Nvidia Unveils its New Blackwell GPU Architecture Optimized for Generative AI Workloads

At its GPU Technology Conference yesterday, Nvidia revealed its new Blackwell GPU architecture designed to accelerate the generative AI revolution. The company’s new flagship AI product is the Nvidia GB200 Grace Blackwell superchip, which combines two B200 Tensor Core GPUs with 208 billion transistors with an NVIDIA Grace CPU to make it the “world’s most powerful chip.”

Nvidia’s GB200 Grace Blackwell superchip will enable organizations to develop trillion-parameter-scale AI models while reducing costs and energy consumption by an up to 25x factor. OpenAI, Microsoft, Google, or any other company working on generative AI today already needs Nvidia’s state-of-the-art AI GPUs to train their large-language models. And Nvidia has already announced that AWS, Google, Meta, Microsoft, OpenAI, Oracle, Dell, Tesla and xAI will be the first companies to adopt the company’s new Blackwell GPU architecture.

“Generative AI is the defining technology of our time,” said Jensen Huang, founder and CEO of NVIDIA. “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.”

In a separate announcement yesterday, Microsoft detailed how it’s going to leverage Nvidia’s latest AI GPUs to improve its AI offerings. The company plans to make NVIDIA’s new Grace Blackwell GB200 and Quantum-X800 InfiniBand switch (for connecting GB200-powered systems) to Azure. Nvidia’s DGX Cloud AI platform for enterprise developers is also getting a native integration with Microsoft Fabric, Microsoft’s AI-powered analytics platform.

“We are committed to offering our customers the most advanced infrastructure to power their AI workloads,” said Satya Nadella, executive chairman and CEO of Microsoft. “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.”

In addition to Azure, Nvidia also announced yesterday that AWS, Google Cloud, and Oracle will be among the first cloud providers to provide Blackwell-powered instances to customers.