Amazon and Nvidia have announced an expanded partnership that includes Amazon’s commitment to purchase an additional 2 million Nvidia GPU chips for deployment in Amazon Web Services (AWS) data centers in 2027 and 2028. This order triples Amazon’s recent procurement of over 1 million Nvidia GPUs, following “surging demand” from startups, enterprises, AI labs, and governments. The GPUs in question include Nvidia’s Blackwell Ultra, Rubin, and Rubin Ultra models, designed for the intensive compute requirements of training and running AI models. Neither company disclosed the financial terms of the deal, though industry estimates suggest it could be worth tens of billions of dollars given the unit costs of the GPUs.
The partnership extends beyond mere hardware procurement, with Nvidia’s technology—including networking hardware, open models, CPUs, data processing software, and robotics platforms—to be integrated across AWS. This integration includes the deployment of Nvidia’s Vera CPUs, some of which will be integrated with Rubin GPUs, while others will operate standalone. Nvidia CFO Colette Kress indicated that Nvidia expects Vera CPUs to be adopted by major hyperscalers, AI labs, and system OEMs, with shipments already underway to partners like Oracle and SpaceXAI. Additionally, the collaboration will see Amazon adopting Nvidia’s full physical AI stack to power its fleet of warehouse robots, utilizing platforms such as Omniverse, Cosmos, Isaac, and Jetson. On the enterprise side, AWS will serve Nvidia’s Nemotron family of open models through Amazon Bedrock and SageMaker. Nvidia reported record second-quarter sales of $96.2 billion, with data center revenue constituting the majority at $89 billion, reflecting a 117% year-over-year increase. The company anticipates revenue of $108 billion in the third quarter, partly driven by the production shipments of its next-generation Rubin GPUs. Nvidia has committed $279 billion to secure supply and manufacturing capacity for current and future data-center projects, underscoring its strategic focus on meeting the burgeoning demand for AI infrastructure. Source