A recent survey of data center professionals highlights significant obstacles in scaling AI data center capacities, primarily due to supply chain limitations and insufficient power availability. The Turner & Townsend 2025-2026 Data Center Construction Cost Index, built from insights into over 300 projects in 20+ countries by 280 experts, identifies critical hindrances in meeting the demand for AI infrastructure. Nearly half of the respondents express that grid access delays are a significant impediment, often extending to years. In the U.S., new grid requests face wait times up to seven years, a challenge echoed in the UK where developers encounter costly substation upgrades.

With AI infrastructures planned to consume monumental power, such as projects by OpenAI expected to use 55.2 gigawatts, the strain on the grid is intensified as data centers compete against housing and manufacturing sectors. This consumption could power millions of homes, exacerbating grid allocation conflicts across the Western hemisphere.

Deloitte foresees AI data center energy demands to balloon thirtyfold within a decade, while 5 GW facilities are already in the pipeline. Turner & Townsend advise on-site power generation and energy storage solutions, as alternative sources are imperative for AI facilities. Though renewable energy is suggested, practicality leans towards gas-driven turbines for immediate needs.

Beyond energy, 83% of industry professionals believe the current local supply chains fall short of providing the sophisticated cooling technology critical for high-density AI projects. Turner & Townsend’s analysis indicates a noticeable cost difference in AI-optimized liquid-cooled systems compared to their air-cooled counterparts, attributed to their higher initial design and construction costs.

Data center operators are encouraged to rethink their procurement strategies to strengthen and sustain supply chain efficiency, ensuring timely AI infrastructure delivery. Strategic adjustments are essential as AI data facilities escalate in both priority and complexity, posing greater energy needs and requiring more advanced cooling solutions.