Amazon's AI Gold Rush: Tech Giant Triples Nvidia Chip Orders Amid Skyrocketing Demand!

Amazon and Nvidia have announced a significantly expanded partnership, marking a deeper collaboration between the two technology giants. This deal includes Amazon adding another 2 million Nvidia GPU chips, specifically the advanced Blackwell Ultra, Rubin, and Rubin Ultra GPUs, to its Amazon Web Services (AWS) data centers. These chips are scheduled for deployment between 2027 and 2028. This substantial order follows an earlier agreement made just five months prior, where Amazon committed to deploying over 1 million Nvidia GPUs across its AWS infrastructure starting this year, indicating that demand has rapidly outpaced initial expectations. While specific financial terms were not disclosed, the sheer volume of GPU units suggests this deal is worth tens of billions of dollars.

The scope of this expanded partnership extends beyond just GPU acquisition. Nvidia’s comprehensive technology stack, encompassing networking hardware vital for connecting thousands of GPUs into a unified system, its open models, CPUs, advanced data processing software, and its cutting-edge robotics platform, will be integrated across AWS. Both companies highlighted “surging demand” from a diverse range of clients—including startups, large enterprises, leading AI labs, and even governmental entities—as the primary catalyst for this intensified collaboration.

Notably, this close alliance with Nvidia is occurring concurrently with Amazon's considerable investments in its own AI chip development efforts. Amazon is actively developing proprietary chips, particularly CPUs (Graviton) and AI-specific processors (Trainium), with the strategic aim of reducing its reliance on and potentially competing with chip giants like Nvidia. Amazon’s AI chief, Peter DeSantis, has indicated that AWS is exploring opportunities to sell its Trainium chips—which serve as a direct alternative to Nvidia’s H100 or Blackwell chips for deep learning workloads—to other companies for deployment in their own data centers. Amazon’s custom chip business is reportedly experiencing robust growth, having crossed a $25 billion annualized revenue run rate, supported by substantial commitments totaling $225 billion from prominent AI labs such as Anthropic and OpenAI.

As part of the latest agreement, Nvidia also plans to supply an unspecified number of its Vera CPUs to AWS. Some of these Vera CPUs will be integrated with the Rubin GPUs, while others will be deployed as standalone units. Nvidia CEO Jensen Huang has expressed ambitious plans for the company’s Vera CPUs, projecting a new market opportunity (Total Addressable Market, or TAM) of $200 billion. Beyond AWS, Nvidia CFO Colette Kress affirmed that Vera CPUs are expected to be deployed by “every major hyperscaler, neocloud, AI lab, and system OEM,” with initial shipments already underway to key partners like Oracle and SpaceXAI.

The partnership’s reach further extends into Amazon’s vast logistics and enterprise sectors. Amazon intends to adopt Nvidia’s full physical AI stack to enhance its fleet of warehouse robots. This comprehensive stack includes Omniverse, Nvidia’s simulation and digital twin platform; Cosmos, its world model platform; Isaac, its robotics development platform; and Jetson, its computing hardware designed for robots and edge AI applications. Coinciding with this, Nvidia recently launched a new version of Jetson, designed to offer a more accessible robotics computer for “entry-level edge AI.” In the enterprise domain, AWS will host Nvidia’s Nemotron family of open models on Amazon Bedrock, its managed foundation model platform, and on SageMaker, its managed cloud service for machine learning.

In a related development, Nvidia reported impressive financial results, with sales reaching $96.2 billion for the second quarter, surpassing analyst expectations. Data center revenue constituted the vast majority of these sales, hitting $89 billion, an increase of 117% compared to the previous year. Nvidia projects its revenue to climb to $108 billion in the third quarter, partially driven by sales of its next-generation Rubin GPUs, for which production shipments commenced this quarter. Investors are closely monitoring Rubin’s early sales performance for indications of sustained demand for Nvidia’s future hardware generations. To meet anticipated AI demand over the coming years, Nvidia has significantly increased its commitment to secure supply and manufacturing capacity, allocating $279 billion for current and future data-center projects, a substantial increase from $119 billion last quarter. This commitment includes projected spending of $92 billion for the remainder of the current fiscal year and an additional $87 billion for fiscal year 2028.

During Nvidia’s earnings call, CEO Jensen Huang articulated his vision, stating that