AI’s Growing Share of Capacity
AI workloads are projected to increase from about 25% of total global capacity in 2025 to 50% by 2030, with a significant shift expected around 2027 when inference workloads are predicted to surpass training requirements. This shift is already impacting site selection and facility design, as environments focused on training demand higher power per rack and often command premium lease rates.
JLL’s research suggests that without a secure power source, the risk of financing a new data center build is too high. This emphasizes the critical role of power availability in the development of data center facilities.
Power Constraints Drive ‘Opportunistic’ Siting
Data center operators are shifting towards a “power opportunistic” approach for site selection, focusing on areas where power supply is available rather than traditional geographic preferences. This change has led to increased development activities in states like Wisconsin, Indiana, Louisiana, and Pennsylvania, driven by extended grid connection timelines in primary markets.
Design Adapts to Rising Density
As data center floor density rises, operators are reevaluating how power is distributed within cabinets and high-voltage setups. The traditional Basis of Design is now more dynamic, requiring teams to accommodate changing processor densities throughout the facility’s construction process. This flexibility in design is a significant shift for the industry.
Regional Outlook: Americas Lead
The Americas are expected to maintain their position as the largest and fastest-growing region for data center capacity expansion, with the US leading the way. Asia-Pacific and Europe, the Middle East, and Africa are also projected to see substantial growth in data center capacity, driven by increasing demand from hyperscalers and digital transformation initiatives.
Build Timelines and Equipment Supply
Global equipment lead times have extended to an average of 33 weeks, causing delays in more than half of the projects initiated in 2025. To mitigate these delays, developers are focusing on securing power sources early on and integrating design, sourcing, and delivery processes more tightly.
Financing Matures Around the Asset Class
Capital strategies in the data center industry are evolving to support the scale of expansion. Core investment strategies, asset-backed structures, and commercial mortgage-backed securities are becoming more prevalent, reflecting the growing institutional interest in the sector. This trend is driven by the increasing demand for AI technologies and greater institutional involvement in data center investments.
