Amazon triples Nvidia GPU order to 3 million chips amid surging AI demand

Aug 26, 2026

Amazon and Nvidia expanded their partnership on August 26, 2026, with Amazon ordering an additional 2 million Nvidia GPUs for AWS data centers, tripling its total commitment to 3 million chips. The deal comes amid surging AI demand and record Q2 revenue of $96.2 billion for Nvidia. While Amazon continues to develop its own competing Trainium and Graviton chips, which reached a $25 billion annualized revenue run rate, it remains heavily reliant on Nvidia's hardware, including its upcoming Vera CPUs and robotics stack.

Amazon GPU order expansion

▪The expanded Amazon Web Services order includes Nvidia Blackwell Ultra, Rubin, and Rubin Ultra GPUs, prompted by surging demand from startups, enterprises, AI labs, and governments
▪The August 2026 agreement to add 2 million Nvidia GPUs is on top of a prior agreement made five months earlier for Amazon to deploy more than 1 million Nvidia GPUs starting in 2026
▪Amazon and Nvidia announced an expanded partnership on August 26, 2026, to add 2 million Nvidia GPU chips to Amazon Web Services data centers in 2027 and 2028

Nvidia quarterly revenue growth

▪Nvidia projected third-quarter revenue of $108 billion, supported by initial sales of its next-generation Rubin GPUs which began production shipments in the second quarter of 2026
▪Nvidia reported second-quarter sales of $96.2 billion on August 26, 2026, with data center revenue making up $89 billion of the total

AWS infrastructure integration

▪Amazon plans to adopt Nvidia's physical AI stack, including the Omniverse, Cosmos, Isaac, and Jetson platforms, to power its fleet of warehouse robots
▪Amazon is developing its own custom AI chips, including Trainium deep learning chips and Arm-built Graviton CPUs, to compete with Nvidia and reduce its hardware dependence

Amazon custom chip competition

▪Nvidia has committed $279 billion to secure supply and manufacturing capacity for data-center projects, up from $119 billion in the previous quarter
▪Amazon reported that its custom chip business crossed a $25 billion annualized revenue run rate, driven by $225 billion in commitments from AI labs like Anthropic and OpenAI

Nvidia supply chain commitments

▪Amazon Web Services will serve Nvidia's Nemotron family of open models on its Amazon Bedrock and SageMaker cloud platforms
▪Nvidia's supply and manufacturing commitments include $92 billion in projected spending for the remainder of the fiscal year and $87 billion in fiscal year 2028