Amazon Web Services has agreed to deploy an additional 2 million Nvidia GPUs across its global infrastructure, extending a hardware partnership the two US companies have run for 16 years. The rollout is scheduled through 2027 and 2028 and covers Blackwell Ultra, Rubin, and Rubin Ultra processors.
The agreement reaches past accelerator supply. AWS will bring Nvidia Vera CPU-based infrastructure onto its platform, extend Nvidia NVLink Fusion with custom high-bandwidth memory, and build AI factories for the US government running 100,000 GPUs on secure AWS infrastructure. Nvidia Nemotron models will be offered through AWS, and Amazon Robotics will adopt the Nvidia physical AI platform. The company is also expanding Blackwell capacity with Nvidia RTX Pro 4500 Blackwell Server Edition GPUs for its EC2 G7 instances, which AWS reports deliver 4.6 times the AI inference performance and 2.1 times the graphics performance of the previous G6 generation.
At Nvidia GTC 2026, AWS said it planned to bring more than one million Nvidia GPUs online during 2026. Customer demand has since run past that figure. AWS chief executive Matt Garman said, "Customers want the freedom to choose the best tools for their AI workloads, and they want confidence that everything works seamlessly together." Garman has previously said the cloud operator has never retired an A100 server, keeping six-year-old accelerators in service alongside new silicon.
Nvidia reported second quarter 2026 revenue of $96.2 billion, up 18 percent from the prior quarter and 106 percent from a year earlier, with GAAP and non-GAAP gross margins both at 75 percent.
Source: Data Center Dynamics - https://www.datacenterdynamics.com/en/news/aws-to-deploy-2-million-additional-nvidia-gpus/