Amazon Web Services (AWS) has nearly exhausted its cloud computing capacity for 2027, with much of its 2028 capacity already booked as demand for AI infrastructure continues to rise.
Amazon CEO Andy Jassy says the company is struggling to keep up with requests from AI research laboratories and enterprise customers despite plans for massive infrastructure investments.
Speaking on a video podcast with Andreessen Horowitz co-founder Ben Horowitz and general partner Erik Torenberg on October 23, Jassy said AWS had already sold out its major capacity allocations for 2026, while most of its 2027 capacity was also committed. AWS has also reserved a substantial amount of capacity for 2028. Read the full report from Reuters.
The growing demand reflects the rapid expansion of AI services, which require enormous computing resources to train models, process information and respond to user requests. AI developers and businesses are competing for access to advanced processors, servers and data center infrastructure as they build and deploy increasingly sophisticated systems.
Amazon has announced plans to spend approximately $220 billion in capital expenditure in 2026, reflecting the scale of its investment in cloud computing, artificial intelligence and related infrastructure. However, Jassy said the spending would not be enough to satisfy all the demand for AWS capacity this year. He added that Amazon could justify investing even more if customer demand continues to grow.
The capacity shortage does not mean every AWS service or computing instance is unavailable. Jassy explained that the company continues to reserve some resources for on-demand customers and services such as Amazon Bedrock and Amazon Elastic Compute Cloud (EC2). Bedrock provides access to AI models for building generative AI applications, while EC2 offers computing capacity for running applications and workloads.
Amazon is also working to improve the economics of AI computing through its own semiconductor development efforts. Its Trainium chips are designed for AI workloads, while Graviton processors target general-purpose cloud computing. These in-house technologies help Amazon compete on performance and cost while reducing some of its dependence on external chip suppliers.
Jassy also highlighted the importance of making AI inference more affordable. Inference refers to the computing required when an AI model responds to a prompt or performs a task after training. Lowering these costs could help businesses deploy AI more widely and increase demand for cloud infrastructure even further.
The company faces several practical challenges as it tries to expand capacity. Building data centers requires suitable locations, reliable electricity supplies, networking equipment and large quantities of specialized hardware. Even with substantial financial resources, these components cannot always be delivered and deployed quickly enough to meet immediate demand.
The pressure on AWS highlights how AI is changing the cloud computing industry. Large research laboratories need substantial computing capacity for advanced models, while businesses are increasingly adopting AI for customer service, fraud detection, software development and workflow automation. As more applications move from experimentation into everyday use, demand for computing resources could continue to rise.
Amazon’s comments also suggest that major customers are committing to infrastructure well in advance to secure the capacity they need. Smaller businesses and developers may still be able to access AWS services, but availability can vary by region, instance type and service. Customers planning large deployments may need to evaluate their capacity requirements earlier and consider alternative configurations or cloud providers.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
Amazon was scheduled to report its third-quarter financial results on October 29, 2026. Investors are expected to watch for further details about AWS growth, infrastructure spending and the company’s ability to turn rising AI demand into revenue. The results could provide additional insight into how quickly Amazon can expand capacity and whether its substantial investments will be sufficient to meet customer needs.
Amazon AWS Capacity Nearly Sold Out Through 2027 as AI Demand Surges
