OpenAI struck a $38 billion cloud partnership with Amazon Web Services, sending Amazon’s stock higher as investors bet the deal will funnel more AI workloads onto AWS and bolster its standing in the cloud race.
On November 3, 2025, OpenAI and Amazon announced a multi‑year agreement that makes AWS a core provider of compute for OpenAI’s models, including ChatGPT. The seven‑year arrangement represents a $38 billion commitment for cloud capacity and gives OpenAI access to clusters built with “hundreds of thousands” of Nvidia AI chips, with additional capacity slated to come online through 2026 and room to expand into 2027 and beyond, the companies said. OpenAI can begin using AWS immediately, with deployment targeted to ramp significantly by the end of 2026.
Investors cheered the news: Amazon shares rose as much as about 5% following the announcement, reflecting expectations that the tie‑up will drive additional demand for AWS and reaffirm its competitiveness against Microsoft and Google in AI infrastructure.
The deal underscores the industry’s overarching constraint: compute. As model training and inference grow more intensive, leading AI developers are diversifying across multiple cloud providers to secure reliable capacity at scale. For AWS, bringing OpenAI’s workloads onto its infrastructure is a marquee win that may help counter the perception that rivals have pulled ahead in generative AI. For OpenAI, the pact diversifies cloud supply after changes to its Microsoft relationship and supports plans to scale next‑generation systems. Sam Altman captured the backdrop succinctly, saying, “Scaling frontier AI requires massive, reliable compute.”
The partnership arrives amid a broader build‑out of data center capacity and chip supply to meet surging AI demand. Amazon highlighted specialized EC2 UltraServers networked for low‑latency performance, while OpenAI emphasized the ability to scale into vast CPU and GPU clusters for training and inference. Capacity targeted by the end of 2026 suggests near‑term pressure on supply chains and continued capital spending across the AI stack.
The move also reflects a multi‑cloud strategy taking hold across the sector, even as the largest players maintain deep strategic ties. Amazon, for instance, also backs Anthropic, another leading AI lab. OpenAI’s arrangement with AWS signals that competition for the most compute‑hungry workloads remains wide open—and expensive.
Watch for AWS to detail additional chip deployments and data center expansions, and for OpenAI to outline how new capacity advances model capabilities and product performance. Market attention will likely focus on revenue impacts for AWS in upcoming quarters and whether other AI labs follow with similarly large, multi‑year capacity commitments.