Dominion Energy says its upcoming $11-a-month rate hike is driven by grid upgrades and inflation — not Virginia’s rapidly growing data-center industry.
Critics note that Virginia hosts one of the nation’s largest concentrations of data centers, which consume massive amounts of electricity and have raised public concern about grid strain.
Modern data centers — especially those built for AI — can use power equivalent to tens of thousands of homes, prompting ongoing efforts to improve efficiency, cooling, and reliance on cleaner energy.
The Virginia State Corporation Commission has approved a utility rate increase for Dominion Energy customers that will add about $11 to the average monthly electric bill starting in January.
Critics were quick to point out that Virginia has one of the highest concentrations of data centers in the country, but Dominin said that had little effect on the rate. Instead, it said grid upgrades and inflation are responsible for the increase, not data centers – which have been a hot topic of conversation lately.
If you aren’t exactly clear about what data centers do, you probably aren’t alone. A data center is a specialized building full of computers (servers), storage systems, and network equipment that store, process, and move digital information for things like websites, cloud storage, streaming, and AI. It’s essentially the “physical internet” behind online services.
Data centers, by their very nature, increase electricity demand. They use a lot of electricity mainly because:
They run thousands to tens of thousands of powerful servers at once, often 24 hours a day, so the core computing load is huge.
All those servers generate a lot of heat, so large cooling systems (air conditioning, fans, or liquid cooling) must run constantly to keep the equipment from overheating.
They need extra infrastructure that also draws power: backup batteries and generators, power conversion equipment, networking gear, security systems, and building systems.
Modern AI and other “high‑performance computing” tasks are especially energy‑intensive, so newer data centers built for AI can use as much power as a small city.
What’s typical?
A “typical” modern data center uses on the order of a few to a few dozen megawatts of power continuously, which is like tens of thousands of homes running at once. Operators are trying to make them more efficient by improving cooling, tightening how power is delivered and measured (PUE), using smarter software, and shifting to renewable energy.
How much energy is “typical”?
Large “hyperscale” facilities can easily use 20–100+ MW, and some planned AI-oriented campuses are being designed for hundreds of megawatts or more. Small commercial data centers with a few hundred to a couple thousand servers usually draw about 1–5 megawatts (MW) of power. That is roughly comparable to the average usage of several thousand U.S. homes, depending on local consumption
Across the whole country, U.S. data centers together used about 183 terawatt-hours (TWh) of electricity in 2024, a bit over 4% of all U.S. power use, and that share is expected to more than double by 2030.
Key efficiency metric (PUE)
The main yardstick for efficiency is Power Usage Effectiveness (PUE), which is total facility power divided by the power used just by the IT equipment (servers, storage, networking).
A perfect data center would have a PUE of 1.0; many older facilities were around 2.0 or worse, while efficient modern and high‑performance centers try to get close to 1.2 or below, meaning most power goes into actual computing rather than overhead like cooling and lighting.
What’s being done to cut energy use?
Better cooling: Operators are switching from basic air conditioning to more efficient methods such as hot/cold aisle containment, liquid cooling directly to chips, and using outside air or water (“free cooling”) when climate allows, which can cut cooling energy by 10–50%.
Smarter operations: Software and AI are used to balance workloads, power down idle servers, and tune cooling in real time so the data center does not run all systems at full power unnecessarily.
Cleaner power and design: Many large data center operators sign long‑term contracts for wind or solar, build on‑site generation or storage, and design buildings, power distribution, and equipment layouts specifically to reduce losses and improve efficiency.