On August 4, the head of Amazon’s cloud business disclosed that customers are transitioning from leveraging Amazon’s services for AI model training to incorporating these models into their business operations—a shift that is fueling a surge in demand for inference computing. Matt Garman, CEO of Amazon Web Services (AWS), stated on Monday that while some companies continue to rely on large-scale training clusters, the growing popularity and sophistication of AI models are prompting more businesses to integrate inference capabilities into their workflows. He emphasized that the AI sector holds immense potential, and Amazon is committed to scaling up its capital investments to meet this rising demand. As the world’s leading provider of computing power and data service rentals, Amazon announced last week that its capital expenditures are expected to hit $220 billion by 2026, up from the prior estimate of $200 billion. This increase reflects the escalating costs of components like storage chips, which are essential for data center operations.