Google Cloud and NVIDIA have announced their partnership expansion to scale AI

Google also plans to implement NVIDIA NIM inference microservices, expanding on their recent partnership

Today, Google Cloud and NVIDIA announced a consolidated partnership to provide the machine learning (ML) community with tools that facilitate the rapid development, scalability, and management of generative AI applications. Google announced the deployment of the NVIDIA DGX Cloud service on Google Cloud and the new NVIDIA Grace Blackwell AI computing platform in order to continue providing AI advances to its businesses and developers. Furthermore, the DGXTM Cloud platform, which is powered by the NVIDIA H100, is now widely accessible on Google Cloud. Google also plans to implement NVIDIA NIM inference microservices, expanding on their recent partnership to improve the Gemma family of open models. This will give developers an open and flexible platform to train and deploy using their choice of tools and frameworks.

Additionally, both companies announced support for Vertex AI instances powered by NVIDIA H100 and L4 Tensor Core GPUs, as well as JAX on NVIDIA GPUs. The companies' long-standing dedication to offering the AI community advanced features at every level of the AI stack is strengthened by the new integrations between NVIDIA and Google Cloud. For a long time, Google Cloud has provided GPU virtual machines (VM instances) that combine cutting-edge NVIDIA technology with cutting-edge Google technologies. The Google Cloud AI Hypercomputer is a supercomputing architecture that combines open software, performance-optimized hardware, and flexible consumption patterns. One of its main components is an NVIDIA GPU.