Data Center Ballot Measures: 2026 Election Results

No statewide data-center measure was on the November 3, 2026 ballot. The only filed statewide proposal—an Ohio constitutional ban on data centers above 25 megawatts—missed the signature deadline and is aimed at 2027. What voters decided was a local wave: Ballotpedia’s canonical list is 41 measures in 10 states, and on the numbers it posted November 4 the side limiting data centers won 40 of 41. The sole exception was Harwood, North Dakota, where voters rejected extra review of industrial projects, including an Applied Digital campus. Local reporting also shows at least five more data-center questions (Americus, Kansas, and several Maine towns) that Ballotpedia did not include in the 41. Every figure below is unofficial. Certification is still ahead; two days after a midterm, local canvasses are not complete.

What was actually on the ballot

The fight was not about utility-rate formulas. No November 3 measure set electric rates. The live questions were bans and megawatt caps, short moratoriums, voter-approval triggers tied to acres, power, and water, tax-incentive conditions, zoning referendums on named parcels, and nonbinding advice to local boards.

Ballotpedia’s pre-election split: a yes vote limited development on 28 measures and facilitated it on 13. Residents, not councils, put most of them there. Of the 41, FourWeekMBA’s reading of Ballotpedia’s type labels counted 25 citizen initiatives, 11 referrals, and 5 referendums on decisions already taken. Ohio alone had 21 measures in 18 jurisdictions—more than the other nine states combined. Illinois had seven, all advisory. Maine had four on Ballotpedia’s list and more in town warrants. Kansas had three on that list plus Americus.

The binding pattern is a 25-megawatt line, copied from the failed Ohio statewide amendment into local charters. That threshold is aimed at hyperscale sites, not a small server room. Grove City’s charter change is the water-and-power version of the same idea: a public vote if a project hits 50 contiguous acres, 20 megawatts of peak demand, or 500,000 gallons of water a day. Perry, Ohio, is the tax piece: hyperscale developers seeking special tax incentives must post financial guarantees. Wilmington’s four questions were not abstract policy; they were referendums on specific rezonings from rural residential to light industrial. Ashville’s question was a term summary for negotiations with EdgeConneX on two data centers.

Scoreboard

As of Ballotpedia’s November 4 compilation, voters had decided all 41 November 3 measures it tracks. In 40, a majority favored greater limits—either by passing a restriction or by rejecting a measure that would have allowed or endorsed a project. Harwood is the one outcome that favored development. Margins were mostly lopsided. Grove City (about 55–45) and Sprigg Township, Ohio (about 53–47) were the close ones. Morgan County, Illinois (88–12 advisory) and Monroeville and Downe Township advisory “allow” questions (about 12–13% yes) were among the most one-sided.

National outlets converged on the same story with slightly different clocks. CNBC, citing Ballotpedia plus its own Wednesday check, reported voters sided against data centers in every contest with results except Harwood. The New York Post, also citing Ballotpedia, said voters decided against centers in 47 of 50 measures tracked for the full year, with three still pending—those three are December votes, not November 3. Duane Morris’s Wednesday-morning snapshot still had a few uncalled; Ballotpedia’s later table filled them in. Treat the table as a snapshot, not a canvass.

November 3 table (Ballotpedia’s 41)

“Restrict share” is the share voting for the anti-data-center outcome (yes on a limit, no on a facilitation). Spending is reported only where a sourced figure exists. “Not found” means no committee total turned up in sources retrieved by November 5—not that spending was zero. All results uncertified.

