SAN JOSE, Calif. - Google (NASDAQ:GOOGL) Cloud and NVIDIA (NASDAQ:NVDA) have announced an expansion of their partnership aimed at providing the machine learning community with advanced tools to build and manage generative AI applications more efficiently. This collaboration includes Google Cloud's adoption of NVIDIA's new Grace Blackwell AI computing platform and the NVIDIA DGX Cloud service.
The partnership also sees the general availability of the NVIDIA H100-powered DGX Cloud platform on Google Cloud, providing developers with a robust and open platform for training and deploying AI with preferred tools and frameworks. Additionally, support for JAX on NVIDIA GPUs and Vertex (NASDAQ:VRTX) AI instances powered by NVIDIA H100 and L4 Tensor Core GPUs has been announced.
Google Cloud CEO Thomas Kurian emphasized the comprehensive nature of the partnership with NVIDIA, noting the aim to deliver a highly accessible and complete AI platform for machine learning developers. NVIDIA's CEO, Jensen Huang, highlighted the solutions provided to enterprises for leveraging generative AI swiftly and effectively.
Key elements of the partnership expansion include the adoption of NVIDIA Grace Blackwell for real-time inference on large language models, the introduction of Grace Blackwell-powered DGX Cloud to Google Cloud, and support for JAX on NVIDIA GPUs to facilitate large-scale AI model training.
Moreover, NVIDIA NIM inference microservices will be integrated into Google Kubernetes Engine (GKE), speeding up the deployment of generative AI in enterprises. Google Cloud has also streamlined the deployment of the NVIDIA NeMo framework across its platform, enhancing the ability to automate and scale the training and serving of AI models.
Vertex AI and Dataflow have expanded support for NVIDIA GPUs, providing scalable infrastructure for managing and deploying AI applications. The partnership underscores the commitment to offer leading AI capabilities at every layer of the AI stack.
This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.