JLL Report: H1 2026 Data Center Demand Hits Record 25 GW
AI infrastructure demand drives record capacity absorption while public opposition exposes a 65-point acceptance gap.
North America's data center market experienced an unprecedented surge in demand during the first half of 2026, absorbing a record-breaking 25 GW of capacity. This staggering milestone, revealed in JLL's North America Data Center Report – Midyear 2026, highlights the fierce race among hyperscalers, neoclouds, and artificial intelligence firms to secure scarce computing infrastructure. However, the report also warns that accelerating growth is colliding with severe local bottlenecks, particularly around community acceptance of water, energy, and noise impacts.
Key details
According to JLL's Midyear 2026 analysis, the 25 GW of capacity absorbed across North America in the first six months of the year represents a doubling year-over-year and is five times the absorption level recorded just two years prior. This rapid buildout has occurred in an environment where vacancy has remained locked at a historical low of approximately 1% for the third consecutive year.
To satisfy this unrelenting demand, developers have pushed construction pipelines to new heights. The report notes there are currently 66 GW of capacity under construction across North America, 95% of which is already pre-committed to tenants. Geographically, Texas leads the market with 26 GW of existing and under-development capacity, followed by Virginia at 13 GW. This geographic concentration has forced a major structural shift, with energy-rich, build-friendly frontier markets now representing 77% of all capacity under development.
Why this matters
The JLL findings underscore that the AI revolution is fundamentally an infrastructure and resource scaling challenge. The sheer volume of power and cooling water required to support high-density AI clusters is shifting developers away from traditional tech hubs toward frontier markets that can accommodate massive energy requirements. As developers contract space as far out as 2028, the massive capital commitments—evidenced by $250 billion in AI-related bond issuance in the first half of 2026 alone—illustrate the colossal financial footprint backing this resource expansion.
Context
The report exposes a stark, defining paradox for the next phase of AI expansion: while 79% of Americans support U.S. leadership in artificial intelligence, only 14% support data center development in their own communities. This 65-point support gap highlights growing public anxiety over local resource consumption, grid strain, and noise pollution. As a result, developers are being forced to pivot toward greater transparency and early community engagement, particularly regarding how they source electricity and manage water use.
Risks and open questions
The primary risk highlighted by JLL is the severe bottleneck of community resistance. With a massive 65-point gap in public acceptance, developers face the risk of costly delays, moratoriums, and local regulatory pushback if resource concerns are not addressed. Furthermore, enterprise users requiring smaller deployments of 500 kW to 3 MW are struggling to secure any capacity at all, as hyperscalers and neoclouds dominate pre-commitments, potentially stifling broader economic integration of AI technologies.
What happens next
To sustain this growth, the data center industry must actively build trust with local communities. Developers are expected to adopt more transparent approaches to mitigate local impacts, such as partnering on water restoration, self-funding grid improvements, and utilizing cleaner power options. With rent growth averaging 9% annually since 2020 and expected to maintain momentum through 2030, landlords hold significant leverage, but actual delivery of the 66 GW under-construction pipeline will ultimately depend on navigating local utility and community bottlenecks.
Source: JLL Published on AI Usage Global, author: AUG Bot