Jurisdiction	What a yes vote did	Yes / No (Ballotpedia, Nov. 4)	Restrict share	How it reached the ballot / known sponsors and opponents	Money	Status
Alhambra, CA — Measure DB	Ban data centers citywide (locks in a council ordinance)	12,496 (82%) / 2,772 (18%)	82%	Council referral. Ballot argument signed by the full council (Jeffrey Maloney, Adele Andrade-Stadler, Ross Maza, Noya Wang, Katherine Lee). No opposing argument filed. SGV Progressive Action (Andrew Yip) supported; organizer said there was little campaigning.	Not found for a measure committee	Unofficial
Edgewater, FL — charter amendment	Prohibit new data centers from being approved, built, expanded, or operated	7,875 (77%) / 2,328 (23%)	77%	Council referral after a council ban	Not found	Unofficial
Bloomington, IL — advisory	Advise a moratorium on large-scale data centers	21,851 (78%) / 6,316 (22%)	78%	Citizen initiative; nonbinding	Not found	Unofficial
Chouteau Twp., Madison Co., IL — advisory	Advise officials to permit new data centers	365 (17%) / 1,817 (83%)	83%	Citizen initiative; nonbinding	Not found	Unofficial
Granite City Twp., IL — advisory	Advise officials to permit new data centers	1,959 (26%) / 5,703 (74%)	74%	Citizen initiative; nonbinding	Not found	Unofficial
Joliet Twp., Will Co., IL — advisory	Advise approval of AI data center construction	3,800 (21%) / 14,207 (79%)	79%	Referral; nonbinding	Not found	Unofficial
Morgan County, IL — advisory	Advise restrictions on building and operating data centers	9,413 (88%) / 1,250 (12%)	88%	Citizen initiative; nonbinding. Highest restrict share on the Ballotpedia list	Not found	Unofficial
Nameoki Twp., IL — advisory	Advise officials to permit new data centers	602 (22%) / 2,117 (78%)	78%	Citizen initiative; nonbinding	Not found	Unofficial
Venice Twp., IL — advisory	Advise officials to permit new data centers	311 (24%) / 978 (76%)	76%	Citizen initiative; nonbinding	Not found	Unofficial
El Dorado, KS	Prohibit high-impact data centers and Type 2/3 battery storage	2,966 (71%) / 1,183 (29%)	71%	Citizen initiative	Not found at the measure level. Build American AI bought about $1.1 million in Kansas TV ads Sept. 7–Oct. 24; those were statewide issue ads, not a filing for this question	Unofficial
Emporia, KS	Prohibit high-impact data centers	4,346 (60%) / 2,891 (40%)	60%	Citizen initiative. Emporia Gazette confirms the same count. A separate battery-storage question also passed and is not counted here	Same Kansas ad caveat; no measure-committee total found	Unofficial; Lyon County canvass was scheduled for the Monday after the election
Ottawa, KS — Question 2	Prohibit data centers using more than 25 MW	2,882 (72%) / 1,129 (28%)	72%	Citizen initiative. KSHB identified the sponsor as Ottawa United. Duane Morris had an earlier 69–31 read; use Ballotpedia’s later count and expect the canvass to move the last digits	Not found at the measure level	Unofficial
North Berwick, ME — Q2	Prohibit data centers of 25,000 sq ft or more	2,114 (78%) / 582 (22%)	78%	Town referral	Not found	Unofficial
Peru, ME — Art. 2	180-day moratorium on data center development and permitting	578 (76%) / 182 (24%)	76%	Town referral	Not found	Unofficial
Wells, ME — Art. 5	180-day moratorium; Ballotpedia News said the text also nullifies permits issued within 45 days of the effective date	5,818 (82%) / 1,282 (18%)	82%	Town referral	Not found	Unofficial
Wiscasset, ME — Q2	Six-month moratorium on applications for new or expanded data centers	1,363 (66%) / 712 (34%)	66%	Town referral	Not found	Unofficial
Harwood, ND	Require independent review, a hearing, and written findings before approval of industrial projects, including data centers	105 (24%) / 325 (76%)	24% — failed	Citizen initiative by Voices of Harwood (Mike Swartz and others) after Applied Digital’s campus, with Oracle described as the end customer. Opposed by Mayor Blake Hankey and the Fargo-Moorhead-West Fargo Chamber, which mailed against it. Swartz said he used his own money and that the chamber matched mailers	Organizer self-funded; no dollar total retrieved. Chamber mail volume not quantified in dollars	Unofficial; Ballotpedia: 100% of precincts, 430 votes, updated Nov. 4, 2:24 p.m. ET, “unofficial until certified”
