AI is no longer a niche topic. It’s fueling everything from advanced logistics and predictive analytics to real-time medical imaging and autonomous systems. But this growth comes with a hidden cost: power. The computational needs of AI are enormous, and as a result, demand for data center power is rising faster than many regions can keep up with. The infrastructure required to support AI—from model training to inference—has created new challenges for commercial real estate.
For commercial property managers, particularly those overseeing buildings designed for tech tenants or high-performance computing, data center power limitations are not a distant issue. They’re a very real, immediate factor that can influence property value, tenant retention, leasing potential, and even regional economic activity.
This article breaks down why data center power is becoming the bottleneck for AI growth, what this means for your properties, and the proactive steps you can take to stay ahead of the curve.
AI’s Appetite for Power Is Surging
AI models, especially those in the large language model (LLM) category like GPT-4 and beyond, require massive computational resources. Training a model of that size can consume millions of kilowatt-hours of energy. Even running inference—producing outputs for user prompts—draws significant power. As AI adoption increases across industries, the physical footprint and electrical load of the compute infrastructure behind it will only grow.
According to a 2024 report from the International Energy Agency (IEA), global electricity consumption from data centers, AI, and cryptocurrency combined could double by 2026, reaching over 1,000 terawatt-hours (TWh) annually—roughly equivalent to the electricity demand of Japan.
This surge is due not only to the growth of AI, but also the nature of the chips powering it. Processors like the NVIDIA H100 require far more electricity than traditional CPUs. A single server rack loaded with GPUs can draw 10 to 50 times more power than a typical enterprise server rack.
Regions Are Already Hitting Power Limits
This is no longer hypothetical. Multiple U.S. markets are already seeing data center power limitations stall new development and strain the electrical grid.
For example, in Northern Virginia—home to the world’s largest concentration of data centers—new projects have been paused or delayed due to regional power shortages. Utility companies in the area have told developers it could take 5-8 years to connect large-scale data center projects to the grid because the infrastructure simply isn’t ready.
Phoenix and Santa Clara have experienced similar constraints. In both areas, the local utilities are struggling to keep up with surging demand driven by high-density data workloads. Developers are being forced to relocate or delay projects simply because they can’t secure power.
For property managers, especially those operating in power-constrained regions, this directly affects leasing timelines and the kinds of tenants your buildings can realistically support.
Data Center Power Is Now a Leasing Factor
In this environment, data center power capacity is no longer just an engineering or IT concern—it’s a commercial real estate issue. High-growth tenants like AI labs, biotech firms, video processing companies, and financial tech startups are now assessing power availability during site selection, right alongside square footage, fiber access, and rent.
If your building can’t support 1 to 3 megawatts of additional power—or if your region’s utility can’t approve new capacity within 12–18 months—you may be disqualified from the running before a lease is even negotiated.
How This Will Shift Commercial Real Estate Demand
The rise of AI and the resulting data center power crunch will create a new segmentation within commercial properties—those that are “AI-capable” and those that are not. The shift is already happening, and it will become more pronounced in the next two to three years. Here’s how:
1. Properties With Ample Power Will Command Premiums
Buildings that already have high power density, robust backup systems, and available load capacity will become highly valuable to compute-heavy tenants. These properties will likely command higher rents, longer leases, and stronger tenant demand.
2. Locations Near Substations or Transmission Lines Will Surge
Proximity to electrical substations or high-capacity transmission infrastructure will become a key site selection factor. In some cases, it may outweigh other traditional criteria like parking or walkability.
3. Retrofits Will Become More Common
To stay competitive, some properties will need to undergo substantial upgrades to their electrical systems, transformers, backup generators, and cooling infrastructure. This creates both an opportunity and a cost for property managers.
4. Low-Power Buildings Will Lose Tech Tenants
Properties that can’t support the growing power needs of modern businesses may lose current tenants or fail to attract new ones. Over time, this could lead to reduced lease rates or longer vacancy periods.
What Commercial Property Managers Should Do Now
If you manage commercial properties that cater to—or want to attract—tech tenants, cloud services, or AI firms, there are steps you should be taking now to prepare for a future shaped by data center power needs.
Conduct a Power Audit
Bring in an electrical engineer or MEP (mechanical, electrical, and plumbing) consultant to assess your building’s current power usage, available capacity, and expansion potential. Know your building’s limits before a tenant forces you to find out.
Open a Dialogue With Your Utility
Establish a relationship with your local power provider. Ask about load capacity, timelines for service upgrades, and any programs available to help commercial properties increase electrical infrastructure. Utilities are already overwhelmed—getting in early matters.
Explore Onsite Power Options
In some cases, it may be possible to add supplemental systems like solar arrays, battery storage, or even small-scale natural gas generators. While these won’t replace utility power, they can add redundancy or offset peak usage demands.
Evaluate Cooling Infrastructure
AI workloads don’t just draw power—they also generate heat. Make sure your HVAC systems can handle the load. Consider upgrading to support liquid cooling or expanded airflow requirements for high-performance computing tenants.
Start Marketing Power Capacity
If your building already has strong data center power infrastructure, don’t bury that in the spec sheet. Market it aggressively. Use it to attract the next wave of tenants that need more than just square footage—they need scalability.
Closing Thoughts
The growth of AI is exciting, but it also introduces real challenges for commercial property managers. Chief among them is data center power—a resource that is rapidly becoming the most important factor in high-tech leasing decisions.
Those who invest early in evaluating and upgrading their power infrastructure will be in a position to capitalize. Those who wait may find themselves managing buildings that no longer meet the expectations of tomorrow’s tenants.
For more insights into how technology is changing commercial property management, visit Property Manager Insider for the latest trends, analysis, and industry news.
