Read the threshold, not the headline

Most coverage will call these data center bans. The operative number is 25 megawatts. Pataskala's ballot language covers buildings and infrastructure used primarily or exclusively for digital information services with aggregate monthly demand or peak load above that line. Sunbury's charter text uses the same figure. Facilities below it are untouched, and that distinction is not a loophole: small and enterprise data centers make up 52% of the facilities in Ohio, according to the industry tracker Reason cited.

The threshold is designed to capture large, power-intensive facilities rather than every data center. A large hyperscale facility can use as much electricity as 100,000 homes, according to the Ohio Consumers' Counsel. For projects above the specified threshold, these measures function as prohibitions or consent requirements, depending on the municipality. So a measure written as a size limit functions, for the projects the AI buildout actually wants to place in these towns, as a ban. Developers reading these results should price them that way. Readers should too, in the other direction: a town that "banned data centers" has usually banned one size of data center, and the smaller facilities that were already uncontroversial remain legal.

The margins say this was not a set of coin flips. Granville passed its amendment with 82.45% on the county's unofficial count. Trenton, where a ban aimed at a proposed 18-building Amazon Web Services campus passed 79.68% on Butler County's unofficial count of 3,541 votes to 903, had to go to the state supreme court to get the question in front of voters at all. Sunbury's amendment, aimed at a possible hyperscale Amazon project, passed with 78.64%. Pataskala's passed with 73.62%. In Piqua, voters approved a charter ban above 25 MW by 66.7% to 33.3% on the unofficial count and, in the same election, recalled the mayor and two city commissioners, the Dayton Daily News reported.

Two of those cases carry an asterisk worth keeping. In Piqua, a Meta data center planned for about 607 acres had divided the city, and it is not yet clear whether the amendment stops it, because the city had already signed development and utility agreements for the campus. In Trenton, a separate four-building AWS site already under construction nearby is likely unaffected. Amazon has declined to say whether Tuesday's measures will change its expansion plans. A charter amendment is a strong instrument against a project that has not signed anything. Against a project with agreements in hand, it is the start of an argument, not the end of one.

Ashville presents a different case. Its ballot question asked voters to approve a resolution covering an EdgeConneX development and supply agreement for a campus already under construction. Voters rejected it, with 81.61% voting no on the county's unofficial count. That result blocks approval of the resolution, but neither Fortune's account nor the county results establish that construction itself has stopped.

The one close vote is the interesting one

Grove City did not ban anything. Its Community Consent Amendment, which unofficial county results show passing with 54.60%, or 11,317 votes to 9,411, requires voter approval before the city can approve heavy industrial development, including data centers, that meets any one of three thresholds: 50 or more contiguous acres, more than 20 megawatts of peak electrical demand, or more than 500,000 gallons of water a day. The applicant pays for an independent impact study and gets no control over who writes it or what it finds. The study, covering water, grid demand, noise, traffic, jobs and tax revenue against infrastructure costs, must be public at least 60 days before the vote. Lose the vote and a substantially similar application is barred for 24 months. Win it and the project still has to go through normal zoning, with no permit guaranteed.

That is a materially different offer to voters than a ban, and it nearly lost where the bans won by thirty points. One reading is that Ohio voters were not choosing between yes and no so much as refusing to hand the decision to councils and developers in advance. Another is simpler: a ban is an easy thing to vote for in a town with no project in front of it, and several of the towns that passed bans, in Ohio and elsewhere on Tuesday, have no data center operating, planned or applied for. Both readings can be true at once. What Grove City established is that the consent route is available and passable. It is also a different proposition for developers: a project can proceed to the normal zoning process if voters approve it, although that approval does not guarantee a permit.

Where this leaves the buildout

Ohio is not a marginal market being closed off. It has more than 200 operating data centers, the fifth-highest count of any state, concentrated in Central Ohio, and companies plan up to $40 billion more in investment there by 2030. Tuesday's votes sit on top of 125 local moratoriums already in place. Nationally, 41 measures across 10 states were on ballots, according to the initial Axios/Ballotpedia count reported by ASTIG, with some Arkansas and North Dakota votes still being counted. NBC's exit polls found that artificial intelligence and data centers were a crucial or major factor for 56% of voters nationwide. In Ohio, a pre-election survey by Ohio Northern University's Institute for Civics and Public Policy and SurveyUSA, of 1,650 adults in mid-October, found 55% opposing data center construction in the state against 22% supporting it, and 79% expecting higher local water and electricity costs.

None of that stops a single project that already has its agreements, its power and its permits. What it changes is the sequence for the next ones. Land, grid capacity and a welcome from state officials used to be most of the work. In at least 14 Ohio municipalities, unofficial results show voters approved new restrictions or approval requirements for certain data center projects. The rules differ by location, but the 25 MW threshold is a recurring dividing line. The Washington Post reported that Conserve Ohio, the group that helped residents draft 10 of the charter amendments, says all 10 passed and now wants a statewide ban on large data centers on next year's ballot. The immediate next step is certification. The Ohio Secretary of State expects certified results approximately three weeks after the election. Piqua also faces a county certification deadline of November 24, with the question of its previously signed agreements still unresolved.

Sources

  1. [1] Reason, "Local data center bans pass in Ohio, Florida, California, and elsewhere", November 4, 2026 (read first-hand at Gate 1; Ballotpedia measure counts, 25 MW hyperscaler definition, 52% small/enterprise share via trackdatacenters.com, NBC exit-poll figure)Read source
  2. [2] Unite.AI, "Ohio Voters Back Local Data Center Limits in Unofficial Results", November 4, 2026 (read first-hand at Gate 1; county board figures for Butler, Licking, Delaware, Franklin and Pickaway counties, Pataskala and Sunbury measure text, Grove City amendment terms, Trenton supreme-court history, Ohio Consumers' Counsel context, Ohio Northern University polling, and the Ohio Secretary of State's certification timing and Ashville ballot-catalog wording via its linked official sources)Read source
  3. [3] Fortune, Sasha Rogelberg, "Ohio voters approved data center bans or restrictions in at least 14 municipalities across 21 ballot measures", November 4, 2026, as syndicated by ShiftDelete (Piqua/Meta detail, Trenton AWS detail, 125 moratoriums, $40 billion planned investment)Read source
  4. [4] The Washington Post's national count (at least 29 jurisdictions) and Conserve Ohio account, as reported by ASTIG, "Ohio voters restricted data centers in 14 towns. Nationwide, at least 29 places said no", November 4, 2026 (read first-hand at Gate 1; counts kept separate by unit; the Axios/Ballotpedia count of 41 measures across 10 states also via this source)Read source