Nvidia Corp. (NVDA) and Amazon.com Inc. (AMZN) just took their 15-year partnership to a new level. At AWS re:Invent, the two tech giants unveiled a deeper integration that connects Nvidia's accelerated computing technology with Amazon's custom silicon and sprawling cloud infrastructure. The goal? Give enterprises faster, more efficient ways to train and deploy advanced AI systems.

Here's what makes this interesting: Amazon is building Nvidia's technology directly into its own chips. At the conference, the companies announced expanded support for Nvidia NVLink Fusion, a new interconnect architecture that links Nvidia graphics processing units with AWS-designed chips like the next-generation Trainium4 and Graviton processors. Think of it as creating a unified fabric where Nvidia's compute power and Amazon's custom silicon work together seamlessly.

AWS is designing Trainium4 to integrate with NVLink and Nvidia MGX, marking the first step in what both companies describe as a multigenerational collaboration. Amazon is building its future cloud-scale AI platforms around this unified fabric, using NVLink Fusion and Nvidia's MGX rack architecture to deliver better performance across the board.

Sovereign AI and the Latest Hardware

AWS is also expanding access to Nvidia's latest Blackwell GPUs, which will power new AWS AI Factories. These are dedicated sovereign AI clouds that let global organizations run advanced AI workloads locally while maintaining full control of their data to meet increasingly strict regulatory requirements.

Nvidia says this approach accelerates the "AI industrial revolution" by making high-end computing infrastructure available to every company and every country. It's part tech strategy, part geopolitical positioning.

Beyond hardware, Nvidia is embedding more of its software ecosystem inside AWS. Nvidia Nemotron open models are now available on Amazon Bedrock for production-scale generative AI applications. AWS is also the first major cloud provider to offer serverless vector indexing with Nvidia GPUs, an upgrade that significantly speeds up unstructured data processing for retrieval-augmented generation and other agentic AI techniques.

Robots and the Physical World

The collaboration extends into physical AI as well. Nvidia's Cosmos world models and Isaac robotics tools now run natively on AWS services, helping robotics companies simulate, train, and validate real-world automation systems at scale. Because apparently, virtual AI wasn't enough—now we're talking about the machines that move around in actual warehouses and factories.

Nvidia and AWS position this expanded partnership as the next phase of their decade-and-a-half collaboration, one that's accelerating as demand for AI compute explodes.

What the Leaders Are Saying

"GPU compute demand is skyrocketing… The virtuous cycle of AI has arrived," said Jensen Huang, Nvidia CEO. "With Nvidia NVLink Fusion coming to AWS Trainium4, we're unifying our scale-up architecture with AWS's custom silicon to build a new generation of accelerated platforms. Together, NVIDIA and AWS are creating the compute fabric for the AI industrial revolution, bringing advanced AI to every company, in every country, and accelerating the world's path to intelligence."

AWS CEO Matt Garman said the collaboration "will bring new capabilities to customers so they can innovate faster than ever before." He added, "The upcoming support of NVIDIA NVLink Fusion in AWS Trainium4, Graviton and the Nitro System will bring new capabilities to customers so they can innovate faster than ever before."

Nvidia, now valued at $4.4 trillion, has gained more than 35% year-to-date on strong GPU demand. Amazon, valued at $2.5 trillion, is up 7%.

The Dual-Track Strategy

Here's where things get clever. Both companies are executing a dual-track AI strategy that keeps them central to the broader platform race, no matter which direction the market moves.

Amazon is selling large volumes of Nvidia GPUs to OpenAI through AWS, strengthening Nvidia's chip dominance and boosting AWS cloud revenue. At the same time, Amazon is scaling Anthropic on its Trainium2 custom silicon, a move aimed at lowering AI training costs and influencing long-term GPU economics.

If OpenAI continues to scale, Nvidia benefits from premium hardware demand. If Anthropic's lower-cost approach wins out, Amazon gains more control over AI infrastructure margins. Either outcome keeps the Nvidia–Amazon partnership at the center of next-generation AI development and commercialization.

It's hedging at scale, and it's working.

Price Actions: NVDA stock was trading higher by 1.33% to $182.32 at last check Tuesday. AMZN was up 0.80%.