I. Introduction
As the global race for artificial intelligence (AI) hastens, data centers are driving unprecedented demand growth for electric power. Projections show data centers consuming nearly 10% of total U.S. electricity by 2030, a near fourfold increase from 2023 levels. Goldman Sachs projects that AI-driven data center electricity demand alone will soon rise 160%,1 and Bain researchers estimate that utilities will need to increase generation capacity by 7 to 26% to keep pace.2 U.S Senator Jon Husted (R-OH) recently announced a proposed 9.2 gigawatt natural gas facility to help meet projected growth,3 while more than two gigawatts of additional gas-fired generation is being planned behind-the-meter to supply data centers directly, bypassing the grid altogether.4 These developments represent an enormous economic opportunity. Data centers bring communities billions of dollars in private capital investment, thousands of new jobs, and a steady stream of tax dollars to state and local governments. But they also demand new electricity generation and transmission infrastructure—something that regulatory inertia has made difficult for states to deliver quickly. States that can supply reliable, affordable power on a commercially viable timeline will gain significant competitive advantage. At the same time, they must address the public’s fear that households will ultimately subsidize the energy infrastructure that big tech companies need.
Ohio offers a compelling case study of energy policy reforms and failures for other states to consider. House Bill 15, signed into law in 2025, adopted a successful suite of market-oriented energy reforms—site permitting deadlines, behind-the-meter power rights, brownfield siting incentives, and grid transparency measures—that address critical barriers to new generation. But Ohio’s experience also illustrates the limits of legislative reform when monopolistic utilities resist competition. American Electric Power’s (AEP) poorly designed data center tariff, which requires customers to pay for at least 85% of contracted capacity regardless of actual usage, and a unilateral connection moratorium have undermined the investment House Bill 15 hopes to attract. The lesson for other states: permitting reform and pro-market frameworks must be paired with regulatory vigilance. States that adhere to sound economic principles can attract data center investment and protect consumers by trusting markets over government preferences.
II. The Permitting Bottleneck
A natural gas power plant—the workhorse of reliable, dispatchable generation—takes approximately five years to build and costs more than a billion dollars.5 Yet, the most immediate barrier to meeting AI-driven electricity demand is not technology or money, but the time required to navigate multi-layered government approval processes before a single megawatt of new generation can come online.
Regulatory permitting delays operating at every level of government compound costs, erode investor confidence, and reduce investment. Local zoning and siting boards, spurred by community opposition, require public hearings, impose noise ordinances, and often derail projects before they proceed much further. At the state level, the Ohio Power Siting Board regulates energy generation plants and transmission lines, reviewing need, environmental impact, water use, and consistency with regional plans. The board includes the heads of multiple state agencies and the chair of the Public Utilities Commission of Ohio (PUCO), creating a multi-agency review process that can languish indefinitely. Beyond the state, an alphabet soup of federal agencies like the Federal Energy Regulatory Commission (FERC), the Nuclear Regulatory Commission, and the U.S. Environmental Protection Agency create a power-production bottleneck by inefficiently regulating interstate energy transmission, wholesale electricity commerce, emissions, and every facet of environmental compliance.
Environmental reviews under the National Environmental Policy Act alone take an average of 4.5 years,6 and FERC’s dysfunctional oversight epitomizes the problem. PJM Interconnection, the regional transmission organization serving Ohio and 12 other states, has more than 200 gigawatts of proposed generation awaiting approval in its interconnection queue. Median study durations exceed 36 months, and 80% of projects ultimately withdraw—a textbook tragedy of the commons in which speculative applications clog the administrative pipeline for legitimate projects. By comparison, the Electric Reliability Council of Texas (ERCOT), which operates outside FERC’s regulatory framework, connected 14.2 gigawatts of new generation in 2021–22 while PJM, more than twice ERCOT’s size, connected only 5.6 gigawatts. ERCOT processes interconnection in two to three years; PJM requires more than five.
Federal bottlenecks lie largely outside any single state’s control, but states can control their own siting and permitting process. States should correct common regulatory mistakes such as agency reviews without deadlines, agencies acting without accountability, and outmoded regulatory processes designed for flat energy-demand growth. These tractable problems require political will, not technological breakthroughs, and Ohio has already shown what well-focused political will can accomplish.
