How AWS and NVIDIA Are Powering Agentic and Physical AI

NVIDIA CEO Jensen Huang (Credit: Tomohiro Ohsumi/Getty Images)

Jensen Huang’s NVIDIA and AWS scale their global partnership, deploying 2M GPUs to power next-generation agentic and physical AI enterprise infrastructure

AWS and NVIDIA have announced a major expansion of their long-standing strategic collaboration to meet surging global demand for AI infrastructure.

As the AI buildout continues, compute power is emerging as one of the world’s fastest-expanding resources. The consulting firm McKinsey estimates that by 2030, data centres are projected to require US$6.7tn worldwide to keep pace with the unprecedented demand for compute power.

The two leading firms at the centre of the AI boom are building on already-rapid customer adoption of NVIDIA-accelerated compute on AWS; on top of this, the firms plan to deploy two million additional NVIDIA GPUs across AWS’s global infrastructure.

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

In March, AWS announced plans to add more than one million NVIDIA GPUs starting in 2026. Credit: NVIDIA

Powering physical and agentic AI

The companies highlight that they’re deepening work together across AI factories, CPUs, networking, open models, data processing and robotics, with NVIDIA arguing that together the companies will deliver co-engineered AI solutions that enable customers to accelerate “AI development and deployment” at “unprecedented scale”.

Matt Garman, CEO of AWS, says: “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. Credit: Amazon Web Services (AWS)

“That’s why we’ve invested deeply with NVIDIA to make AWS the best place to run NVIDIA AI technologies, optimising performance across our infrastructure from networking and security to deployment. This expanded collaboration gives frontier labs, enterprises and governments even more ways to build and deploy AI on AWS.”

Key partnership specs

• NVIDIA Vera CPUs coming to AWS
• New EC2 G7 instances boost inference 4.6x
• Federal AI factories built for IL6 security
• Amazon Robotics adopts Jetson, Omniverse and Isaac
• Annapurna Labs connects Trainium via NVLink Fusion

What it means

In March, AWS announced plans to add more than one million NVIDIA GPUs starting in 2026. Since then, demand has exceeded those expectations – so AWS plans to deploy an additional two million NVIDIA Blackwell Ultra, Rubin and Rubin Ultra GPUs through 2027 and 2028 – across AWS Global Infrastructure, including AI factories.

“NVIDIA and AWS have built one of the great growth engines of the AI era, and demand is running ahead of every forecast,” says Jensen Huang, Founder and CEO of NVIDIA.

Jensen Huang, Founder, President and CEO at NVIDIA (Credit: Getty)

“For 16 years, we have scaled NVIDIA computing in the cloud together. Now, we are expanding our partnership across the full stack – GPUs, CPUs, networking, open models and software – to make agentic and physical AI real at an unprecedented pace and scale that only AWS and NVIDIA can deliver. This expansion reflects customers’ demand for NVIDIA’s platform on AWS.”

The companies highlight the expanded partnership means:

• Extending NVIDIA NVLink Fusion with custom NVIDIA high‑bandwidth memory
• Building AI factories for the US government, including 100,000 GPUs on secure AWS infrastructure for running federal and national‑security workloads
• Integrating the NVIDIA platform with the AWS Nitro System and Elastic Fabric Adapter for enhanced security and reliability
• Continuingto support NVIDIA Nemotron open models on Amazon Bedrock and Amazon SageMaker
• Acceleratingdata processing and vector indexing on Amazon EMR and Amazon OpenSearch with NVIDIA cuDF and cuVS CUDA-X libraries
• Further advancingrobotics workloads through Amazon Robotics’ adoption of NVIDIA’s physical AI platform

It’s exciting to see AWS and NVIDIA making AI production-ready at scale

Skylar Graika, CTO and Co-founder of Patrol 6

Additional capacity

This additional capacity will help power customer workloads ranging from agentic AI and scientific discovery to enterprise automation and physical AI.

For AWS customers building on the platform, this expanded stack directly targets the industry's primary hurdle – bridging the gap between initial AI pilots and full production scale. Writing on LinkedIn about the development, Skylar Graika, CTO and Co-founder of the AI-first body camera firm, Patrol 6 (which runs on AWS), said: “It’s exciting to see AWS and NVIDIA making AI production-ready at scale.

Skylar Graika, CTO and Co-founder of Patrol 6. Credit: Skylar Graika/LinkedIn

“Moving from pilot to full deployment is where so many teams get stuck, and these integrations could really accelerate real-world impact.”

NVIDIA's key partners

Apart from its obvious long-standing collaboration with AWS, NVIDIA works with a variety of major hardware and cloud providers. Some of them include:

• TSMC: As the world’s leading pure-play semiconductor foundry, TSMC manufactures NVIDIA’s most advanced AI and graphics chips, including the Blackwell and Rubin architectures. Utilising cutting-edge fabrication nodes alongside advanced Chip-on-Wafer-on-Substratepackaging technology, TSMC provides the physical manufacturing foundation and silicon scale necessary to bring NVIDIA's complex processor designs to life.

• Dell Technologies: A global powerhouse in enterprise server and storage infrastructure, Dell collaborates with NVIDIA to deliver integrated enterprise solutions like the "Dell AI Factory with NVIDIA." By pairing NVIDIA GPUs, Spectrum-X networking and AI software with Dell PowerEdge servers, the partnership offers scalable, turn-key AI deployment for enterprises running on-premises data centers and hybrid cloud environments.

• Foxconn: The world’s largest electronics contract manufacturer, Foxconn works with NVIDIA to manufacture high-density AI server systems, including GB200 NVL72 racks. Foxconn also integrates NVIDIA’s physical AI ecosystem, combining Omniverse digital twins and Isaac robotics, to build next-generation automated AI factories, smart manufacturing plants and autonomous vehicle platforms across its operations.