European Investor
27 Aug 2026, 10:22
AWS and NVIDIA Expand AI Partnership With 2 Million Additional GPUs
Amazon Web Services 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 in 2027 and 2028.
The expansion comes as demand for AI computing continues to accelerate. AWS said the new capacity will include NVIDIA Blackwell Ultra, Rubin and Rubin Ultra GPUs, supporting workloads ranging from agentic AI and scientific computing to enterprise automation and robotics.
The partnership extends well beyond GPUs. AWS plans to introduce infrastructure based on NVIDIA’s Vera CPUs, expand NVLink Fusion integration with Amazon’s Trainium chips, and use NVIDIA’s new high-bandwidth memory technology. The companies also plan to build secure AI factories for the U.S. government, including infrastructure incorporating 100,000 GPUs for federal and national-security workloads.
The agreement further expands NVIDIA’s presence across AWS software and services. Nemotron open models will remain available through Amazon Bedrock and SageMaker, while NVIDIA technology will accelerate data processing, vector indexing and Amazon’s robotics development.
The scale of the commitment is another strong signal that hyperscaler AI infrastructure spending remains robust. For NVIDIA, the planned deployment strengthens visibility for demand extending into the Rubin generation through 2028. For Amazon, it significantly expands AWS’s ability to compete for increasingly compute-intensive AI workloads.
Amazon Web Services 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 in 2027 and 2028.
The expansion comes as demand for AI computing continues to accelerate. AWS said the new capacity will include NVIDIA Blackwell Ultra, Rubin and Rubin Ultra GPUs, supporting workloads ranging from agentic AI and scientific computing to enterprise automation and robotics.
The partnership extends well beyond GPUs. AWS plans to introduce infrastructure based on NVIDIA’s Vera CPUs, expand NVLink Fusion integration with Amazon’s Trainium chips, and use NVIDIA’s new high-bandwidth memory technology. The companies also plan to build secure AI factories for the U.S. government, including infrastructure incorporating 100,000 GPUs for federal and national-security workloads.
The agreement further expands NVIDIA’s presence across AWS software and services. Nemotron open models will remain available through Amazon Bedrock and SageMaker, while NVIDIA technology will accelerate data processing, vector indexing and Amazon’s robotics development.
The scale of the commitment is another strong signal that hyperscaler AI infrastructure spending remains robust. For NVIDIA, the planned deployment strengthens visibility for demand extending into the Rubin generation through 2028. For Amazon, it significantly expands AWS’s ability to compete for increasingly compute-intensive AI workloads.
