Table of Contents
In an interview with RCR Wireless News, Beatriz Valle, senior analyst at GlobalData, noted a widening of the gap between data center power needs and available supply
In sum – what to know:
Power is the main constraint – Grid capacity and connection delays are emerging as the biggest barriers to data center growth, outweighing capital availability or customer demand.
Designs are shifting fast – AI workloads are driving extreme rack densities and accelerating adoption of liquid and immersion cooling technologies.
Edge deployment is rising – Enterprise AI inference is increasing demand for localized, modular data centers closer to users and operations.
Power availability is set to become the most significant constraint on data center expansion in 2026, overtaking both access to capital and customer demand, according to Beatriz Valle, senior analyst, enterprise technology and services at GlobalData, in an interview with RCR Wireless News.
Valle said the rapid acceleration of AI workloads, cloud adoption, and broader digital transformation is driving electricity demand far faster than grid infrastructure can be upgraded. As a result, the gap between data center power needs and available supply continues to widen.
“The surge in AI workloads, cloud adoption, and digital transformation has intensified power demand at a rate far exceeding infrastructure upgrades,” Valle said, adding that power availability will “far outweigh” capital access or customer demand as a limiting factor through 2026.
Grid pressure is particularly visible in mature markets. Valle pointed to the U.S. as an example, where rising energy demand is placing strain on transmission networks and contributing to project delays. In parts of Europe, including Ireland, the Netherlands, and the U.K., restrictions on new data center developments have already been introduced due to concerns over grid capacity.
Beyond power itself, regulatory processes are becoming increasingly complex. Valle noted that permitting timelines are lengthening as sustainability standards tighten and community scrutiny grows. These factors are adding cost, uncertainty, and delays to new projects, particularly in congested markets.
Despite these challenges, investment appetite remains strong. The analyst said private capital continues to flow into both new builds and expansions, supported by sustained demand from cloud, hyperscale, enterprise, and AI customers. Financial constraints are unlikely to become a major issue in 2026 unless there is a significant macroeconomic or geopolitical shift.
At the same time, AI workloads are reshaping how and where data centers are built. Valle explained that rack densities are rising sharply, moving well beyond traditional 5–10kW levels toward 30kW, 50kW, and in some cases more than 100kW per rack. This trend is being driven by GPU-heavy architectures and the emergence of large-scale “AI factories.”
Cooling technology is also evolving rapidly. Valle said legacy air-cooled systems are no longer sufficient, with liquid cooling, immersion cooling, and rear-door heat exchangers becoming standard in high-density AI environments.
GlobalData estimates in its recent report, dubbed GlobalData’s 2026 Enterprise Predictions: Data Center, that there is currently a global pipeline of large-scale data center projects with a total value of $2.30 billion.
