Amazon has agreed to add another 2 million Nvidia GPU chips to its data centers, tripling the size of a partnership announced just five months earlier, according to a joint announcement made Wednesday during Nvidia's quarterly earnings call. The new chips, which include Nvidia Blackwell Ultra, Rubin, and Rubin Ultra GPUs designed for training and operating AI models, will arrive at Amazon Web Services' data centers in 2027 and 2028. Nvidia stated that since the companies agreed in March to deploy more than 1 million Nvidia GPUs across AWS infrastructure starting this year, "demand has exceeded those expectations."
The expanded partnership extends well beyond additional chip purchases. Nvidia's technology—including networking hardware that links thousands of GPUs into unified systems, open models, CPUs, data processing software, and its robotics platform—will now be integrated throughout AWS, the companies said. Amazon plans to adopt Nvidia's complete physical AI stack to power its warehouse robot fleet, incorporating Omniverse (simulation and digital twin platform), Cosmos (world model platform), Isaac (robotics development platform), and Jetson (computing hardware for robots and edge AI). On the enterprise side, AWS will offer Nvidia's Nemotron family of open models through Amazon Bedrock and SageMaker. Nvidia CFO Colette Kress said the company expects its Vera CPUs to be deployed by "every major hyperscaler, neocloud, AI lab, and system OEM," with shipments already heading to lead partners including Oracle and SpaceXAI.
Neither company disclosed financial terms for the deal, but considering GPU unit costs, it's worth tens of billions of dollars. Nvidia reported Wednesday that it recorded sales of $96.2 billion for the second quarter, beating analyst estimates, with data center revenue making up the majority at $89 billion—up 117% from a year ago. The company expects revenue to reach $108 billion in the third quarter, some of which will come from its next-generation Rubin GPUs, which began production shipments this quarter. Nvidia has committed $279 billion to secure supply and manufacturing capacity for current and future data center projects, up substantially from $119 billion last quarter. That commitment includes $92 billion in projected spending for the remainder of the fiscal year and another $87 billion in fiscal year 2028.
The announcement arrives as Amazon ramps up its own AI chip development—particularly with CPUs, the general-purpose processors at the core of servers—in an effort to reduce reliance on Nvidia and potentially compete with the chip giant. Amazon's AI chief Peter DeSantis has said AWS is in discussions to sell its Trainium chips, which serve as a direct alternative to Nvidia's H100 or Blackwell chips for deep learning tasks, to other companies for data center use. Amazon's Arm-based Graviton CPU also challenges traditional server chips from Intel and AMD. The company has noted its custom chip business is expanding, crossing a $25 billion annualized revenue run rate driven by $225 billion in total commitments from AI labs like Anthropic and OpenAI. But the massive new Nvidia order suggests the chip giant remains dominant in AI hardware despite Amazon's internal efforts.
The companies attributed the expanded partnership to "surging demand" from startups, enterprises, AI labs, and governments. Nvidia CEO Jensen Huang explained the momentum during Wednesday's call: "The thing that matters for the industry is that AI is now doing productive and useful work. 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." Investors will be monitoring whether additional computing capacity translates so directly into higher profits as AI companies pour hundreds of billions of dollars into infrastructure. The partnership between Amazon and Nvidia continues to deepen even as both companies pursue potentially competing strategies, with Amazon building alternative chips while simultaneously becoming one of Nvidia's largest customers.