AWS will add 2 million Nvidia GPUs for AI infrastructure

AWS and Nvidia announced a major expansion of their AI infrastructure partnership, with AWS planning to deploy 2 million additional Nvidia GPUs across its global infrastructure. The companies say the added capacity will support workloads from agentic AI and scientific discovery to enterprise automation and physical AI.

The new GPU plan builds on an earlier AWS commitment to add more than 1 million Nvidia GPUs. Amazon says demand has run ahead of those expectations, so the 2027-2028 expansion will include AI factories across AWS infrastructure.

The partnership is also moving beyond GPUs. Nvidia Vera CPU-based infrastructure is coming to AWS as another option for agentic AI workloads that need high-performance CPU compute around accelerated systems, including tool use, code execution, sandboxing, data pipelines and reinforcement-learning loops. AWS’s Annapurna Labs will also work with Nvidia custom memory and NVLink Fusion technology for future Trainium chips.

For governments, AWS and Nvidia plan AI factories for U.S. federal and national-security workloads, including 100,000 GPUs on AWS secure infrastructure classified at Impact Level 6 and above. The broader signal is clear: hyperscale AI capacity is becoming a full-stack partnership problem, spanning chips, networking, memory, security boundaries and dedicated government infrastructure.