Commentary: The United States is entering a new phase of the artificial-intelligence race, and the central question is no longer whether data centers will be built. It is who will pay for the electricity, transmission and environmental costs they impose.
That question deserves more than a culture-war answer. Communities are right to demand a say over enormous facilities that consume land and power. Technology companies are also right that delays in connecting new capacity could weaken American competitiveness, raise dependence on foreign infrastructure and slow useful innovation. The policy challenge is to reconcile those claims without quietly transferring private costs to the public.
The political backlash is already visible. A recent POLITICO poll found that half of Americans oppose a data center being built within three miles of their home. New York Gov. Kathy Hochul has approved a one-year statewide moratorium on new data-center projects, while Republican voices, including Michigan Senate candidate Mike Rogers, have also supported pausing construction.
That bipartisan unease should not be dismissed as opposition to technology. Residents see the possibility of higher utility bills, pressure on water supplies, noise, traffic and a landscape reshaped for facilities that may employ relatively few people once construction ends. Those concerns become sharper when companies receive tax incentives while households absorb the cost of grid upgrades.
The case for building
There is a serious argument on the other side. Data centers support cloud computing, scientific research, digital services and the rapidly expanding AI industry. Restricting them indiscriminately could push investment to countries with weaker environmental safeguards or leave the United States dependent on foreign computing capacity. Policymakers also face genuine electricity-demand growth, and postponing infrastructure does not make that demand disappear.
A bipartisan permitting proposal discussed in Congress would accelerate federal approval of major energy and infrastructure projects. It also includes provisions requiring data centers to pay for grid upgrades, according to POLITICO. That is a promising direction: faster approvals paired with clearer responsibility.
Yet speed alone is not reform. If permitting becomes a mechanism for approving projects before regulators can assess water use, emissions or local reliability, the result will be faster mistakes. Nor should every facility be treated identically. A data center powered by new, low-carbon generation and located where transmission capacity exists presents a different public burden from one that requires expensive grid expansion or prolongs dependence on fossil fuels.
Make the bargain explicit
Our position is straightforward: build the infrastructure America needs, but require the beneficiaries to pay their fair share and prove that projects serve a public interest beyond corporate growth.
- Data centers should fund the marginal grid and transmission costs they create, with transparent state oversight.
- Utilities should publish how new projects affect rates, reliability, water consumption and emissions.
- Permitting deadlines should be shorter, but environmental review and meaningful community participation should remain intact.
- Tax incentives should be conditional on measurable jobs, local investment, efficiency standards and clean-power procurement.
- States should coordinate siting so that facilities are directed toward regions with available power and suitable infrastructure.
Critics will say these requirements could slow investment or place American firms at a disadvantage. That risk is real. A fragmented maze of local bans and unpredictable rules would be worse than a national framework with clear standards. But the answer to uncertainty is not an open-ended subsidy. It is a stable set of rules that companies can price into their plans.
The debate also exposes a larger weakness in American energy policy. Lawmakers want more generation, faster permitting and global leadership, but they have repeatedly avoided deciding how to allocate costs. A recent analysis from Harvard Business School’s Business and Government Innovation Solutions program notes that affordability, competitiveness and national security are increasingly shaping the energy debate, while bipartisan permitting efforts remain difficult to enact. Those priorities should be connected, not traded against one another.
AI may transform the economy. It should not transform ordinary electricity customers into involuntary venture capitalists. A durable computing boom requires public consent, and public consent depends on an honest bargain: companies may build, communities may scrutinize, and ratepayers should not be left holding the bill.