On August 26 local time in Seattle and Santa Clara, California, Amazon Web Services (AWS) and NVIDIA jointly announced an expansion of their strategic collaboration: AWS plans to deploy an additional 2 million NVIDIA graphics processors across its global infrastructure between 2027 and 2028, including models such as Blackwell Ultra, Rubin, and Rubin Ultra. The contract value was not disclosed by either party. On the same day, NVIDIA stated during its earnings call that Amazon will adopt its full physical artificial intelligence technology stack to support its warehouse robotics operations. Approximately five months earlier, AWS had already announced the addition of over 1 million NVIDIA GPUs starting from 2026. The newly added 2 million GPUs will be deployed across AWS's global infrastructure, covering AI factories and customer-facing cloud instances, primarily for areas such as agentic artificial intelligence, scientific computing, enterprise automation, and physical AI. Additionally, NVIDIA will provide AWS with an undisclosed quantity of Vera central processors, some integrated with Rubin and others delivered separately. The two companies also plan to build AI factories for the U.S. government, delivering 100,000 GPUs on AWS's secure infrastructure to run federal and national security workloads.