AWS and Nvidia announced a major expansion of their strategic partnership on August 26, 2026. The two companies plan to deploy 2 million additional Nvidia GPUs across Amazon's global infrastructure and deepen their integration across the entire AI stack. They're responding to surging demand for AI compute from frontier labs, global enterprises, startups, and governments.
Key Facts
- 2 million GPUs to be deployed on AWS infrastructure between 2027 and 2028
- Nvidia brings Vera CPUs, NVLink Fusion with custom high-bandwidth memory, and Nemotron models to AWS
- AWS and Nvidia building AI Factories for the U.S. government, including 100,000 GPUs for federal and national security workloads
- Partnership includes robotics integration via Amazon Robotics and Nvidia's Physical AI platform
What's Changing in Practice
The collaboration goes beyond mere GPU provisioning. Nvidia will make its Vera CPU-based infrastructure available on AWS and extend NVLink Fusion technology with custom high-bandwidth memory (NVHBM). Both companies are integrating Nvidia's platform deeper with the AWS Nitro System and Elastic Fabric Adapter (EFA) – technologies for enhanced security and reliability.
For enterprise customers: Nvidia's Nemotron model family will continue to be available on Amazon Bedrock and Amazon SageMaker. Additionally, the companies are accelerating data processing and vector indexing through Nvidia's cuDF and cuVS CUDA-X libraries on Amazon EMR and Amazon OpenSearch.
The AI Compute Arms Race
"Customers want the freedom to choose the best tools for their AI workloads, and they want confidence that everything works seamlessly together. That's why we've invested deeply with NVIDIA to make AWS the best place to run NVIDIA AI technologies, optimizing performance across our infrastructure from networking and security to deployment."
This is how AWS CEO Matt Garman describes the strategy. The announcement reveals how intense the competition for AI infrastructure has become. Workloads are scaling rapidly – from model training through agentic AI to physical AI and robotics. Customers are moving from pilots to production and need not just more chips, but also security, reliability, and breadth in model selection.
Particularly noteworthy: The partnership includes AI Factories for the U.S. government, including 100,000 GPUs on secure AWS infrastructure for federal and national security workloads. This signals that AI compute is now treated as strategic infrastructure.
Implications for European Enterprises
This expansion has indirect but significant implications for European companies. First: The massive capacity increase at AWS/Nvidia could put medium-term pressure on GPU prices – but only if the new chips are actually deployed by 2028. Until then, compute remains scarce and expensive. Second: Companies running AI workloads on AWS benefit from better Nvidia integration and new model options. Third: The focus on robotics and physical AI suggests this sector is becoming mainstream – potentially relevant for German automation specialists. At the same time, dependence on two U.S. corporations intensifies, which poses a risk for European enterprises as long as no genuine European alternatives exist.
Sources
Editorially owned by Ideal Syka. Sources and method: Newsroom & method. Tips and corrections: ai@i6eal.de.
