A Smarter Deal for the Data Center Boom

Data centers have a target on their backs. In blue and red states, cities and small towns, communities are pushing back. It’s not hard to see why. Residents worry that data centers will drive up electric bills, strain water supplies, create noise and other local impacts, while delivering too little in return. They have also become a symbol of the divide between the booming AI economy and everyone else. Arguments about national competitiveness will only go so far. But the AI boom is an opportunity states should negotiate, not fear. The challenge is capturing the upside without shifting costs and risks to consumers. The good news? Cities and states have the tools to do that. 

The goal should be a more reliable and affordable power system that benefits everyone. AI and cloud computing bring capital and demand to the power sector. Used well, that demand can finance generation, transmission, and grid modernization; create early markets for advanced nuclear, enhanced geothermal, storage, and other clean resources; and accelerate better chips and computing architectures.

The benefits can also show up in consumers’ electric bills. The affordability debate usually starts with the risk that data centers will raise rates. With the right policies, however, they can put downward pressure on them. A large customer that pays its full connection costs and stays for the long term can bring in more revenue than it costs to serve. That means more customers sharing the grid’s fixed costs and better use of infrastructure that is already in place. The gains can extend to local budgets too with sound tax policy: in Loudoun County, data centers generate 38 percent of General Fund revenue, helping fund schools and public services while allowing the county to lower homeowners’ property-tax rate every year for the past decade. But it depends on good policy: realistic, independently vetted load forecasts; long-term customer commitments; and rules that make data centers pay for the infrastructure built to serve them, including when a project never materializes. Otherwise, consumers can be left paying for power plants and infrastructure built for demand that never arrives. 

States have leverage for reasons that go beyond permits, speed, and certainty. Hyperscalers need reliable and affordable power, clear interconnection rules, and confidence that the grid can support their growth. Everyone benefits from a stronger grid and that shared interest creates an opening for a smarter deal.

States have started with large-load tariffs that keep the cost of serving major energy users from spilling onto other consumers. Developers must provide credible load forecasts, post collateral, and make firm commitments, including covering customer-specific infrastructure and the cost if they walk away. Done right, these tariffs protect consumers from stranded investments and give serious projects a faster path forward.

States should also ask large customers to help supply the capacity needed to serve them. Bring Your Own Generation is one option , but it is too narrow as a general rule. A data center might bring generation, storage, flexible demand, distributed resources, or some combination. What matters  is not which technology it uses. It is whether the capacity is real, available when the grid needs it, and backed by  enforceable commitments.  The same principle should apply to water: projects should account for local watershed conditions, pursue alternatives such as recycled water, closed loop systems, and air cooling where water risk is high, and be transparent about expected use.

States should also find ways to spread the benefits beyond the data center. A large customer could meet part of its capacity obligation by helping homes and businesses reduce demand during the year’s most expensive hours  Illinois already has a useful building block: under ComEd’s net-metering tariff, solar and storage can push a customer’s Peak Load Contribution below zero, turning a capacity charge into a credit. Illinois should use that structure  in its BYOC policy, and other states should create similar pathways.  

States should also encourage innovation. Large customers can help emerging energy technologies move from demonstration to scale, while data centers can shift workloads away from constrained hours, make better use of existing grid infrastructure, and reduce the energy needed for each unit of computing. States should reward those results rather than prescribe the technologies used to achieve them.

Moratoriums are a blunt answer to a problem that requires better policy. They do not produce power, modernize the grid, lower bills, resolve water concerns, or earn community trust. They freeze bottlenecks in place while sending investment elsewhere.

The test is whether people can see the bargain working: consumers are protected, water and land are managed responsibly, workers and communities benefit, and each project leaves the grid stronger and more flexible than it found it.

States should say yes, but strike a smarter deal. Data centers should pay their way, help build the power system they need, and produce benefits people can see in their communities and electric bills.

Get that bargain right, and the data-center boom will not just place more demand on the grid. It'll help finance a better one.