Amazon and Nvidia announced Wednesday an expanded partnership that includes a deal to add 2 million more Nvidia GPU chips to Amazon's data centers over the next two years, the companies said during Nvidia's quarterly earnings call.
The GPUs, which include Nvidia Blackwell Ultra, Rubin, and Rubin Ultra chips, will be deployed to Amazon Web Services data centers in 2027 and 2028. The announcement comes just five months after Amazon agreed to deploy more than 1 million Nvidia GPUs across AWS infrastructure starting this year, and Nvidia said "demand has exceeded those expectations."
Neither company shared financial terms, but considering GPU unit costs, the deal is worth tens of billions of dollars. The expansion is notable not just for its size and speed, but because it extends beyond chip purchases to include Nvidia's networking hardware, open models, CPUs, data processing software, and robotics platform.
The partnership is significant because it comes even as Amazon invests in its own competing AI chips. Amazon's AI chief Peter DeSantis has said AWS is in talks to sell its Trainium chips to other companies, and the company's custom chip business crossed a $25 billion annualized revenue run rate.
Nvidia CEO Jensen Huang said the company expects revenue to reach $108 billion in the third quarter, some of which will come from next-gen Rubin GPUs now in production. Nvidia has committed $279 billion to secure supply and manufacturing capacity for current and future data-center projects.
"The thing that matters for the industry is that AI is now doing productive and useful work," Huang said. "AI is generating profitable tokens... If we had more compute, we could generate more profitable tokens, which results in more profit for all of the services."
Amazon also plans to adopt Nvidia's full physical AI stack — including Omniverse, Cosmos, Isaac, and Jetson — to power its fleet of warehouse robots, while AWS will serve Nvidia's Nemotron open models on Amazon Bedrock and SageMaker.
