AWS and NVIDIA expand cloud infrastructure partnership

AWS and NVIDIA are expanding their strategic collaboration to meet rising demand for AI infrastructure across frontier labs, enterprises, startups and governments. AWS plans to deploy an additional 2 million NVIDIA Blackwell Ultra, Rubin and Rubin Ultra GPUs in 2027-2028 across its global infrastructure, including AI factories for workloads such as agentic AI, scientific discovery, enterprise automation and physical AI.

The companies are broadening integration across the stack with NVIDIA Vera CPU-based infrastructure, NVIDIA NVLink Fusion with custom NVIDIA high-bandwidth memory, NVIDIA Spectrum networking, and tighter support for the AWS Nitro System and Elastic Fabric Adapter. AWS and NVIDIA also plan AI factories for the U.S. government, including 100,000 GPUs on secure AWS infrastructure for federal and national-security workloads.

NVIDIA Nemotron open models will continue to be supported on Amazon Bedrock and Amazon SageMaker. The collaboration also extends to GPU-accelerated data processing on Amazon EMR, where NVIDIA cuDF can deliver up to 3.7x faster processing speeds and a 30% better price performance compared with CPU-based configurations. GPU-accelerated vector indexing on Amazon OpenSearch Service can deliver up to 9x faster vector indexing at a quarter of the cost, while Amazon Robotics is adopting NVIDIA physical AI tools for warehouse automation and next-generation robots.