Amazon’s AWS is on a global expansion spree, aggressively building out its server infrastructure and vying for a lion’s share of NVIDIA’s latest AI chips, according to remarks from the cloud computing giant’s leadership.
Speaking to reporters last Friday, AWS CEO Matt Garman highlighted the recent opening of a data center cluster in Mexico, alongside ongoing construction in Chile, New Zealand, and Saudi Arabia, underscoring the company’s commitment to meeting surging demand.
The move comes as Amazon anticipates its AI business to generate billions of dollars in revenue annually. Garman confirmed this significant figure “comes from AWS” representing usage of its on-demand AI services by customers.
As the industry’s leading provider of computing power and data storage services, AWS maintains a commanding lead over rivals Microsoft Azure and Google Cloud. In its most recent financial report released earlier this month, Amazon reported AWS revenue of $29.267 billion for the first quarter, a 17% year-over-year increase.
However, this growth rate lags behind the performance of its competitors. Microsoft’s Azure reported a 21% revenue increase, while Google Cloud saw a 28% jump in its top line.
Mirroring strategies of both Microsoft and Google, Amazon is accelerating investment in its AI computational capabilities. Garman noted AWS’s collaboration with NVIDIA to secure a substantial supply of the chipmaker’s next-generation GB200 chips, already available to AWS customers for testing.
NVIDIA’s supply chain partners, including those companies that build the machines, have reportedly overcome the technical hurdles in producing GB200 AI server racks, leading to a sharp increase in shipments.
Each GB200 AI rack integrates 36 Grace CPUs and 72 Blackwell GPUs, all interconnected through NVIDIA’s NVLink communication system.
“Demand is very strong,” Garman stated. He has been working up the ladder at AWS and became the CEO replacing Adam Selipsky a year ago.
Garman also shared his willingness to see OpenAI’s models deployed on AWS, even though they currently operate on Microsoft’s infrastructure.
While earlier this year, OpenAI sought approval to explore further partnerships, on the condition that Microsoft was not willing to provide the service.
“We encourage all of our partners to work across platforms,” Garman said. “I would love to see other companies take that approach.”
He added that he had no issues with Amazon partner Anthropic’s AI model Claude being available on Microsoft’s cloud platform.
“It’s great if Claude can be deployed elsewhere, but we see the vast majority of its usage on AWS,” Garman said.
Original article, Author: Tobias. If you wish to reprint this article, please indicate the source:https://aicnbc.com/1365.html
