Data center capacity constraints in North America have driven net new demand to 25 gigawatts in the first half of 2026, while vacancy rates remain stuck at 1% for the third consecutive year, according to JLL’s 2026 North America Data Center Midyear report released Tuesday.
Grid Constraints and Stabilized Asset Protection
According to Morningstar, development pipelines are facing disruptions due to difficulties in securing and provisioning new data center capacity, leading to a greater reliance on operational, stabilized facilities.
In an August 5 commentary, the firm indicated that because current facilities already possess secured capacity under contract, they remain largely shielded from new interconnection moratoriums and grid limitations.
“Once operational, data centers are substantially protected from grid constraints and new interconnection moratoriums through binding utility agreements that fix capacity rights, allocate infrastructure risk, and defined service priority,” Morningstar noted.
Operators are capitalizing on this dynamic by modernizing older footprints. Facilities that are 10, 15, or 20 years old still retain functional power, fiber, and mechanical plants with years of useful life remaining, according to Sean Farney, vice president of data center strategy at JLL. Hyperscale, colocation, and enterprise organizations are actively updating these aging sites to support rising artificial intelligence computing requirements, Farney told Facilities Dive.
These established clusters also benefit from a self-reinforcing concentration effect. Operators seek close proximity to dense fiber routes, cloud availability zones, carrier hotels, and deep pools of skilled labor situated in major metropolitan areas, Morningstar reported.
Power Scarcity and Emergency Curtailment Risks
Despite strong contractual protections, existing facilities are not entirely immune to ongoing power scarcity issues. This insulation is not absolute, according to Morningstar, because utility companies and system operators retain ultimate control over physical power delivery, transmission, and grid balancing.
As electricity demand from data centers outpaces infrastructure expansion, regulators are increasingly introducing curtailment provisions, demand response obligations, and load-shedding mechanisms. These frameworks can force operators to reduce or temporarily disconnect power usage during grid emergencies.
A January decision by the U.S. Department of Energy granted authority to PJM Interconnection to redirect electricity away from large industrial factories and data centers toward residential homes during severe cold weather events to stop blackouts, as highlighted by Morningstar. Additionally, rules enacted in Texas require data centers with loads exceeding 75 megawatts to participate in mandatory demand management programs.
Consequently, Morningstar classifies stabilized data centers as contractually secured but systemically contingent assets. They remain protected from routine grid constraints under existing agreements yet exposed to tail-risk scenarios involving emergency curtailment or structural electricity market changes.
Surging Rents and Enterprise On-Premise Retrofits
Tight supply dynamics are driving significant financial gains for landlords. Data center rents have increased approximately 9% year over year since 2020, according to JLL. Larger deployments exceeding 20 megawatts now achieve average rents of roughly $141 per kilowatt-hour, excluding energy costs, reflecting a clear market premium for high-density, hyperscaler-ready capacity, Morningstar reported.
Landlords are capturing average rent increases of up to 70% on lease renewals since 2020, JLL’s midyear report found. Most new leases carry annual escalations of 3% or more with zero concessions.
Artificial intelligence drives more than 50% of new data center capacity, but even energy-abundant markets face commercial roadblocks if local substations are full or transmission upgrades are backlogged, according to Morningstar. To bypass these delays, enterprise users managing hybrid portfolios are increasingly retrofitting on-premise data centers that were emptied over the past decade as workloads migrated to the cloud, Farney explained.
“Some enterprises are beginning to retrofit their on-premise data centers, which were largely made empty over the last 10 years, moving compute loads to the cloud,” Farney said. “This means they have some stranded power, which is perfect for building out a test and AI environment.”
These on-site projects reduce long-term spending compared to relying entirely on external cloud services for AI inference. They also eliminate latency bottlenecks for internal staff, according to Farney.
Did You Know? Enterprise site requirements for hybrid infrastructure typically hover between 500 kilowatts and 3 megawatts, a much lower threshold than massive hyperscale deployments, yet these smaller projects still face severe equipment and capacity bottlenecks.
Frequently Asked Questions
Why are vacancy rates remaining at 1% despite heavy construction?
Net new demand across North America doubled to 25 gigawatts in the first half of 2026, outpacing unprecedented construction activity and keeping vacancy flat at 1% for the third consecutive year, according to JLL.
Are existing data centers completely safe from power shutoffs?
How are enterprises solving capacity shortages on their own?
Many enterprises are retrofitting previously emptied on-premise data centers to utilize stranded power for internal AI inference testing labs, reducing latency and avoiding high long-term cloud computing costs, according to JLL and Sean Farney.
Join the Conversation: How is your organization addressing data center capacity limits and AI infrastructure bottlenecks? Share your strategy or questions in the comments below.
Worth a look