On April 27, the Wisconsin Public Service Commission spent more than six hours debating a question that utility regulators across the country have been quietly circling for two years. When a data center consuming as much electricity as a mid-sized American city connects to the grid, who pays for the new power plants, substations, and transmission infrastructure required to serve it? The utility’s existing customers, who had no say in the decision to attract the facility? Or the data center itself, which captures the economic benefit of the infrastructure investment? Wisconsin chose the data center. The decision is being described by commissioners, regulators, and industry analysts as one of the first of its kind in the Midwest, and the implications extend well beyond the state’s borders.

The Public Service Commission approved a new Very Large Customer tariff for We Energies, the utility subsidiary of WEC Energy Group that serves eastern Wisconsin. Under the revised structure, any data center or similarly large industrial customer drawing at least 100 megawatts must pay 100% of the generation and grid infrastructure costs required to serve it. The commission lowered the eligibility threshold from We Energies’ original proposal of 500 megawatts to 100 megawatts, immediately widening the scope to capture a much broader class of AI data center developments. Additionally, it extended the minimum contract term from 10 to 15 years, rejected a capacity-only payment option that would have allowed data centers to cover just 75% of generation costs, and required new reporting mechanisms to monitor how the tariff works in practice.

Why This Decision Is Different From What Came Before

Utilities have historically provided electricity to data centers under the same large industrial customer frameworks used for manufacturers, cold storage facilities, and other high-consumption businesses. They designed these frameworks for a system where the largest industrial customers drew tens of megawatts under relatively stable load profiles. By contrast, an AI data center drawing 500 megawatts to 1 gigawatt under dynamic GPU workload patterns represents a fundamentally different type of grid customer. It requires infrastructure investments at a scale and on a timeline that existing frameworks cannot accommodate, and regulators and utilities have not yet resolved who should pay for that infrastructure in most US markets.

Wisconsin’s decision matters because it answers that question clearly and unambiguously. Commissioner Kristy Nieto stated during the hearing that data center customers must pay their own way, fully and transparently. PSC Chairperson Summer Strand framed the goal as increasing transparency and visibility while ensuring existing customers are protected. Furthermore, the commission’s rejection of We Energies’ original 75% payment proposal over the objections of the utility itself signals that regulators are willing to impose stricter cost allocation than industry prefers when the public interest argument for doing so is sufficiently clear. That regulatory posture is new. It reflects a shift in how utility commissions across the country are beginning to think about data center load as a special category requiring special treatment rather than simply very large conventional industrial demand.

What the 35-State Picture Looks Like

The Smart Electric Power Alliance, which has been tracking data center utility policies across the US, confirmed that at least 35 other states have established or are actively developing new utility rate structures for data centers. Wisconsin is therefore not acting in isolation. It is acting at a specific moment when the regulatory landscape is in transition, and its decision will inform how other state commissions approach the same question. Specifically, the elements of the Wisconsin decision that other regulators are likely to study most closely are the 100-megawatt threshold, the 100% cost allocation requirement, the 15-year minimum term, and the reporting mandate.

Each of these elements represents a policy choice that other commissions must make in their own proceedings. A state that sets its threshold at 500 megawatts, as We Energies originally proposed, effectively exempts most data centers from dedicated cost allocation frameworks. A state that sets it at 100 megawatts captures the majority of commercially significant AI data center developments. The threshold decision is therefore not a technical question but a political one about how much the state is willing to prioritise data center attraction over ratepayer protection. As covered in our analysis of the time-to-power crisis as AI’s hidden scaling ceiling, the grid infrastructure costs of AI data center development are substantial and growing, and the question of who bears those costs is becoming one of the most politically consequential issues in US energy regulation.

What the Data Center Industry Says

The Wisconsin Data Center Coalition has pushed back on the narrative that data centers are automatically a burden on ratepayers. A study it commissioned from Charles River Associates found that rate increases across the US are driven by localised factors specific to northeastern and California markets rather than by data center development broadly. The coalition argues that well-designed large load tariffs, precisely like the one Wisconsin just approved, can significantly reduce or eliminate cost impacts on other customers by requiring data centers to fund their own infrastructure needs. From the industry’s perspective, the Wisconsin decision is therefore not a defeat but a workable framework that provides regulatory certainty in exchange for cost transparency.

That framing is plausible but incomplete. The Wisconsin decision protects existing ratepayers from the direct cost of new generation and grid infrastructure, but it does not resolve the transmission cost allocation question. The commission acknowledged this gap and directed We Energies to address it in a revised tariff. It also leaves open the broader issue of whether tariff structures adequately reflect the environmental and community costs of data center development in specific localities. By requiring new reporting mechanisms, the commission signals that the full impact of data center costs will take time to become clear. As a result, the Wisconsin framework represents a first step in an evolving regulatory conversation rather than a final resolution of the data center ratepayer question.