Nvidia Supplies 1 Million AI Chips to Amazon in Multi‑Year Cloud Deal

Nvidia’s commitment to ship a full million of its next‑generation GPUs to Amazon marks the most sizable single‑supplier deal ever announced for a public cloud provider, according to BusinessToday Malaysia. The agreement, described as a “multi‑year cloud deal,” bundles the chips with Nvidia’s broader AI software stack, including the company’s inference‑optimised libraries and its DGX cloud‑ready systems. By locking in such volume, Amazon can scale its AI‑as‑a‑service offerings—particularly generative‑model workloads that have surged in demand since the launch of large‑language‑model APIs—without the lead‑time and price volatility that typically accompany spot‑market GPU purchases.

The deal also dovetails with broader industry talks about a massive joint investment in AI infrastructure. Reuters reported that Nvidia, Microsoft and Amazon are in discussions to pour up to $60 billion into shared compute capacity, a figure that underscores how the three firms view cloud‑based AI as a strategic growth engine. While the $60 billion figure refers to a separate, still‑unfinalised investment framework, the Nvidia‑Amazon shipment is the first concrete component of that larger collaboration, giving Amazon a guaranteed supply of the “H100”‑class GPUs that power the most demanding transformer models.

From Nvidia’s perspective, the contract secures a critical revenue stream as the company pushes beyond the consumer graphics market into enterprise AI. The firm’s fiscal reports have shown GPU sales climbing at double‑digit rates, driven largely by data‑center demand, and a million‑chip order from Amazon would represent a multi‑hundred‑million‑dollar boost to its top line. Nvidia’s CEO Jensen Huang has repeatedly emphasized that scaling AI compute is the “new electricity” for the tech economy, and the Amazon partnership is a direct embodiment of that vision, delivering the hardware foundation needed for Amazon’s Bedrock and SageMaker services to compete with Microsoft’s Azure OpenAI offering.

Amazon, for its part, is positioning the influx of Nvidia hardware as a catalyst for new AI products aimed at both enterprise customers and developers. The cloud giant’s recent announcements have highlighted a suite of generative‑AI tools that rely on high‑throughput inference, and the guaranteed supply of Nvidia’s latest GPUs will allow Amazon to promise lower latency and higher throughput than competitors still dependent on mixed‑vendor fleets. Analysts cited by Reuters note that such hardware depth could translate into pricing power for Amazon’s AI services, especially as customers increasingly seek “pay‑as‑you‑go” compute that can handle the spikes typical of model training and fine‑tuning.

The partnership also hints at a shifting balance of power in the AI‑cloud ecosystem. While Google’s Tensor Processing Units have long dominated its own cloud, the Nvidia‑Amazon tie‑up signals that the industry’s de‑facto standard for high‑performance AI compute is consolidating around Nvidia’s GPU architecture. This alignment may accelerate the adoption of Nvidia‑centric software stacks, from CUDA to the newer Nvidia AI Enterprise suite, across Amazon’s vast customer base. As BusinessToday Malaysia notes, the scale of the deal “could set a new benchmark for cloud‑provider hardware procurement,” potentially prompting rivals to secure similar long‑term supply agreements to stay competitive in the rapidly evolving AI market.