Amazon Web Services and NVIDIA Expand Collaboration to Add GPUs and Vera CPUs
Amazon Web Services (AWS) and NVIDIA outlined an expanded, multi-year collaboration to add more NVIDIA GPUs and extend infrastructure work across CPUs, networking, open models, data processing, and robotics. The update addresses rising demand for AI compute capacity and related platform capabilities across AWS’s global infrastructure.
The companies said they planned to deploy 2 million additional NVIDIA GPUs across AWS’s global infrastructure in 2027-2028 and to deepen their work across “AI factories,” CPUs, networking, open models, data processing and robotics. They also described plans to expand the NVIDIA GPU fleet, extend NVIDIA Vera CPU-based infrastructure to AWS, and continue support for NVIDIA Nemotron open models on Amazon Bedrock and Amazon SageMaker.
As part of the collaboration, AWS and NVIDIA said they would bring “NVIDIA Vera CPU-based infrastructure” to AWS, extend NVIDIA NVLink Fusion with custom NVIDIA high-bandwidth memory (NVHBM), and integrate the NVIDIA platform with the AWS Nitro System and Elastic Fabric Adapter (EFA). They also described GPU-accelerated data processing on Amazon EMR using Amazon EC2 G7 instances and the NVIDIA cuDF library, and GPU-accelerated vector indexing on Amazon OpenSearch Service using dedicated GPUs.
For robotics and public-sector work, the companies said they would build AI factories for the U.S. government, including plans for 100,000 GPUs on secure AWS infrastructure for federal and national-security workloads. AWS also described robotics collaboration plans with Amazon Robotics and NVIDIA’s physical AI platform, including NVIDIA Jetson, NVIDIA Omniverse libraries and the NVIDIA Isaac open robotics development platform. “Customers want the freedom to choose the best tools for their AI workloads, and they want confidence that everything works seamlessly together,” Matt Garman, CEO of AWS, said. “NVIDIA and AWS have built one of the great growth engines of the AI era, and demand is running ahead of every forecast,” Jensen Huang, founder and CEO of NVIDIA, said.
The companies also cited ongoing plans to expand NVIDIA Blackwell capacity in 2027-2028, support NVIDIA Spectrum networking for large-scale AI training workloads, and continue integrating the AWS platform with NVIDIA’s software and open models.
Provided by Globe Newswire on behalf of Amazon Web Services (AWS). Read the original.