The expanded AWS deal, announced on Nvidia’s earnings call, adds GPUs for 2027 and 2028 and follows a prior 1 million chip order.
Amazon will add 2 million Nvidia GPUs to AWS data centers in 2027 and 2028, tripling its prior order from five months ago. Nvidia announced the expanded partnership during its quarterly earnings call, and the deal includes Blackwell Ultra, Rubin, and Rubin Ultra chips.
The expanded order follows Amazon’s earlier commitment to deploy more than 1 million Nvidia GPUs starting this year. Nvidia says demand has exceeded expectations, and the companies plan to integrate Nvidia networking hardware, open models, CPUs, data processing software, and robotics platform across AWS.
For builders, the deal means more AWS capacity for training and serving AI models, but it also signals that Nvidia remains the dominant supplier despite Amazon’s custom chip efforts. AWS also continues to develop its own Trainium and Graviton chips, and Amazon’s AI chief says AWS is in talks to sell Trainium to other companies.
Watch for Nvidia to ship Vera CPUs alongside the GPUs, with some integrated with Rubin and others standalone. Nvidia expects every major hyperscaler, neocloud, AI lab, and system OEM to deploy Vera, and CEO Jensen Huang has said the chip opens a $200 billion market.
What matters
- Amazon will add 2 million Nvidia GPUs to its AWS data centers during 2027 and 2028.
- For builders, the deal means more AWS capacity for AI workloads, but GPU supply remains tight.
- Watch for AWS to sell its Trainium chips externally and challenge Nvidia’s dominance.
Why it matters
Watch for AWS to sell its Trainium chips externally and challenge Nvidia’s dominance.
This GenAI News article was prepared in original wording using reporting and materials published by TechCrunch AI. Source reference: https://techcrunch.com/2026/08/26/amazon-just-tripled-its-order-of-nvidia-chips-over-surging-demand/.
Drafted by the GenAI News review pipeline.