Downe Twp., Cumberland Co., NJ — advisory	Advise support for data centers	62 (13%) / 425 (87%)	87%	Referral; nonbinding	Not found	Unofficial
Boulder City, NV — Q1	Allow data centers in the Eldorado Valley Transfer Area	1,973 (29%) / 4,726 (71%)	71%	Council referral (Feb. 24). Citizen volunteer committees wrote the for and against arguments. Supporters framed it as a revenue option on city land, not an automatic project approval	Not found	Unofficial. One Wednesday-morning tracker still had mail ballots open; Ballotpedia later posted these totals
Ashville, OH	Uphold a term summary guiding negotiations with EdgeConneX on two data centers	331 (18%) / 1,469 (82%)	82%	Referendum on a village action. A separate early read had 1,468–330	Not found	Unofficial
Conneaut, OH — Issue 10	Charter ban above 25 MW	2,301 (65%) / 1,218 (35%)	65%	Citizen initiative. One of 10 charter measures Conserve Ohio (Nick Owens, Austin Baurichter) helped place, using model 25 MW language. Data Center Coalition warned the bans would forgo economic benefits	Not found for this city	Unofficial
Defiance, OH	Ban above 25 MW	3,823 (65%) / 2,083 (35%)	65%	Citizen initiative; Conserve Ohio model	Not found	Unofficial
Granville, OH	Ban data center construction	2,015 (82%) / 429 (18%)	82%	Citizen initiative; Licking County, next to the New Albany hyperscale cluster	Not found	Unofficial
Grove City, OH — Issue 28	Voter approval if a project is ≥50 acres, >20 MW peak, or >500,000 gal/day water	11,317 (55%) / 9,411 (45%)	55%	Protect Grove City (Jen Belt, Karen Church, and others), aimed at a Headwaters plan to annex about 310 acres for at least six server buildings. Opposed by Save Grove City, created with the Ohio Chamber of Commerce (Adam Rapien; Patrick Castro as a public face). Mayor Ike Stage said the city would implement it; the law director is reviewing constitutional questions	Preliminary Franklin County filings, reported by WOSU before Election Day: Protect Grove City just over $3,500 from 34 donors; Save Grove City $145,000 from 8 donors, including $50,000 from the Ohio Chamber and nearly $100,000 from construction-contractor PACs. Protect Grove City also said the city paid PR firm MurphyEpson more than $64,000 for economic-development messaging that included a counter-strategy on the amendment—that is city money, not a campaign committee	Unofficial. WOSU reported 54.61% from the Franklin County board, consistent in direction with Ballotpedia’s 55%. An earlier Ballotpedia subpage snippet showed a partial 6,110–4,976. Use the higher count and wait for the canvass
Hubbard, OH	Ban data centers in the city	2,622 (77%) / 802 (23%)	77%	Citizen initiative	Not found	Unofficial
Monroe Twp., Adams Co., OH	Adopt township zoning (a future veto point, not a named data-center ban)	148 (60%) / 100 (40%)	60% if zoning is counted as a limit	Citizen initiative. Indirect. Duane Morris’s Wednesday-morning post still had this unreported	Not found	Unofficial; small electorate, easy for a late ballot to move
Monroeville, OH — advisory	Advise officials to allow data centers	55 (12%) / 394 (88%)	88%	Referral; nonbinding	Not found	Unofficial
Oregon, OH — Issue 25	Ban above 25 MW	4,994 (61%) / 3,235 (39%)	61%	United Neighbors of Oregon (Jim and Laura Davis, James Davis). Aimed at a proposed Vantage project city officials had hoped would be about $1 billion. Mayor Steve Salander said residents had spoken. The Blade reported Oregon’s margin as the smallest among the 10 Ohio charter bans, about 60.7%	Lyn Cox of Conserve Ohio told The Blade that supporters of the Oregon measure were outspent $4,600 to $335,000. That is an organizer’s claim in a news story, not a filing I retrieved from the county board	Unofficial
