Amazon will add two million additional Nvidia GPUs to its cloud between 2027 and 2028, both companies announced on August 26 during the chip maker's quarterly earnings presentation. The order builds on the one million processors that Amazon Web Services agreed to deploy in March and signals that compute demand for artificial intelligence continues to accelerate.
The partnership expands the agreement announced in March and goes beyond chip purchases: the networking technology connecting thousands of GPUs, Nvidia's open models, and its robotics platform will also be integrated into Amazon's infrastructure, according to Nvidia's press release. The company cited growing demand from startups, enterprises, AI labs, and governments as the driver behind the arrangement. The scale of the commitment, which will bring a combined total of three million GPUs to the AWS cloud, is a relevant reference point for Mexico and Latin America, where data center construction has become a thermometer for global competition over artificial intelligence.
Nvidia reported revenue of 96,200 million dollars in the quarter ending in June, with 89,000 million from its data center business, up 117% year over year, and projected sales of 108,000 million dollars for the current quarter. The additional chips include the Blackwell Ultra, Rubin, and Rubin Ultra models, which will arrive at Amazon Web Services data centers in 2027 and 2028, according to Bloomberg. The chip maker has committed 279,000 million dollars to secure production capacity and supply, and the agreement includes AI factories for the U.S. government with 100,000 GPUs in AWS secure infrastructure. All of this is unfolding as Amazon develops its own chips, a business that surpasses 25,000 million dollars annualized, according to TechCrunch, though the partnership with Nvidia remains in place.
The deployment of the new chips is scheduled for 2027 and 2028, and Nvidia has already begun production shipments of its Rubin line this quarter. The trend confirms that compute demand continues to define investment decisions among major cloud providers, a pattern that will determine where the next generation of data centers gets built.
This article was drafted with artificial intelligence assistance from verified sources and reviewed by a human editor before publication.