Scioto Analysis polled 14 Ohio economists about the state's data center incentive structure and the proposed statewide construction ban. The Ohio Capital Journal reported the findings: 10 of 14 agreed that tax incentives for data centers are not an efficient use of public funds. 8 of 14 also agreed that the economic costs of a statewide ban would outweigh the benefits.
The two findings sit together cleanly even though they read as contradictory on first pass. The economists are saying that public money should not go into subsidizing facilities that produce few permanent jobs and substantial externalities, but that explicit prohibition is the wrong remedy when the underlying issue is unpriced cost. The position they converge on is the same position cooling industry economists have been quietly making for two years: let data center operators pay the actual cost of the resources they consume, and the buildout will rebalance itself.
Two of the externalities Ohio economists named most often were power costs imposed on residential rate payers and water consumption from cooling systems. Albert Sumell of Youngstown State described data centers as one of the worst possible uses of public funds because of the combination of low job creation and high external costs. Michael Jones of the University of Cincinnati framed the right response as forcing data centers to internalize and pay for the energy and land use they consume.
That framing is where the cooling industry sits in the middle of the policy fight. The externality is not the data center as such. It is the resource draw. Cooling water consumption from open-loop evaporative towers is the most measurable externality at a typical hyperscale facility, because the water comes from a municipal system that has finite capacity and other claimants. If facilities had to pay the true marginal cost of that water at peak summer load, the financial case for evaporative cooling would shift sharply against it. Closed-loop and dry cooling architectures that look expensive on a capex basis become competitive on a TCO basis once water is priced correctly.
The same applies to power. Cooling typically consumes 30 to 40% of the electricity used by a data center. If the facility had to pay grid-impact-adjusted rates rather than industrial flat rates, the marginal value of every percentage point of PUE improvement would rise. Vendors offering high-efficiency cooling architectures would face less price pressure because the operator's incentive to minimize total cooling load would intensify.
The economists' position is roughly the policy path the cooling industry should want. A regime that prices water, electricity, and grid impact at marginal cost is a regime that rewards the cooling architectures that already perform best on water and energy efficiency. There are 67% energy savings on the table from component-level cooling efficiency that most data centers leave untouched. The reason they leave them is that the resources are cheap enough that the capex investment to capture the savings does not pencil. Change the price of resources, and the capex investment penciled overnight.
The contrast with the alternative paths is instructive. A blanket construction ban freezes the existing inefficient capacity in place because operators have no incentive to retrofit. A subsidy-driven buildout encourages operators to build cheap and forget about long-term operating costs because the tax structure favors capex deployment over efficiency. The economists' middle path forces a continuous optimization on cooling spend that the industry has been wanting for a decade.
The political math in Ohio suggests that some version of the economists' framework is the most likely outcome over a 24-month horizon. The state legislature is hearing from both sides of the existing fight: the operators who want continued abatement structures and the activists who want a ban. The middle position the economists describe is the only one that can plausibly hold a majority.
The cooling vendor base should be making the case for that middle position directly. Operators who internalize their cooling resource costs become customers for higher-performance equipment. The marketing argument writes itself: "When water and power are priced correctly, our equipment pays back faster than the incentive structure ever did." That is a different conversation than the one cooling vendors are currently having, which is mostly about capex on a PUE basis. Repricing the resource environment makes the cooling industry the obvious beneficiary, and the cooling industry should be willing to say so out loud.