The Economics of Data Centers in the AI Era: A Survey of Cost Structure, Power Constraints, Market Organization, and Financial Risk

This review examines the economics of modern data centers as AI demand reshapes what they cost to build, where they can operate, and how they are financed. I focus on five connected issues: capital costs, operating costs, power and grid constraints, market structure, and financing. The central argument is that AI infrastructure no longer behaves like a conventional real estate-and-utilities business. Its economics increasingly depend on the useful life of expensive chips, access to power that may take years to secure, and financing structures that can hide risk behind long-term contracts. The final sections offer a practical framework for judging whether a data-center investment is resilient or simply riding a capacity boom.