Amazon Web Services CEO Matt Garman said customers are shifting from using the company’s infrastructure to train AI models toward embedding those models into their own business processes, a change that is driving higher demand for inference computing. Speaking on Monday, Garman said Amazon still sees some companies relying on large training clusters, but wider adoption and stronger model capabilities are pushing more businesses to bring inference workloads into production. He described the potential size of Amazon’s AI business as “very large” and said the company will keep raising capital spending to meet demand. Amazon, the world’s largest provider of rented computing capacity and data services, said last week that it now expects capital expenditures to reach $220 billion in 2026, up from an earlier forecast of $200 billion. The increase reflects higher prices for storage chips and other components needed for data centers.
Amazon (AMZN.O) cloud chief Matt Garman said customers are beginning to move from using the company’s services to train AI models toward integrating those models into their own business processes, a shift that is increasing demand for inference computing.
Inference workloads are taking a larger share
Speaking on Monday, Garman, the chief executive of Amazon’s cloud computing unit, said: “We still see some companies using large training clusters, but as these models become more popular and more capable, more companies are integrating that inference capability into their own workloads.”
He said the potential for the company’s artificial intelligence business is “very large,” adding that Amazon will continue to increase capital expenditures to meet rising demand.
Amazon raised its 2026 capex outlook
As the world’s largest provider of rented computing capacity and data services, Amazon said last week that it expects capital expenditures to reach $220 billion in 2026, above its previous forecast of $200 billion.
The higher spending reflects rising prices for storage chips and other components required for data centers.
This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan. Disclaimer:
The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.
Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.
