February 20, 2026
Brookings has a new prescription for easing public resistance to data centers: negotiate community benefit agreements, or CBAs, that spell out what companies owe local communities in exchange for building AI infrastructure. The proposal calls for legally binding, transparent contracts that define mitigation measures, workforce commitments, and enforceable accountability mechanisms. That framework is laid out clearly in Brookings’ January 2026 report, “Why community benefit agreements are necessary for data centers”.
On the surface, the pitch sounds reasonable. Spell out the impacts. Promise transparency. Lock in commitments. Reduce friction.
In reality, this approach points American AI infrastructure in the wrong direction.
Data centers are not ordinary development projects. They are the backbone of AI, cloud computing, and the modern digital economy. Even mainstream reporting now describes digital infrastructure as foundational to economic competitiveness and national growth.
Scale matters. When infrastructure is foundational, governance must be coherent.
The Brookings model centers on localized, project-by-project negotiation. Community benefit agreements are presented as collaborative tools to address concerns about electricity rates, water consumption, noise, tax incentives, and corporate power. But once these agreements become the expectation, they stop being optional tools and start functioning as gatekeeping mechanisms.
Recent data center fights make that risk obvious. In Prince William County, Virginia, a massive “Digital Gateway” project was blocked by a state judge after intense political and legal conflict.
Across the country, resistance has hardened into moratoria, zoning fights, and litigation. WIRED recently described the growing “data center resistance” movement as a defining political flashpoint in multiple states.
Layering formalized, legally enforceable CBAs on top of already contentious permitting fights risks institutionalizing that friction. What begins as mitigation quickly becomes negotiation leverage. A zoning change depends on concessions. A permit requires a new package of commitments. Each locality defines its own standard.
That is how a quasi regulatory regime forms without ever being debated as regulation.
States and local regulators are not bystanders in this process. They make real decisions about grid planning, generation siting, and how infrastructure fits into their communities. That’s by design. Under federal law, states oversee retail electricity rates and power plant siting, while the Federal Energy Regulatory Commission handles interstate transmission and wholesale markets. It’s a shared system, and that division of responsibility matters.
That shared system works best when everyone knows the rules going in. States and local regulators handle the real, on-the-ground decisions. That’s appropriate. But the rules shouldn’t change every time a new project shows up. When every permit turns into a fresh round of bargaining, it doesn’t strengthen oversight. It just creates uncertainty.
Electricity cost allocation illustrates the broader point. Public anxiety about rising power bills is real. But resolving that issue through dozens of localized contracts is not governance. It is improvisation.
When large new data center loads connect to the grid, questions about stranded costs and ratepayer exposure should not be settled in ad hoc bargaining sessions. They should be addressed within a coherent framework that sets clear guardrails and applies them consistently.
In Ohio, for example, the Public Utilities Commission approved a tariff structure requiring large new data center customers to pay for at least 85 percent of their subscribed electricity usage for up to 12 years, regardless of actual consumption.
That debate took place in a formal regulatory venue and resulted in a system-level rule rather than a project-specific bargain. The merits of any individual tariff can be debated. The structural lesson is clearer: cost allocation questions belong in transparent rulemaking processes, not embedded inside one-off development contracts.
The same principle applies to tax incentives and economic development policy. Debates over subsidies, abatements, and workforce projections should not be negotiated anew for every individual project. They should be governed by clear, nationally coherent standards that provide certainty to communities and investors alike.
Project-by-project bargaining encourages escalation. Each locality competes to extract more concessions, each negotiation introduces new conditions, and over time the cumulative effect is unpredictability. Infrastructure at national scale cannot operate under a constantly shifting set of local demands.
The deeper risk is fragmentation. The United States is already grappling with the consequences of inconsistent AI policy across jurisdictions. Executive branch analysis has emphasized the need for a coherent national policy framework and warned against burdensome, fragmented approaches that stifle innovation.
Industry leaders have echoed similar concerns about patchwork governance creating compliance risk and slowing investment.
Extending that fragmentation into infrastructure through localized community benefit agreements moves in the opposite direction of clarity. AI infrastructure is national in consequence. The policy framework governing it should be national in design.
Community concerns deserve serious treatment. Electricity rates, water usage, noise mitigation, and environmental monitoring are not trivial matters. But those issues belong within a coherent national framework that sets consistent guardrails, implemented through transparent and predictable regulatory processes.
Brookings envisions enforceable local contracts governing AI infrastructure. The alternative is simpler and stronger: nationally consistent guardrails, standardized transparency, and the confidence to build AI infrastructure at scale without renegotiating the rules in every jurisdiction.
AI capacity determines where research clusters form, where startups scale, and where advanced manufacturing anchors. Slowing deployment through regulatory layering carries real economic consequences. Leadership in AI requires building at scale. Scale requires predictability. Predictability requires coherence.
Turning every data center into a negotiated exception risks entrenching the very patchwork that policymakers elsewhere are trying to unwind.
Innovation does not accelerate through localized bargaining sessions. In a world where authoritarian competitors are racing to define the future of AI, America cannot afford to slow its own infrastructure buildout. Leadership requires building.
Jay Burstein is a fellow with Build American AI.
.png)
