's partners starting later this year. The Blackwell product portfolio is also part of the NVIDIA AI Enterprise.