III. Ohio’s House Bill 15—The Permitting Reform Model
Following several recommendations outlined in The Buckeye Institute’s “Better Energy Policy for Ohio,” House Bill 15, signed into law in 2025, aggressively pursues four related reforms to strategically address the regulatory bottleneck that impedes power production in Ohio.7 First, the law introduced “shot-clock” provisions, behind-the-meter generation rights, and tax reductions for new power plants. So-called shot clocks set binding deadlines for the Ohio Power Siting Board’s review of energy project applications. Previously, without enforceable deadlines, the board’s project reviews could extend indefinitely as agencies requested additional information, scheduled hearings, and deliberated without consequence for delay. House Bill 15 defines a timeframe within which the board must act, transforming the permitting process from an open-ended negotiation into a more predictable procedure. For developers weighing billion-dollar capital commitments, the difference between an uncertain multi-year review and a defined timeline may determine whether a project proceeds in Ohio or moves to another state.
Second, House Bill 15 establishes brownfield siting incentives that steer data center and energy infrastructure toward previously developed sites. Brownfield locations offer multiple advantages: existing road access and utility connections that reduce development costs, less community resistance to already industrialized sites, and streamlined environmental reviews that begin with well-documented baseline conditions. By creating explicit incentives for brownfield development, House Bill 15 simultaneously accelerates permitting timelines and advances land-use goals that enjoy broad political support.
Third, the legislation addresses information asymmetry by requiring publication of transmission capacity heat maps that show where the public grid has available capacity and where it remains constrained. Before this required mapping, developers lacked vital information about grid conditions, often investing months of planning for sites that proved unviable once transmission constraints became apparent. Public heat maps allow investors to self-select sites with available capacity, reducing wasted applications and focusing development where infrastructure can most readily support it.
Finally, House Bill 15 provides tax relief for new power plants, reducing the upfront cost burden that discourages energy investment in Ohio relative to competing states. Tax relief, along with the law’s procedural reforms, helps align Ohio policy with the critical time and cost aspects of power generation across the state.
Other states can follow Ohio’s lead and adopt House Bill 15’s reforms. Creating a “shot clock” may prove the simplest correction so far as any state with a siting board, energy facility review process, or public utility commission can implement binding review timelines without restructuring its regulatory apparatus. The new rule follows the straightforward principle that regulatory agencies should exercise authority within defined periods. And, by adhering to this simple rule, Ohio and other states can meaningfully accelerate the application-to-construction timeline.
IV. Behind-the-Meter Generation—Unlocking a Market Solution
The electric grid cannot keep pace with the buildout of artificial intelligence. A modern data center can be sited, permitted, and operational within eighteen to twenty-four months.8 Connecting that facility to the grid through PJM’s interconnection queue takes four to five years.9 For technology companies competing to deploy in the global AI race, that gap creates significant disadvantages. Behind-the-meter (BTM) generation resolves the mismatch. Instead of waiting for a utility connection, a data center developer builds onsite generation, typically a natural gas turbine, on the customer’s side of the meter. The facility produces its own power without passing through the utility’s distribution system or the Regional Transmission Organization’s interconnection process. An onsite gas turbine can be operational in 12 to 18 months, or less time than it takes to build a data center.10 The months saved could prove decisive in a burgeoning sector that measures competitive advantages in weeks, not years.
The logic extends beyond speed. BTM generation provides the uninterrupted reliability that AI infrastructure demands—power available every minute regardless of grid congestion, transmission constraints, or queue delays. Fortunately, Ohio’s abundant natural gas reserves and pipeline and storage infrastructure make onsite generation viable at scale. BTM is not a permanent substitute for a grid connection, but it allows developers to move with the market while transmission infrastructure expands. The facilities can connect when the grid is ready, and, in the meantime, the investment, jobs, and tax base remain in Ohio.
House Bill 15 removed regulatory and legal ambiguities to make this possible. With no statutory framework governing industrial onsite generation, it was unclear whether Ohio law permitted BTM arrangements, and that uncertainty alone deterred developers from making the necessary capital commitments to proceed. House Bill 15’s clear legal structure for behind-the-meter generation authorizes data center developers and other large industrial users to build and operate onsite power plants. And it did not take long for developers to act on the clarified authority. Within mere months, Ohio saw $4 billion in data center investments, more than 2 gigawatts of approved BTM projects, and individual power plants costing upward of $1.6 billion each.11 For perspective, the entire PJM grid, serving 67 million people across 13 states, added just 2,669 megawatts of new dispatchable capacity in its most recent reliability auction, the first increase after four consecutive years of decline. Ohio’s BTM approvals alone nearly matched that figure, driven entirely by private capital responding to a single regulatory reform.12
By removing regulatory barriers that prevented willing buyers and sellers from transacting, Ohio blazed a market-oriented reform trail that should be easy for other states to follow.
