Amazon has significantly expanded its partnership with Nvidia, announcing a deal to triple its order of GPU chips to meet surging demand for artificial intelligence (AI) capabilities. The agreement, revealed during Nvidia’s quarterly earnings call, will see Amazon add an additional 2 million Nvidia GPU chips to its data centers, specifically designed to handle the heavy computational requirements of training and running AI models.

The new GPUs, which include Nvidia’s Blackwell Ultra, Rubin, and Rubin Ultra models, are scheduled to be deployed in Amazon Web Services (AWS) data centers in 2027 and 2028. This announcement follows a previous agreement made just five months ago, where Amazon committed to deploying over 1 million Nvidia GPUs across its infrastructure starting this year. Nvidia noted that demand has exceeded initial expectations since that agreement.

Financial Implications and Market Context

While neither company disclosed the financial terms of the expanded partnership, industry analysts suggest that the deal could be worth tens of billions of dollars, given the unit costs of the GPUs. The rapid growth of this partnership is noteworthy, especially as Amazon simultaneously invests in developing its own AI chips, which could compete with Nvidia’s offerings.

Nvidia’s technology will not only include the GPUs but also its networking hardware, open models, CPUs, data processing software, and robotics platform, all of which will be integrated into AWS. The companies cited “surging demand” from various sectors, including startups, enterprises, AI labs, and government entities, as a driving factor behind this expanded collaboration.

Amazon’s AI Chip Initiatives

Amazon is actively working on its own AI chip solutions, particularly focusing on CPUs that serve as general-purpose processors in servers. The company has been developing its Trainium chips, which are designed as direct alternatives to Nvidia’s H100 and Blackwell chips for deep learning workloads. Additionally, Amazon’s Graviton CPU is positioned as a competitor to traditional server chips from Intel and AMD.

Amazon’s custom chip business has seen significant growth, recently reporting a $25 billion annualized revenue run rate, driven by commitments from AI labs such as Anthropic and OpenAI. Despite these advancements, Nvidia remains a dominant player in the AI chip market.

Future Developments and Broader Implications

With the addition of 2 million GPU chips starting in the third quarter, Nvidia also plans to supply an unspecified number of Vera CPUs, which will be integrated with some of the new GPUs. Nvidia’s CEO Jensen Huang has expressed ambitious plans for the Vera CPUs, identifying a potential $200 billion market opportunity.

The partnership will extend to Amazon’s warehouse robots and enterprise offerings, with plans to adopt Nvidia’s full physical AI stack to enhance its robotic fleet. This includes platforms such as Omniverse, Cosmos, Isaac, and Jetson, which are designed for simulation, robotics development, and edge AI computing.

As Nvidia continues to report strong financial performance, with $96.2 billion in sales for the second quarter, the company anticipates revenue to reach $108 billion in the third quarter. This growth is largely attributed to the increasing demand for AI capabilities, prompting Nvidia to commit $279 billion to secure supply and manufacturing capacity for current and future data center projects.

As the AI landscape evolves, the collaboration between Amazon and Nvidia highlights the growing importance of advanced computing capabilities in driving innovation across various sectors. For further details, you can read the full article on TechCrunch.