Although industry expansion is, of course, excellent news for the sector, there are several factors at play slowing its progression. As the power required to operate AI-enabled data centers grows, so too does the strain on aging electrical infrastructure, which was never designed to, nor is it capable of handling, the kind of fluctuating electrical loads posed by the kinds of applications we’re seeing today.

Upgrading is the obvious answer to this dilemma, but with development periods for a typical greenfield data center around two to three years, and the time needed for necessary infrastructure upgrades spanning four to eight, the math doesn’t match the reality. This leaves data center developers contending with challenges around getting access to the power required, with lengthy wait times for interconnection and bottlenecks in power infrastructure making solving the power puzzle all the more complex.

Compounding the issue, supply chain constraints, a lack of skilled labor, and a renewed onus on the sustainability of these facilities via ever-evolving rules and regulations, make constructing a data center on time and to budget seem nigh on impossible. In this survey report, we examine how priorities in data center construction have changed as a result of the AI era, exploring the barriers to development, as well as some of the sentiments and solutions surrounding this newfound data center landscape.