The increased for data centers have significantly increased in the last year as the use of artificial intelligence becomes more mainstream. According to a report by JLL, in the next four years nearly 100 GW of new data centers will be added, doubling global capacity.
Currently, the Americas represent the largest data center region, accumulating about 50 percent of global capacity, with the United States accounting for 90 percent of the capacity. The region also has the fasted growth, with a project 17 percent supply CAGR to 2030, the JLL report found.
Even with AI becoming more mainstream, it still only represented a quarter of all data center workloads last year. According to the report, training drove most of the demand, however, a shift is expected in 2027 when inference workloads could overtake training as the dominant AI requirement.
The report explains that once an AI model is created, inference generates continuous revenue through application usage. Every AI model deployment creates demand that will grow through user adoption. However, the rapid implementation of inference application don’t exist at the scale of which they are needed. This will require geographical distribution to reduce latency and serve users effectively, according to the report.
The increase in data center construction is resulting in extended lead times and increasing costs. According to the report, the average global data center construction cost increased 7 percent CAGR to $10.7 million per MW, JLL reports.
Mackenna Moralez is the associate editor for the facilities market and the host of the Facilities in Focus podcast.
