The driver is data centers — huge, power-hungry facilities that process and store digital information. A report released earlier this year by the International Energy Agency found that data centers will account for nearly half of electricity demand growth in the U.S. between now and 2030. By then, the country will use more electricity for data centers than it does for producing aluminum, steel, cement, chemicals and other energy-intensive materials combined.

Zooming in from that national perspective, we can also look at the energy cost of something familiar. Asking a chatbot like ChatGPT for an answer uses roughly ten times the electricity of a typical Google search. And it adds up. OpenAI CEO Sam Altman recently said that all those polite prompts, users adding “please” and “thank you”, cost the company tens of millions of dollars in electricity. In fact, a single day’s worth of electricity to run ChatGPT equals the power used by 180,000 U.S. households.

As AI tools are built into more apps and services, the energy demands from data centers will continue to grow. Here in Utah, we’re seeing a rapid buildout of these facilities. In West Jordan, a 100-acre data center complex is underway with a price tag of $2 billion. When it’s completed, it will consume 175 megawatts of electricity. It’s already fully leased, a sign of just how much demand exists for the computing power AI requires.

Other developments are following. BluSky AI Inc., for instance, has secured land and power deals for modular data centers across the state. These projects combine grid electricity and renewable sources to meet their energy needs. All of this adds up to a lot more electricity demand in a relatively short amount of time.

So what does that mean for business and homeowners? Will regular customers get stuck with higher electric bills as tech companies race to build AI infrastructure?

Earlier this year, the Utah Legislature passed a law designed to prevent that. Senate Bill 132 allows large data centers, which it defines as those needing at least 100 megawatts within five years, to find their own power if a utility like Rocky Mountain Power can’t meet their demand without making infrastructure upgrades. That includes building their own power plants or signing contracts with independent providers. These centers can also buy power through the existing utility grid, but they have to pay their full share of transmission costs.

The bill was crafted with the goal of protecting households and small businesses from rate hikes tied to these large new energy users. It requires strict separation of costs. If a data center needs a new substation or transmission line, it pays for it, not the public. The law directs utilities to keep separate accounts for these large load customers, making it harder for costs to be hidden or folded into general rates.

The Utah Public Service Commission oversees the process, reviewing each case to ensure the rules are followed. Utilities get 90 days to review a large energy request and assess whether they can meet the need without significant new investment. If not, the company must seek another option and absorb the cost of doing so.

That’s the theory. In practice, it will take time to see how well the protections work as more data centers come online. One risk is that a sharp increase in demand could raise wholesale power prices, if that demand exceeds new power facilities. Another potential challenge is that some data centers may choose to generate their own electricity and operate entirely off the grid. That can make demand harder to predict, complicating future infrastructure planning for utilities. But even in those cases, the law is structured to ensure that utilities aren’t left on the hook for facilities they don’t serve.

What sets Utah apart is the clarity in who pays for what. If a data center needs a new substation or line, it pays for it. If it wants to tap into renewable sources, it negotiates the contract. And if it needs to go entirely private, it builds its own system. These costs are not spread to residential and small business customers.

That transparency and separation of costs are what make Utah’s approach notable. It doesn’t attempt to slow AI expansion but ensures that the responsibility for powering it stays with those who benefit most. As AI infrastructure continues to grow, Utah’s experience could become a model for other states navigating similar issues.

Leslie Meredith has been writing about technology for more than a decade. As a mom of four, value, usefulness and online safety take priority. Have a question? Email Leslie at asklesliemeredith@gmail.com.