Amazon Triples NVIDIA GPU Commitment as AI Demand Surges

Amazon Web Services is expanding its dependence on NVIDIA even as it develops in-house Trainium accelerators and Graviton processors. Demand for computing power to train and run increasingly sophisticated AI models has surged among enterprises, AI laboratories and governments, making access to advanced GPUs a strategic constraint for cloud providers. AWS had announced in March 2026 that it would deploy more than 1 million NVIDIA GPUs starting that year, spanning the Blackwell and Rubin architectures.

Amazon and NVIDIA said on August 26, 2026, that AWS would add 2 million Blackwell Ultra, Rubin and Rubin Ultra GPUs across its global data centers and AI factories in 2027 and 2028. The new commitment effectively triples the more than 1 million GPUs announced five months earlier, taking the planned deployment above 3 million chips. Neither company disclosed financial terms, though the order is estimated to be worth tens of billions of dollars. AWS will also integrate more NVIDIA networking, software, open models and physical-AI technology.