Amazon AWS CEO: The focus of AI applications is shifting from model training to inference, with strong and sustained enterprise demand
I'm LongbridgeAI, I can summarize articles.Amazon AWS CEO Matt Garman stated that as AI models mature, the focus of enterprises is shifting from training to inference deployment, driving rapid growth in AI inference demand. Driven by this, Amazon has raised its capital expenditure forecast for 2026 to $220 billion and signed long-term agreements to secure data center construction needs, expressing strong confidence in the long-term prospects of its AI business
According to Zhitong Finance APP, Matt Garman, CEO of Amazon's (AMZN.US) cloud computing business AWS, stated that as artificial intelligence (AI) models become more mature, enterprise customers are gradually shifting their focus from training large AI models to deploying these models in practical business scenarios, which is continuously driving rapid growth in AI inference demand.
In an interview on Monday, Garman said, "We still see many companies using large training clusters for model training, but as these models become more mature and capable, more and more enterprises are beginning to integrate AI inference into their business workloads."
He explained that model training refers to the process of building artificial intelligence models using a large amount of computing power, while inference is the use of a trained model to respond to user requests and execute practical applications. As AI gradually enters the stage of commercialization, inference demand is becoming an important force driving the growth of cloud computing resources.
As the world's largest cloud service provider, Amazon announced last week that it has raised its capital expenditure forecast for 2026 from the previous $200 billion to $220 billion. The company stated that the increase in capital expenditure mainly reflects the rising costs of memory chips and other key hardware needed for data center construction.
Regarding next year's capital expenditure plans, Garman did not make specific predictions but emphasized that the future market size for AI business is "extremely large," and the company is confident about the long-term demand outlook.
He stated that AWS is currently securing demand for future data center construction by signing long-term cooperation agreements with customers. "As we continue to expand our investments, we are signing five-year long-term agreements with customers," Garman said. "Currently, market demand still far exceeds supply, and we are continuously building and investing to meet the growing needs of our customers."
Amazon's previously announced second-quarter financial report showed that AWS revenue grew by 37% year-on-year, marking the fastest growth rate since the end of 2021 and the fifth consecutive quarter of accelerating growth. Driven by the better-than-expected performance of its cloud computing business, the company's stock price has recently continued to rise, and its market value has surpassed $3 trillion, making it the fifth publicly traded company in the world to reach this milestone
