At AWS re:Invent, NVIDIA and Amazon Web Services unveiled a major expansion of their long-standing collaboration, integrating NVIDIA’s NVLink Fusion interconnect technology across AWS’s custom silicon platforms including the upcoming Trainium4 AI chips, Graviton CPUs, and the Nitro System to accelerate cloud-scale AI performance and streamline deployment and systems management, marking a “new milestone” in their 15-year journey together, as highlighted by Matt Garman, CEO of AWS. Using NVLink Fusion alongside NVIDIA’s MGX rack architecture, AWS aims to enhance its next-generation AI infrastructure, bolster time to market, and support cutting-edge capabilities across inference and agentic model training, while offering customers advanced GPU options like the Blackwell architecture in AWS AI Factories and ensuring sovereign AI deployments that maintain data control and regulatory compliance.
Also Read: Amazon to Invest Up to $50 Billion to Power U.S. Government AI and Supercomputing Infrastructure
Jensen Huang, founder and CEO of NVIDIA, emphasized that “GPU compute demand is skyrocketing more compute makes smarter AI, smarter AI drives broader use and broader use creates demand for even more compute. The virtuous cycle of AI has arrived,” adding that the NVLink Fusion integration will help “unify our scale-up architecture with AWS’s custom silicon to build a new generation of accelerated platforms,” thereby democratizing advanced AI globally. The partnership also extends into software, with NVIDIA’s Nemotron open models now integrated into Amazon Bedrock to support enterprise-scale generative AI applications, while collaborations on services like serverless GPU-powered vector indexing via Amazon OpenSearch and co-engineered development tools aim to simplify developers’ AI workflows. Beyond digital AI, the collaboration pushes into physical AI via platforms like NVIDIA Cosmos WFMs and Isaac Sim to accelerate robotics simulation and deployment on AWS. This expanded full-stack integration underscores the companies’ commitment to delivering efficient, scalable, and secure AI infrastructure to enterprises worldwide.
