What is the connection between AI data center power demand and your electricity costs? AI data center electricity costs refer to the upward pressure that artificial intelligence computing places on regional power grids, wholesale electricity prices, and ultimately the utility bills paid by commercial real estate owners and tenants. As hyperscalers including OpenAI, Anthropic, Google, and Microsoft race to train and serve models such as ChatGPT, Claude, and Gemini, the megawatts required have started to reshape the operating cost structure of nearly every property type. For the broader landscape, see our guide to AI commercial real estate.
Key Takeaways
- AI data center power demand is pushing wholesale and retail electricity prices higher in several grid regions, raising operating expenses for commercial properties that have nothing to do with AI.
- Electricity is a controllable line item in NOI, so even a 10 to 20 percent rate increase can compress net operating income and pressure valuations at today's cap rates.
- Capacity prices in PJM, the largest U.S. grid, jumped more than tenfold for the 2026 to 2027 delivery year, with data centers cited as a major driver.
- Multifamily, office, retail, and industrial owners in strained grids face the most near term exposure to AI driven utility cost inflation.
- Submetering, energy benchmarking, and AI powered building management systems are the fastest levers CRE owners have to protect margins.
AI Data Center Electricity Costs Explained
The artificial intelligence boom runs on electricity. Training a frontier model and then serving billions of queries through tools like ChatGPT, Claude, and Gemini requires dense clusters of NVIDIA accelerators that draw enormous and continuous power. Lawrence Berkeley National Laboratory estimates that U.S. data center electricity demand will grow from about 176 terawatt hours in 2023, roughly 4.4 percent of national consumption, to between 325 and 580 terawatt hours by 2028. The U.S. Energy Information Administration forecasts residential electricity prices will rise about 4 percent on average in 2026 after climbing roughly 5 percent in 2025. For context on the asset class itself, see our analysis of the CBRE data center boom.
Here is the part many CRE investors miss. When a gigawatt scale data center campus connects to a regional grid, it does not only affect that one developer. It raises demand across the shared transmission system. In the PJM Interconnection, which serves the Mid Atlantic, capacity auction prices for the 2026 to 2027 delivery year rose to 329.17 dollars per megawatt day, more than ten times the 28.92 dollars from the 2024 to 2025 year, with rapid data center growth identified as a major contributing factor. Those costs flow downstream to every ratepayer on the grid, including the multifamily, office, and industrial properties in your portfolio.
How Rising Electricity Costs Hit NOI
Net operating income (NOI) is gross revenue minus operating expenses, and it does not include debt service, capital expenditures, or income taxes. Electricity sits squarely inside operating expenses, which makes it one of the levers that directly moves NOI. Consider a 200 unit multifamily property where common area and house electricity runs roughly 180,000 dollars per year. If grid stress pushes utility rates up 15 percent, that is an extra 27,000 dollars in annual operating expense. At a 6 percent cap rate, losing 27,000 dollars of NOI implies roughly 450,000 dollars of lost value, because value equals NOI divided by cap rate.
The math is unforgiving in commercial properties where tenants do not reimburse utilities. Office landlords on full service gross leases absorb electricity directly. Retail centers with common area maintenance structures can pass some costs through, but only up to what the lease and the market will bear. Industrial and logistics facilities are increasingly power hungry themselves as they add robotics and automated material handling. For leveraged assets, a higher expense base also lowers the debt service coverage ratio (DSCR), which is NOI divided by annual debt service and the first number a lender watches.
Which Markets Face the Most Exposure
Geography matters. The grid regions absorbing the largest data center buildouts are seeing the most pronounced rate pressure. Northern Virginia, often called Data Center Alley, sits inside PJM and has become a focal point, with the watchdog Monitoring Analytics attributing roughly 23 billion dollars of capacity costs to data centers. ERCOT in Texas, MISO in the Midwest, and pockets of the Southeast are also experiencing concentrated demand growth. CRE owners with assets in these regions should model utility inflation as a base case, not a tail risk.
This dynamic intersects with the land and resource competition we examined in data centers versus housing land competition, and with the power infrastructure deals reshaping utilities, such as the merger we covered in the NextEra and Dominion power story. It is worth noting the debate is not settled: a Charles River Associates study for the Edison Electric Institute argued data centers were not the sole cause of rate increases in many regions, while market design and policy also play a role. Either way, the cost trend for owners in affected grids is up.
What CRE Owners Can Do About It
- Benchmark and submeter: Use ENERGY STAR Portfolio Manager and unit level submetering to isolate where consumption and cost are growing.
- Deploy AI building management: AI powered building management systems can cut HVAC and lighting energy use by optimizing in real time. This is one of the clearest cases where AI helps offset an AI driven cost.
- Lock in supply: In deregulated markets, fixed rate electricity supply contracts can hedge against volatile capacity charges.
- Reprice leases: At renewal, revisit expense pass through structures so rising utility costs are shared appropriately with tenants.
- Add on site generation: Solar plus storage can reduce grid dependence and improve resilience.
For investors who want a structured plan to deploy these tools and protect NOI, The AI Consulting Network specializes in exactly this kind of operational AI implementation. The same data forces raising electricity costs are also reshaping who controls real estate information, a trend we cover in our companion article on CoStar's $800M Zonda acquisition.
The Underwriting Implication
Forward looking underwriting should treat utility inflation as a structural assumption in markets exposed to heavy data center growth. A pro forma that holds operating expenses flat while the local grid absorbs gigawatts of new AI load is understating risk. Stress test your NOI against a 10 to 20 percent multi year utility increase, check the effect on DSCR, and confirm the resulting cap rate and valuation still support your basis. You can review authoritative grid and consumption data through the U.S. Energy Information Administration and track CRE specific implications through CBRE Insights. If you are ready to build that capability into your portfolio, connect with Avi Hacker, J.D. at The AI Consulting Network.
Frequently Asked Questions
Q: Why are AI data centers raising electricity costs for other properties?
A: Data centers connect to shared regional grids and add large, continuous demand. In capacity markets like PJM, that demand drives up auction prices, and those costs are passed to all ratepayers, including commercial properties unrelated to AI.
Q: How much can rising electricity costs affect property value?
A: Because value equals NOI divided by cap rate, even a modest expense increase compounds. A 27,000 dollar annual NOI loss at a 6 percent cap rate translates to roughly 450,000 dollars of lost value.
Q: Which CRE markets are most exposed to AI driven utility inflation?
A: Regions with concentrated data center growth face the most pressure, including Northern Virginia and the broader PJM footprint, ERCOT in Texas, MISO in the Midwest, and parts of the Southeast.
Q: Can AI also help reduce these energy costs?
A: Yes. AI powered building management systems optimize HVAC, lighting, and equipment in real time, often reducing energy consumption enough to partially offset rising utility rates.