AWS to deploy 2 million more NVIDIA GPUs by 2028

Amazon Web Services and NVIDIA ↗ 9.06% announced on Wednesday a major expansion of their strategic collaboration, with AWS planning to deploy 2 million additional NVIDIA GPUs across its global infrastructure in 2027 and 2028 to meet what both companies described as surging demand for AI computing power.

A Deepening Commitment to NVIDIA Hardware

The deal, which includes NVIDIA Blackwell Ultra, Rubin, and Rubin Ultra GPUs, comes on top of the more than 1 million NVIDIA GPUs that AWS announced at NVIDIA GTC 2026 in March it would begin installing this year. As Bloomberg reported, the commitment signals Amazon's Amazon.com, Inc. ↘ 1.64% continued reliance on NVIDIA's AI hardware despite its own competitive chipmaking efforts through its Annapurna Labs division, which produces the Trainium line of custom AI chips.aboutamazon+2

"NVIDIA and AWS have built one of the great growth engines of the AI era, and demand is running ahead of every forecast," said Jensen Huang, founder and CEO of NVIDIA. Matt Garman, CEO of AWS, said the expanded collaboration gives "frontier labs, enterprises and governments even more ways to build and deploy AI on AWS".nvidianews.nvidia+1

Beyond GPUs: CPUs, Networking, and Government AI

The partnership extends well beyond raw GPU capacity. AWS and NVIDIA are working to bring NVIDIA's Vera CPU-based infrastructure to AWS, designed to support agentic AI workloads requiring high-performance CPU compute alongside GPU acceleration. The companies also plan to build AI factories for the U.S. government, delivering 100,000 GPUs on AWS's secure infrastructure for federal and national-security workloads classified at Impact Level 6 and above.aboutamazon+1

Additionally, NVIDIA and Amazon's Annapurna Labs are expanding their work on NVLink Fusion interconnect technology with NVIDIA's custom high-bandwidth memory, which would allow AWS's Trainium chips and NVIDIA GPUs to operate within a common rack-scale architecture.nvidianews.nvidia

Complementary Rather Than Competing

AWS framed the expansion as complementary to its own custom silicon strategy rather than a concession. The company positioned itself as offering the broadest choice of compute — from AWS Trainium chips to the latest NVIDIA accelerators — letting customers select the best tools for their specific workloads. AWS is also the first major cloud provider to offer compute instances accelerated by NVIDIA's RTX PRO 4500 Blackwell Server Edition GPUs, which the companies said deliver 4.6 times the AI inference performance of previous-generation instances.aboutamazon+1