Candidates in at least 23 governor races took a position on pausing new data centers, many of which won their respective races on November 3rd.

Of the 633 GW of data centers tracked by Modo Energy, 165 GW (26%) sits in states whose incoming governor backs an immediate moratorium. When counting conditional moratoria and pauses already in force, 256 GW is exposed. However, very little of that headline number is under construction or permitted. Pauses would mostly bar announced projects.

Key takeaways

• State-level data center pauses could push capacity to planned projects in other markets. This could lower the net buildout through 2035 as other states and markets pick up the slack.
• Texas will have the most consequential impact on the national data center buildout under a pause or slowdown. 124 GW of operating and planned capacity are located there.
• Taken on their own terms, the pauses would stop about 11 GW of the capacity Modo Energy expects to be built in these states, delay 1 GW and put 4 GW on new terms. Another 39 GW is under construction, and its outcome under a pause is uncertain.
• Most of the headline impact is paper. Of 242 GW not yet operating in these states, 41 GW is under construction and 201 GW is proposed.

Where did governors win with a position on data centers?

Five winners back an immediate moratorium:

• Gina Hinojosa in Texas
• Keisha Lance Bottoms in Georgia
• Deb Haaland in New Mexico
• Hannah Pingree in Maine and
• Kelly Ayotte in New Hampshire

Three more winners back conditional moratoria. Amy Acton in Ohio would pause new campuses unless they cover 100% of their utility costs, build with union labor and meet Ohio's environmental standards. Jocelyn Benson in Michigan and David Crowley in Wisconsin set similar conditions.

Three states already have a governor's pause in force. New York's Executive Order 62 pauses state environmental permits for projects of 50 MW or more. Oregon is holding state land and environmental permits until July 2027. Illinois has stopped signing new incentive agreements.

Kansas is still too close to call at the time of writing. Cindy Holscher (D) announced a plan to block approvals until guardrails exist. Ty Masterson (R) would leave the decision to local communities.

Would a moratorium slow national data center growth?

A moratorium starts by pushing grid studies and permitting back. In Texas, that may clear when the legislative session ends, which is May 31, 2027. Not every pause is meant to permanently bar data centers. Most candidates focused on keeping costs off ratepayers, or on the environmental impact of energy use and on-site generation.

Even so, a short pause can last longer than intended. A governor’s announcement becomes tied to the legislative session, court schedules, and future elections. Projects are tied to grid study cycles, permitting, procurement and supply chains. Permitting for on-site generation could add friction after a moratorium lifts. Disputes over legal definitions, protections and cost allocation could spill into litigation, rulemaking, public comment periods and legislative efforts that may not end up in a single passed bill.

Capital constraints are also affected. Developers have a finite pool to deploy, and costly GPUs lose value every month they sit idle. Capital flows to jurisdictions where interconnection and permitting is not on hold. With that much friction, momentum and inertia, a temporary delay can end up looking permanent.

Given that, Modo expects projects elsewhere to develop while paused projects wait, and the delayed projects rejoin the queue behind them. For example, a one-year pause in ERCOT ends about 9 GW behind the forecast by 2035, and MISO, SPP and the West make up the capacity. A pause of two years or more looks more like an open-ended moratorium, with ERCOT about 13 GW behind.

Texas is the swing case in this forecast. Hinojosa's pledge has no firm end date, with a range between the end of the legislative session or until favorable terms are met. If it lifts when the legislative session ends in May 2027, Batch Zero would likely slip to summer 2027, a six- to nine-month delay rather than a stop. Modo Energy's national forecast then falls from 147 GW to 136 GW of capacity in 2030 if nothing replaces the lost projects.

However, when projects in the pause states drop out or face delays, the expectation is that projects in line elsewhere take their place. A data center in MISO from the same developer that faces a pause in ERCOT may reallocate GPUs. With that reallocation, 2030 capacity falls by only 3 GW. By 2035 the gap is 4 GW, about 2% of the forecast. There is a significant degree of uncertainty here, pursuant to political appetite, litigation, and ultimate form of any pauses.

How much data center capacity is now in question?

Those 11 states’ races could affect 256 GW of the development pipeline, or 40% of tracked US capacity. Most of that is still just announced. Of the 242 GW not yet operating or expanding, 41 GW is under construction. The other 201 GW is proposed, permitted, waiting on a grid connection, or simply announced.

Taken on each pause's face-value terms:

• 131 GW of that pipeline would stop
• 14 GW would be delayed and
• 54 GW would face new terms.

The 41 GW under construction is not explicitly targeted, but the outcome of early-phase projects will remain unclear. This is particularly so if on-site generation or other non-construction permits are halted, as has been noted in New Mexico.

Open-ended moratoria in Texas, Georgia, New Mexico, New Hampshire and Maine could stop proposed projects from inauguration in January. That includes permitted projects, which still need grid connections and further permits. Together, they may stop up to 11 GW of expected build, 10 GW of it in Texas.

Time-limited pauses in New York, Oregon and Wisconsin delay about 1 GW. Conditional pledges in Ohio and Michigan, plus Illinois's incentive pause, put another 4 GW on new terms without stopping it.

Under-construction projects are an open question. 39 GW of expected builds in these states is already under works, with 24 GW of it in Texas alone. Most pledges target new projects, and Texas Governor-elect Hinojosa's sets conditions such as no subsidies and proof that household bills won't rise. However, these campuses still need permits and grid approvals to finish, so their outcome under a pause is uncertain.

How does this look on a market level?

By market, ERCOT has the most capacity under uncertainty with all 119 GW of its tracked capacity sitting under a governor who backs a pause. In PJM, 60 GW, or 26%, is exposed, mostly in Ohio and the ComEd portion of Illinois.

Can new governors turn pledges into permanent law?

That depends on legislatures, court battles, and other mechanisms. It’s also not clear that there is a desire for permanent pauses. Whatever the ultimate path, most of the capacity in reach is still press releases. In every pause state, at least 61% of tracked capacity is only announced or waiting on a grid connection.

What happens next?

New governors take office in January, and most legislatures convene the same month. Before then, ERCOT's audit is planned will show whether Texas's permit freeze lifts or stays. Loudoun's vote will show whether the country's largest data center market joins the local pauses. Other governors that have conditions for data center developers will have to run the gauntlet of implementation or legislation. For developers, politics are another factor to consider in execution risk, along with the interconnection queue and supply chains. System operators will need to adjust forecasts and timing as well, with potential reallocations of projects around the country.