Amazon and Nvidia expanded their AI infrastructure partnership on Wednesday, with Amazon committing to add another 2 million Nvidia GPUs to its data centers. The chips — Blackwell Ultra, Rubin, and Rubin Ultra — are designed for the heavy compute demands of training and running AI models, and will head to Amazon Web Services data centers in 2027 and 2028, starting in the third quarter. The announcement, made during Nvidia‘s quarterly earnings call, comes just five months after Amazon agreed to deploy more than 1 million Nvidia GPUs across AWS infrastructure starting this year. Nvidia said demand has since exceeded those expectations. Neither company shared financial terms, but given unit costs the deal is worth tens of billions of dollars. The announcement is notable not just for its size and the speed at which it grew, but because it extends beyond Amazon buying more Nvidia chips.

The partnership stretches well beyond GPU purchases. Nvidia said its networking hardware, CPUs, data processing software, open models, and robotics platform will be integrated across AWS. Nvidia CFO Colette Kress said an unspecified number of Vera CPUs will ship to AWS, some integrated with Rubin GPUs and others standalone. She said Nvidia expects Vera to be deployed by every major hyperscaler, neocloud, AI lab, and system OEM, with shipments already underway to lead partners such as Oracle and SpaceX AI. Nvidia CEO Jensen Huang said in May that he had found a ‘brand new $200 billion TAM’ for the company. Amazon also plans to adopt Nvidia‘s full physical AI stack for its warehouse robots, using Omniverse (simulation and digital twin), Cosmos (world models), Isaac (robotics development), and Jetson (robotics and edge AI computing). This week, Nvidia introduced a new version of Jetson designed as a more accessible robotics computer for entry-level edge AI. On the enterprise side, AWS will host Nvidia‘s Nemotron family of open models on Amazon Bedrock and SageMaker. Nvidia said ‘surging demand’ from startups, enterprises, AI labs, and governments drove the closer cooperation.

The deal comes as Amazon accelerates its own chip efforts to reduce dependence on Nvidia. Its Trainium chips are a direct alternative to Nvidia‘s H100 or Blackwell for deep learning workloads, and AWS has been in talks to sell Trainium to other companies for data center use. Amazon’s Arm-built Graviton CPU is seen as a challenger to Intel and AMD server chips. Amazon has said its custom chip business crossed a $25 billion annualized revenue run rate, supported by $225 billion in total commitments from AI labs such as Anthropic and OpenAI. Even so, Nvidia remains the dominant player in AI chips.

Nvidia also released financial results alongside the announcement. It recorded $96.2 billion in sales for the quarter, beating analyst estimates, with data center revenue of $89 billion, up 117% from a year earlier. Nvidia guided to $108 billion in revenue for the next quarter, some of which will come from its next-generation Rubin GPUs, which began production shipments this quarter. Investors have been looking out for Rubin‘s initial Q3 sales for signs that demand will continue into Nvidia‘s next generation of hardware. To secure supply and manufacturing capacity for current and future data center projects, Nvidia has committed $279 billion, up from $119 billion last quarter. That commitment includes $92 billion in projected spending for the rest of the fiscal year and another $87 billion in fiscal year 2028. CEO Jensen Huang said on the 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. This is the exact phase where we’re at, which is the reason why everybody’s leaning in.’ Investors will be watching whether additional compute indeed translates into additional profits as AI companies pour hundreds of billions of dollars into infrastructure.