V. The Subsidy Trap—What Ohio Got Wrong Before House Bill 15
Ohio’s energy reform record is not unblemished. Before House Bill 15 corrected course, the state’s most consequential recent energy legislation was House Bill 6 of 2019. That infamous law13 introduced ratepayer-funded subsidies for two aging nuclear plants and several coal facilities, and added new market distortions to an already complex regulatory environment.14 By distorting the energy supply mix and subsidizing inefficient plants, House Bill 6 hindered investment in new dispatchable generation and replaced reliable baseload with intermittent sources that cannot serve around-the-clock industrial demand. When data center investment arrived and electricity demand surged, Ohio’s grid lacked sufficient capacity. Predictably, electricity prices spiked, capacity auctions tightened, and data centers took the political blame for a supply shortfall years in the making. And the fallout is palpable: nearly 20 Ohio municipalities have enacted or are considering moratoriums on data center construction; packed public meetings and referendum threats have spread across the state; and legislators have responded with a flurry of restrictive proposals. The backlash is not unexpected but misdirected.
Data centers did not create Ohio’s energy supply problem—they exposed it. In fact, aware of their sector’s disruptive energy demands, some of the industry’s largest players have signaled their intent to pay their own way: Anthropic committed $50 billion to building custom data centers with private capital,15 and Meta has pledged more than $600 billion in infrastructure investment over three years.16 House Bill 15 is unwinding Ohio’s earlier mistakes, but course correction takes time and market participants still suffer consequences.
VI. Unfinished Reform: AEP’s Data Center Tariff Dulls a Competitive Edge
House Bill 15 addresses permitting timelines, BTM rights, and subsidy distortions, but it does not fix rent-seeking behavior by AEP, Ohio’s dominant utility. In March 2023, AEP imposed a unilateral moratorium on new data center connections, citing transmission constraints, without prior PUCO approval and without adequate evidence that data centers were the primary cause of those constraints.17 PJM Interconnection already maintains transparent processes for identifying and approving transmission needs. AEP bypassed those processes in favor of a blanket ban—then used the ban to extract contractual terms no competitive market would produce.18
The moratorium was ultimately replaced with something even worse. In July 2025, PUCO approved AEP’s data center tariff, a mandatory rate plan governing how large customers connect to and pay for electricity service. The plan imposes an 85% minimum demand charge, 12-year contract terms, and steep exit penalties for customers exceeding 25 megawatts.19 The commission had an opportunity to require genuine negotiation—to weigh competing interests, examine AEP’s cost justifications, and arrive at terms reflecting the actual economics of serving large loads. Instead, it approved a tariff that treats Ohio’s fastest-growing class of electricity customers as a liability rather than an asset.
This tariff is economically self-defeating. Adding large, predictable loads—like data centers—to a grid reduces per-unit costs for all ratepayers by spreading fixed costs across a larger base. Driving such customers away has the opposite effect. AEP’s tariff does not protect residential customers from cost-shifting, instead it guarantees that they will bear a larger share of a $72 billion capital expansion project.20 AEP initially projected 30,000 megawatts of data center demand, but, once developers faced the tariff’s terms, demand estimates collapsed to 5,700 megawatts.21 The $72 billion infrastructure will still likely be built and the costs recovered, but with fewer customers to share them.
By treating data centers as a categorically distinct customer class subject to uniquely punitive terms, AEP courted a legal challenge from the Ohio Manufacturers’ Association Energy Group over discriminatory pricing. The matter is before the Ohio Supreme Court22 and the case’s outcome may determine how Ohio and other states define permissible utility conduct toward large-load customers. The broader lesson reveals that legislative reform and monopoly utility conduct are distinct problems requiring distinct solutions. A utility with exclusive service territory and guaranteed cost recovery has structural incentives to resist large customers and negotiate aggressively. So, when PUCO fails to scrutinize unilateral restrictions, even well-designed legislative reforms can be undermined.
House Bill 15’s behind-the-meter framework provides a solid short-term correction. While the legal and regulatory fights over AEP’s tariff work through the courts, data center developers can pursue BTM generation on timelines that do not depend on AEP’s cooperation or PUCO’s pace. That is not a permanent solution—the long-run efficiency of a unified grid serving diverse loads remains superior to a patchwork of isolated generators—but it ensures Ohio’s data center investment does not stall while utilities catch up. The market found a way around the obstacle, and now the state must remove it.
