The global data center sector is expected to add almost 100 GW between 2026 and 2030 amid an “infrastructure investment supercycle,” according to JLL’s 2026 global data center outlook released this week.
The growth will equate to $1.2 trillion in global real estate asset value creation, with most of that in the Americas, which represents about 50% of global capacity, the report says.
Faced with development delays, market leaders are primarily focused on speed to power, community support and latency, among other priorities, according to JLL. Cost is less of a concern.
“Power, not location or cost, will be the primary site selection criteria due to multiyear wait times for a grid connection,” the report says.
Data center rents in the Americas are projected to increase at 7% CAGR through 2030, fueled by “aggressive hyperscale expansion and severe supply bottlenecks from power constraints,” JLL says. The primary factor containing rent growth is developer competition to sign new tenants, the report says.
Given multiyear wait times for grid connection, data center operators are expected to increase behind-the-meter power arrangements, colocated battery storage and natural gas to ensure they have temporary bridge power as well as permanent on-site power generation, the report says.
Due to softening price points for battery energy storage systems, or BESS, operators are increasingly leveraging these systems to firm up power availability and better handle load spikes.
Increasing AI use, which is the main driver of load spikes, still only represents a quarter of data center workloads, the report says. Training the AI models is driving most of that demand.
Looking ahead, there will be a shift to “inference workloads,” the report says. The term refers to revenue-generating uses of the technology when the AI, after specific training, is applied to projects or problem solving.
“In 2026, data centers will shift from simply powering AI to becoming ‘AI factories,’ infrastructure built to continuously train, fine-tune, and infer at scale while generating intelligence as a core output,” Steve Carlini, vice president of data centers and innovation at Schneider Electric, said in an email. “This evolution will accelerate the rise of AI-driven robotics and autonomous systems, whose complex workloads will require high-density, low-latency facilities designed around new performance thresholds.”
From an operational standpoint, as AI chips become more powerful and costly, average rack density will triple to 45 kW, with 80% liquid cooling adoption for new facilities, JLL predicts.
AI semiconductor chips are almost 10-times the price of traditional CPUs. That leads to an increase in shell and core construction and tech fit-out costs. In 2025, AI infrastructure already costs as much $30 million per megawatt, the report says.
JLL estimates $1-2 trillion in spending from tenants to fit out their space with GPUs and networking by 2030.
From a development standpoint, the sector continues to consolidate as the cost of the facilities, coupled with increasing sophistication required to build and operate them, weeds out smaller companies.
Facility design is also undergoing fundamental changes, in part because of increasing rack densities, which are approaching 100 kW and need liquid cooling. To get these designs right, developers are turning more to digital twin technology.
“Digital twins will become standard, allowing operators to validate 240 kW-plus rack densities and fully simulate thermal behavior, energy loads, and system stress ahead of physical deployment,” Carlini said. “And as liquid cooling, renewable-energy strategies, and adaptive retrofits go mainstream, resilience and flexibility will define the next era of digital infrastructure.”
