Amazon is buying 2 million Nvidia GPUs for AWS data center expansion
AWS plans to deploy two million additional Nvidia GPUs by 2028, expanding AI infrastructure across cloud, government, analytics, and robotics workloads.
Amazon plans to acquire 2 million Nvidia GPUs for its AWS data center expansion between 2027 and 2028. This follows an earlier commitment to add over a million chips starting in 2026. The partnership with Nvidia aims to provide more computing power for artificial intelligence workloads. The cloud provider will also introduce new Vera-based CPUs to support increasingly advanced AI applications.
Enhanced networking upgrades, including NVLink Fusion technology and custom high-bandwidth memory, are also part of the plan to boost performance across large computing clusters. Amazon's analytics service, powered by Nvidia's cuDF software, is expected to improve processing speeds by roughly 3.7 times compared to standard configurations.
Vector search capabilities on Amazon's search service are also improving, with index construction now completing nearly nine times faster when run on GPUs. The robotics division is working with Nvidia on simulation tools to speed up training for next-generation warehouse robots. 100,000 chips are reserved for sensitive government and defense computing needs.
The collaboration aims to give customers the flexibility to choose the best tools for their AI workloads while ensuring seamless integration. Matt Garman, CEO of AWS, emphasized deep investment with Nvidia to make AWS the best place to run Nvidia AI technologies, optimizing performance across the infrastructure. Jensen Huang, founder and CEO of Nvidia, highlighted the growing demand for accelerated computing and Nvidia's 16-year partnership with AWS, which is now expanding across the full stack to deliver AI at an unprecedented pace and scale.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
