Spanish
Amazon.com's AWS cloud computing unit announced on Tuesday that it will adopt Nvidia technology in future generations of its artificial intelligence chips, as the company intensifies its efforts to get major AI customers to use its services.
AWS, or Amazon Web Services, said it will adopt a technology called "NVLink Fusion" in a future chip, with no specified release date, known as Trainium4. NVLink technology creates fast connections between different types of chips and is one of Nvidia's crown jewels.
The companies made the announcement at the AWS cloud computing annual conference, which is held for a week in Las Vegas and attracts about 60,000 people. Amazon is also expected to showcase new versions of its Nova AI model, initially presented last year.
Nvidia has been pushing for other chip companies to adopt its NVLink technology, with Intel, Qualcomm, and now AWS joining.
Technology will help AWS build larger AI servers that can recognize and communicate with each other more quickly, a critical factor in training large AI models, in which thousands of machines must be chained together.
As part of the association with Nvidia, customers will have access to what AWS calls AI Factories, an exclusive AI infrastructure within their own data centers for greater speed and preparedness.
"Together, Nvidia and AWS are creating the computing fabric for the industrial AI revolution, bringing advanced AI to every company, in every country, and accelerating the world's path to intelligence," said Nvidia CEO, Jensen Huang, in a statement.
Furthermore, Amazon announced the launch of new servers based on a chip called Trainium3. The new servers, available on Tuesday, contain 144 chips each and have more than four times the computing power of the previous generation of AWS AI, while consuming 40% less energy, according to Dave Brown, vice president of computing and machine learning services at AWS.
Brown did not give absolute figures on power or performance, but said that AWS intends to compete with its rivals —including Nvidia— based on price. "We have to show them that we have a product that gives them the performance they need and get a correct price point so they get that price-performance benefit," Brown said. "That means they can say, "Eh, yes, that's the chip I want to go use.""
