Amazon Triples Nvidia Chip Order as AI Demand Surges

Amazon and Nvidia have significantly deepened their strategic alliance, announcing an expanded partnership that includes a commitment to acquire 2 million additional Nvidia GPU chips for Amazon Web Services, effectively tripling a procurement deal secured just five months ago. The order, revealed during Nvidia's quarterly earnings call, covers Blackwell Ultra, Rubin, and Rubin Ultra architectures, with deliveries scheduled to commence in the third quarter of this year and continue through 2028. While financial terms were undisclosed, the scale of the acquisition points to a value in the tens of billions, a escalation Nvidia attributed to surging demand from startups, enterprises, AI labs, and governments that has exceeded internal expectations. The collaboration transcends hardware volume to encompass broad infrastructure integration. Nvidia will deploy its Vera CPUs across AWS, with some units delivered standalone and others integrated with Rubin GPUs, supporting CEO Jensen Huang's vision of a vast new market for its general-purpose processors. Additionally, Amazon will implement Nvidia's full physical AI stack, including Omniverse, Cosmos, Isaac, and Jetson technologies, to operate its warehouse robotics fleet. In the enterprise segment, AWS will integrate Nvidia's Nemotron family of open models into its Bedrock and SageMaker services. Nvidia reported second-quarter revenue of $96.2 billion, beating analyst estimates, with data center revenue surging 117% year-over-year to $89 billion. The company forecast third-quarter revenue at $108 billion, supported by the ramp of next-generation Rubin GPUs, for which production shipments have already begun. To secure supply chain stability amid intense demand, Nvidia raised its capital commitment for manufacturing and memory capacity to $279 billion, up from $119 billion last quarter, allocating $92 billion for the remainder of the fiscal year and $87 billion for fiscal 2028. Huang emphasized that the industry has reached a phase where AI generates profitable work, asserting that additional compute directly drives service profitability. Despite the reinforced reliance on Nvidia, Amazon continues to advance its custom silicon ambitions. AWS recorded a $25 billion annualized revenue run rate for its custom chip business, fueled by $225 billion in commitments from major AI labs. Amazon's Trainium processors, designed to compete with Nvidia's high-end offerings, and Graviton CPUs remain central to its strategy for infrastructure diversification. Nevertheless, the accelerated order volume and extensive technology integration highlight Nvidia's enduring dominance in the critical hardware layer powering the global AI infrastructure boom.