House Energy and Commerce Chair Brett Guthrie (R-Ky.) has been one of the AI buildout's most reliable backers in Congress. According to Politico's coverage, published June 16, 2026, local officials across his deep-red Kentucky district are now pushing back hard on the data centers that buildout requires. Breckinridge and Daviess counties have enacted one-year moratoriums on data center applications and construction. Meade County rejected a proposal after roughly 3,000 residents signed an opposition petition. Hancock County is still advancing a TeraWulf project at a former aluminum smelter site that could pull in $14 billion in investment and come online by 2027.
The political pressure tracks a national footprint. Politico counts about 1,500 data centers planned or under construction across 232 congressional districts, split nearly evenly between Democratic and Republican territory. In Kentucky specifically, the numbers explain the alarm. LG&E and KU report 11 projects totaling roughly 3.5 GW with a better-than-even chance of advancing, and a total pipeline that could reach 12 GW. East Kentucky Power Cooperative has 11 active projects seeking more than 10 GW. Kentucky utilities generated a maximum of 18.4 GW across the entire state in summer 2024, so the data center queue alone rivals existing peak supply.
That gap is the cooling story in disguise. A 100-megawatt facility draws roughly the continuous average demand of 80,000 homes, and a large share of that draw is heat that has to be moved off the racks. The planned Louisville hyperscale campus started at 335 MW and expanded to 402 MW with an $11 billion lifetime budget. At that density the thermal architecture is no longer optional. Air handling cannot reject heat from racks running modern accelerators, which is why operators are standardizing on direct-to-chip liquid loops. The grid fight and the water-power tradeoff operators keep getting wrong are the same problem viewed from two ends of the facility.
Guthrie's legislative answer is the draft Ratepayer Protection Act, which would amend the Public Utility Regulatory Policies Act of 1978 to make large electricity users bear the cost of the infrastructure built to serve them rather than spreading it across residential bills. He frames it as a balance: do nothing and a community loses out on "the industry of the future." The cost concern is concrete. Virginia projections cited in Kentucky's own ratepayer report show potential residential bill increases of $14 to $37 per month by 2040, with data center growth a contributing driver. The same dynamic is fueling the grid constraints now blocking projects in Washington and elsewhere.
For cooling vendors and developers, the implication is direct. When the operator owns the marginal cost of new generation and transmission, every watt of facility overhead becomes a line item with a clear owner. That sharpens the case for cutting mechanical load: higher coolant supply temperatures that eliminate chillers, hot-water loops, and warm-water rejection that lets a site run closer to ambient. The same logic raises the stakes on water draw, since evaporative cooling trades electricity for gallons and county boards are now scrutinizing both. Moratoriums in Breckinridge, Daviess, and Meade counties are, functionally, pauses on heat rejection capacity. The projects that clear those reviews will be the ones whose cooling design answers the ratepayer question before it is asked, a pattern visible in the early zero-water cooling pilots launching in Phoenix and Mount Pleasant.
Guthrie sits at the center of federal AI energy policy and cannot get a single county in his district to wave projects through without a fight over cost allocation. That is the buildout's pinch point in miniature. Power availability and ratepayer politics now gate site selection as tightly as fiber or land once did, and cooling architecture is the lever operators control to make a project pencil out. Liquid cooling, chiller elimination, and closed-loop water designs stop being efficiency talking points and become permitting strategy.