Data center developers riding the AI demand wave are discovering that strong market fundamentals do not automatically protect individual projects from rising construction costs, according to JLL’s Global Construction Costs Index 2026.
The real estate services firm found that annual construction cost escalation of 3% to 6% has become the new baseline, up from the 2% to 3% range common before 2020. JLL is also forecasting another 3.1% to 4.2% increase in global tender prices in 2027, including 3.2% to 4.2% in North America. Those forecasts apply to construction broadly, not just the data center sector.
Rising rents and asset values have helped many data center projects absorb some of that inflation, but Matthew Giddings, executive director of technical services with JLL’s Data Centers and Critical Environments team, cautioned that market strength does not guarantee any single project will pencil out.
“The economics break down when costs outpace contracted revenue or power delays postpone income while financing costs continue to accumulate,” Giddings said. “Higher asset values alone cannot close that gap.”
Where the Risk Concentrates
According to JLL, the biggest exposure sits with speculative campuses that lack anchor tenants, projects facing uncertain power delivery timelines, and developments locked into fixed-rent agreements that leave no room to absorb cost escalation. Older facilities being repositioned for AI workloads can also run into trouble if the required upgrades turn out to be more extensive than expected.
Giddings recommended developers revisit a project’s viability when capital costs erode expected returns, when utility delivery slips past committed occupancy dates, or when tenant demand no longer supports the planned capacity. Phased construction, where spending is tied to confirmed power supply and customer commitments, can limit how much capital is exposed before those fundamentals are locked in.
Labor and Procurement Bottlenecks
Skilled trades are compounding the pressure. Giddings pointed to electricians, high-voltage specialists, pipefitters, and advanced cooling technicians as the biggest sources of cost and schedule risk, along with controls, commissioning, and startup teams whose availability often determines when a finished facility can actually go into service.
To manage that, JLL suggests bringing trade partners in early, testing schedules against real labor availability, and leaning on standardized designs, off-site prefabrication, and phased builds to cut peak labor demand. In remote markets, that may mean offering housing and transportation incentives to attract workers.
On the procurement side, long lead times have pushed developers to order transformers, switchgear, generators, and cooling equipment earlier, but Giddings cautioned against freezing every design decision too soon given how quickly AI infrastructure requirements can shift. Reserving manufacturing capacity while retaining flexibility to adjust specifications, he said, protects both the schedule and the asset’s long-term value.
Retrofitting existing buildings can offer a faster, cheaper path to delivery when power, structural capacity, and infrastructure can be reused, but those advantages evaporate quickly if a site needs major mechanical upgrades or large-scale liquid cooling retrofits.