Amazon just tripled its order of Nvidia chips over ‘surging demand’
Amazon is adding another 2 million Nvidia GPU chips to its data centers over the next two years. But this extended partnerships stretches beyond buying more chips.
Amazon and Nvidia have just strengthened their partnership by agreeing to add 2 million Nvidia GPUs to Amazon's data centers by 2027 and 2028. This deal comes in the wake of Amazon's earlier commitment to deploy over 1 million Nvidia GPUs across AWS infrastructure. The expansion was announced during Nvidia's quarterly earnings call, just five months after Amazon's initial agreement.
The companies did not disclose specific financial terms, but considering GPU unit costs, the deal is expected to be worth tens of billions of dollars. While the partnership primarily involves Amazon purchasing more Nvidia chips, Nvidia's technology, including networking hardware, CPUs, data processing software, and robotics platform, will also be integrated across AWS.
This collaboration is significant due to its scale, rapid growth, and the fact that it extends beyond Amazon simply buying more Nvidia chips. It also occurs as Amazon continues investing in its own AI chips, particularly CPUs. Amazon's AI chief has noted that AWS is exploring selling its Trainium chips to other companies for use in data centers, while Amazon's custom chip business has been expanding, generating $25 billion in annualized revenue.
Nvidia CEO Jensen Huang envisions a $200 billion potential market for the company's Vera CPUs. The expanded partnership will also impact Amazon's warehouse robots and enterprise offerings, as Nvidia's technology will be integrated into Amazon's Omniverse, Cosmos, Isaac, and Jetson platforms. Nvidia reported $96.2 billion in sales for the second quarter, with data center revenue accounting for the majority at $89 billion, a 117% increase from the previous year.
The company expects revenue to reach $108 billion in the third quarter, with some of that coming from the new Rubin GPUs. Nvidia has committed $279 billion to secure supply and manufacturing capacity for current and future data-center projects.
Written by urgent.news from TechCrunch's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- Hot Chips 2026: Nvidia touts benefits of its DSX MaxLPS site power management approach — tech allows for more compute from fixed data center power budgets tomshardware.com
- AWS plans to add 2M Nvidia Blackwell Ultra, Rubin, and Rubin Ultra GPUs to its data center fleet in 2027 and 2028, in addition to 1M GPUs announced in March (Bloomberg) bloomberg.com
- AMI orders 9,000 Nvidia Vera Rubin chips economictimes.indiatimes.com