Kate Gallego has had it.
In her 2025 State of the City address, Phoenix’s mayor called on lawmakers to eliminate Arizona’s special tax treatment for “new data centers.” Calling it “a holdover from a time before our economy was the magnet for job growth that it is today,” Gallego declared that to “ensure our precious resources are used to the highest benefit and value for our residents, state lawmakers must roll back this antiquated incentive.”
The chief executive of the nation’s fifth-largest city is uniquely qualified to speak on the subject. Few metro areas surpass Phoenix’s status as a data-center boomtown. But for now, at least, Gallego is an outlier. As American Reformer’s Aaron Renn noted: “Nothing makes a city sound like a tech hub like an announcement of billions of dollars of investment by blue chip Silicon Valley firms. Every mayor or governor is interested in news like that.”
And the announcements are frequent, given the rise of artificial intelligence. McKinsey & Co. estimates that globally, by 2030, data centers “equipped to handle AI processing loads are projected to require $5.2 trillion in capital expenditures.” The U.S. is claiming an enormous portion of that sum. In June, Construction Dive reported that “[i]nvestment has poured into creating new massive data centers in 2025”—the “sector accounted for more than 70% of the increase in [America’s] private nonresidential construction spending between March 2024 and March 2025.”
That’s a lot of groundbreaking ceremonies attended by members of the media who rarely ask pesky questions. Little wonder, then, that the law firm Husch Blackwell has documented 36 tax-incentive programs for data centers at the state level. The facilities are unquestionably popular in the glad-handing halls and hearing rooms of capitol complexes. But Gallego’s concern is just one example of the doubts mounting in the communities where data-center construction and operation occur.
Failure to deliver on headline-grabbing promises is a common complaint. As the Center for Economic Accountability’s John C. Mozena puts it, data centers “stand out for being especially abysmal tools for an economic-development strategy.” Not only are the facilities “unimpressive at job creation since they employ few people onsite,” they “offer limited benefits to surrounding communities,” because they are largely bereft of “visiting customers or vendors who might patronize local restaurants, gas stations, convenience stores, dry cleaners or other businesses that benefit from … spillover traffic.”
Adding to the disappointment, the industry’s government goodies (not merely tax breaks, but infrastructure improvements and industrial revenue bonds) sap public coffers, leaving others to make up the difference. In a recent analysis, Good Jobs First’s Kasia Tarczynska and Greg LeRoy concluded that “10 states already lose more than $100 million per year” to “the cloud-computing warehouses that were proliferating before artificial intelligence greatly accelerated their growth.” The states examined—including major players Virginia, Texas, Georgia, Washington, and Illinois—“typically exempt projects from paying sales and/or use taxes on their largest start-up and maintenance expenses.” And some states don’t even bother to “report foregone tax revenues” from data centers.
The hidden tax imposed by such largesse bites, but perhaps not as painfully as ratepayers’ bills will in the years and decades to come. Per The Wall Street Journal, “America’s power bills are rising even faster than the cost of groceries,” with the “average household power bill [reaching] $784 for the combined period from June to September,” an uptick of “4.2% from last year when adjusted for inflation.” Expect more hikes to come.
AI’s thirst for juice is ravenous. In December, the North American Electric Reliability Corp.’s John Moura warned of “a period of profound change,” driven by “demand growth like we haven’t seen in decades … and what we see is the pace only accelerating.” A December 2024 assessment by Lawrence Berkeley National Laboratory found that “data center load growth … is projected to double or triple by 2028.”
Slowly, state legislators are probing the power-deficit quandary, and crafting responses. California legislators have approved legislation ordering the state Public Utilities Commission to issue a report assessing “the extent to which electrical corporation costs associated with new loads from data centers result in cost shifts to other electrical corporation customers” by January 1, 2027. In June, Beaver State lawmakers approved the Protecting Oregonians with Energy Responsibility (POWER) Act. Sen. Deb Patterson, D-Salem, claims the legislation “helps protect everyday users, like families and small businesses, from paying the costs that big businesses are running up.”
Regardless of how the electricity bills are paid, urban air quality is another data-center complication. Cutting-edge nuclear reactors, assuming they are ever built, will not go online anytime soon. And while climate-conscious Silicon Valley adores wind and solar, “green” generation is intermittent, requiring expensive backup systems. That leaves emissions-spewing natural gas and coal, which pose difficulties for nonattainment metro regions such as Los Angeles, San Francisco, Las Vegas, Phoenix, Salt Lake City, Denver, Dallas, Houston and Chicago.
Similar tensions include neighbors’ gripes and noise and light pollution, as well as the loss of open space, farmland and wildlife habitat. Finally, water is a major data-center challenge, especially in the arid Southwest. According to InformationWeek, a U.S. Department of Energy report on Los Angeles County’s “data center energy use pegs total 2023 water use at 66 billion liters, up from 21.2 billion liters in 2014,” and “that’s just for direct consumption to cool data centers themselves—water needed to cool power plants supplying electricity to data centers, also adds to the total.”
AI has immense potential and of course cities should not impose barriers on construction of data centers. But technological progress is not without tradeoffs. Cities need to evolve beyond vacuous boosterism, and examine the unintended consequences of subsidizing the data-center explosion.
D. Dowd Muska is a researcher and writer who studies public policy from the limited-government perspective. A veteran of several think tanks, he writes a column and publishes other content at No Dowd About It.
