AWS and NVIDIA Plan 2 Million GPU Expansion Across Global Infrastructure
AWS and NVIDIA have announced an expansion of their AI collaboration, planning to deploy two million additional NVIDIA Blackwell Ultra, Rubin, and Rubin Ultra GPUs across AWS's global infrastructure in 2027-2028. This includes building U.S. government AI factories with 100,000 GPUs on secure AWS infrastructure.
The companies plan to bring NVIDIA Vera CPU-based infrastructure to AWS, extend NVLink Fusion with NVHBM, and integrate NVIDIA's platform with AWS Nitro and EFA. They also aim to expand support for Nemotron open models, GPU-accelerated data processing, and vector indexing on Amazon EMR and OpenSearch.
AWS is adding NVIDIA RTX PRO 4500 Blackwell GPUs, with G7 instances delivering 4.6x AI inference performance compared to previous-generation G6 instances. The release also reports that GPU-accelerated EMR processing can be up to 3.7x faster with a 30% better price-performance ratio and GPU vector indexing on Amazon OpenSearch can be up to 9x faster at one-quarter the cost.