VII. Conclusion
AI infrastructure will be the defining capital investment cycle of the next decade—the sector driving the largest share of U.S. economic growth in new jobs, tax revenue, and private fixed investment. To win the AI infrastructure race, states must remove barriers to private investment, enforce competitive markets, and hold monopoly utilities accountable. In a global competition for capital deployment and technological advantage, every year spent litigating interconnection rights or defending monopoly tariffs is a year that investment and technology flow to other states and other countries with fewer institutional frictions. The permitting, legal, and regulatory decisions states make today will determine whether adequate generation exists tomorrow. The highest-value reforms available now are at the state level—in utility commissions, permitting offices, and legislative committees where the rules governing energy infrastructure are made. But states do not need to start from scratch.
House Bill 15 provides a workable, replicable framework for energy policy reform: binding permitting timelines, explicit behind-the-meter generation rights, subsidy elimination, and regulatory oversight of utility conduct. Ohio’s successes with BTM authorization and permitting shot clocks offer exportable models. Its ongoing failures with monopoly tariff oversight offer equally instructive warnings. The states that improve fastest will be those that study what other states got right and wrong, adapting those lessons to local conditions, and building the agency capacity to implement them effectively. House Bill 15 remains a work in progress. But its core insight is sound: markets, not mandates, are the most reliable path to abundant, affordable energy. States that embrace that principle and pair it with institutional reforms needed to make markets function will position themselves to capture AI’s enormous economic opportunity.
Aswin Prabhakar is an economic research analyst at The Buckeye Institute, where he conducts original research on energy, innovation, and tax policies and their impact on the economy. The Buckeye Institute is an Ohio-based independent public policy research institute and public-interest law firm. In the energy space, Buckeye advocates for policies that ensure abundant and affordable energy. Learn more at BuckeyeInstitute.org/Energy.
1 AI Is Poised to Drive 160% Increase in Data Center Power Demand, Goldman Sachs Insights, May 14, 2024.
2 Maeghan Rouch et. Al., Utilities Must Reinvent Themselves to Harness the AI-Driven Data Center Boom, Bain & Company, October 2024.
3 Dean Narciso, New $33 billion natural gas power plant could power half of Ohio, The Columbus Dispatch, February 18, 2026.
4 With More Than $4B in Investments, Ohio is Reaping Benefits of Buckeye Institute-Championed HB15, The Buckeye Institute press release, January 15, 2026.
5 Rea S. Hederman Jr. and Greg R. Lawson, Better Energy Policy for Ohio, The Buckeye Institute, January 22, 2025.
6 Council on Environmental Quality, Environmental Impact Statement Timelines (2010–2018), June 12, 2020.
7 Rea S. Hederman Jr. and Greg R. Lawson, Better Energy Policy for Ohio, The Buckeye Institute, January 22, 2025.
8 Alastair Green et. Al., How Data Centers and the Energy Sector Can Sate AI's Hunger for Power, McKinsey & Company, September 17, 2024.
9 Joseph Rand et. Al., Queued Up: 2025 Edition, Characteristics of Power Plants Seeking Transmission Interconnection As of the End of 2024, Lawrence Berkeley National Laboratory, December 2025.
10 Data center power solutions, Wärtsilä (last visited March 6, 2026).
11 With More Than $4B in Investments, Ohio is Reaping Benefits of Buckeye Institute-Championed HB15, The Buckeye Institute press release, January 15, 2026.
12 PJM Auction Procures 134,311 MW of Generation Resources; Supply Responds to Price Signal, PJM Interconnection press release, July 22, 2025.
13 Natalie Fahmy and Maeve Walsh, Ex-Ohio House speaker Larry Householder sentenced in $60 million bribery scandal, nbc4i.com, June 29, 2023.
14 Rea S. Hederman Jr. and Greg R. Lawson, Better Energy Policy for Ohio, The Buckeye Institute, January 22, 2025.
15 Anthropic invests $50 billion in American AI infrastructure, Anthropic press release, November 12, 2025.
16 Russell Brandom, Anthropic announces $50 billion data center plan, TechCrunch, November 12, 2025.
17 Aswin Prabhakar, Undermining Ohio’s Competitive Edge: How AEP’s Data Center Tariff Harms Ohio, The Buckeye Institute, March 16, 2025.
18 Ibid.
19 Ibid.
20 AEP Reports Third-Quarter 2025 Operating Earnings, Updated Capital Plan Drives New Long-Term Growth Rate, American Electric Power press release, October 29, 2025.
21 Ethan Howland, Manufacturers Say AEP Ohio Still Inflating Data Center Demand After Halving Forecast, Utility Dive, February 6, 2026.
22 In the Matter of the Application of Ohio Power Company for New Tariffs Related to Data Centers and Mobile Data Centers, Case No. 2025-1458, Supreme Court of Ohio (filed November 3, 2025).