Story

AWS AI Demand Drives Record Growth, Capacity Sold Out Into 2028, CEO Says

ENTHMSVIIDZHZH-TWJAKOHI
Aug 3, 20262 min read
AWS AI Demand Drives Record Growth, Capacity Sold Out Into 2028, CEO Says
Text size
Background

Amazon Web Services (AWS) is recording its most rapid growth in almost five years, fueled by intense demand for artificial intelligence workloads that has booked much of its compute capacity well into 2028, CEO Matt Garman said in an interview with Bloomberg TV.

Unprecedented Demand and Capacity Constraints

Garman stated that the surge in demand is broad-based, extending beyond major AI research labs to include traditional enterprise clients in sectors like financial services, healthcare, and retail. This widespread adoption is driving a significant supply-demand imbalance.

To manage the high demand, AWS is seeing customers lock in their future needs with long-term contracts. According to Garman, much of the company's capacity is already committed through the end of 2027 and into 2028, with clients signing five-year agreements to secure access to computing power.

AI Business Hits $25 Billion Run Rate

AWS's AI business is now operating at a $25 billion revenue run rate, a figure that reflects a significant shift in how customers are using AI. Garman highlighted a move from model *training* to *inference*—the process of running pre-trained models to power real-world applications.

Sample IUX Markets – In-articleAd

This trend indicates that companies are increasingly deploying AI to generate direct business value. The growth is distributed across a wide customer base, which Garman contrasted with competitors who may rely more heavily on a few large clients.

Strategic Investments in Custom Silicon

To support this growth and offer cost-effective solutions, AWS is heavily investing in its custom silicon, including its Trainium and Graviton chips. Demand for its Trainium-powered capacity is particularly strong, with much of it sold out through the end of next year.

Garman noted that by controlling the full hardware and software stack, AWS can offer customers 20% to 30% in cost savings on inference compared to other market options. Despite the success of its in-house chips, AWS remains one of Nvidia's largest customers and will continue to offer Nvidia GPUs to provide clients with a range of hardware choices. This investment is part of Amazon's overall capital expenditure, which stands at $220 billion this year, and Garman confirmed that heavy capital spending will continue into next year to build out the necessary infrastructure.

Read next

More on Stocks
Back to latest news

LATEST