Amazon CEO Andy Jassy revealed that AWS’s growth has been hampered by capacity constraints, particularly in meeting the surging demand for AI infrastructure. During the company’s Q4 earnings call, Jassy explained that AWS couldn’t fulfill all customer requests for AI training and inference capacity, specifically for GPU clusters needed to run large language models and other AI applications. Despite these limitations, AWS still achieved $24.2 billion in revenue for Q4 2023. Jassy emphasized that Amazon is actively working to secure additional capacity, with significant GPU deliveries expected throughout 2024 and into 2025. He projected that AWS will have substantially more AI infrastructure capacity by the end of 2024. The CEO also highlighted that the company is seeing strong customer interest in its AI offerings, including Amazon Bedrock and its custom AI chips like Trainium and Inferentia. The capacity constraints have led Amazon to prioritize existing customers’ AI workloads over new customers, though Jassy assured that the company is working to accommodate all demand. This situation reflects the broader industry challenge of meeting the explosive growth in AI computing requirements, with major cloud providers competing for limited GPU resources, particularly those from NVIDIA. The company expects AI-related services to be a major growth driver for AWS once capacity issues are resolved.
Source: https://www.businessinsider.com/amazon-ceo-cloud-capacity-issues-affected-growth-ai-2025-2