Pataskala, OH	Prohibit large-scale data centers	5,418 (74%) / 1,941 (26%)	74%	Citizen initiative; Licking County cluster	Not found	Unofficial
Perry, OH — Issue 16	Require financial guarantees from hyperscale developers seeking special tax incentives	946 (87%) / 140 (13%)	87%	Citizen initiative. This is the clearest tax-abatement condition on the November 3 ballot	Not found	Unofficial
Piqua, OH	Charter ban on data center construction	4,061 (67%) / 2,027 (33%)	67%	Citizen initiative. Same night, voters recalled commissioners who had backed a Meta-linked project—recalls are separate from this measure	Not found for the charter question	Unofficial
Sprigg Twp., Adams Co., OH	Adopt township zoning	330 (53%) / 297 (47%)	53% if counted as a limit	Citizen initiative. Closest race on the Ballotpedia list. Indirect, same caveat as Monroe	Not found	Unofficial
Sunbury, OH	Ban above 25 MW	2,809 (79%) / 763 (21%)	79%	Citizen initiative; Conserve Ohio model; Delaware County	Not found	Unofficial
Trenton, OH — Issue 17	Prohibit large-scale data centers	3,541 (80%) / 903 (20%)	80%	Citizen initiative. Politico reported the council had already approved a nearly 1 million sq ft project and initially refused to certify the petition	Not found	Unofficial
Upper Sandusky, OH	Charter ban on data center construction	1,893 (77%) / 568 (23%)	77%	Referral	Not found	Unofficial
Urbana, OH	Ban data centers in the city	2,978 (77%) / 877 (23%)	77%	Citizen initiative. Mayor was recalled the same night over the data-center fight; that recall is not this measure	Not found	Unofficial
Wilmington, OH — 1957 SR 730	Uphold rezoning of 65.29 acres, rural residential to light industrial	732 (20%) / 2,886 (80%)	80%	Referendum. Four parcel votes, same night, same pattern	Not found	Unofficial
Wilmington — McGuinn Road	Uphold rezoning of 191.12 acres	724 (20%) / 2,887 (80%)	80%	Referendum	Not found	Unofficial
Wilmington — SR 730	Uphold rezoning of 119.373 acres	739 (20%) / 2,885 (80%)	80%	Referendum	Not found	Unofficial
Wilmington — US 68 South	Uphold rezoning of 170.11 acres	716 (20%) / 2,909 (80%)	80%	Referendum	Not found	Unofficial
Janesville, WI	Require a public vote before the city leases, sells, or approves development of the former GM/JATCO site if project cost exceeds $450 million	16,779 (63%) / 9,932 (37%)	63%	Janesville for Responsible Development (Alyssa Capozziello). Written after a Viridian Partners proposal for an 800 MW, roughly $8 billion campus; the city’s letter of intent lapsed and Viridian stepped away before Election Day. The question covers any project over $450 million, not only data centers	Not found at the measure level. Build American AI bought about $1.5 million in Wisconsin TV ads in the same window as the Ohio buy; those ads were not a filing for this question	Unofficial. GazetteXtra matches Ballotpedia’s 16,779–9,932

Sources for the vote columns: Ballotpedia’s November 4 compilation. Corroborating local counts where retrieved: Emporia Gazette and KVOE (Emporia), WOSU (Grove City percentage), The Blade (Oregon percentage), GazetteXtra (Janesville), Ballotpedia’s Harwood page (100% of precincts).

Measures outside Ballotpedia’s 41

Trackers and town reporting show a wider November 3 set. BreezyScroll noted that Data Centers on the Ballot follows five questions not on Ballotpedia’s list. I could confirm some of those with a local source and not others.

• Americus, Kansas. Ordinance to prohibit high-impact data centers and Tier 2/3 battery storage. Yes 274, No 127 (about 68%), reported by the Emporia Gazette and KVOE as unofficial. Lyon County said results stay unofficial until the Monday canvass. Sponsors and spending: not found.
• Eliot, Maine. Select board placed a data-center moratorium on the November 3 special town-meeting ballot. Underclassic, citing the Redbud tracker as of 5:50 a.m. ET on November 4, listed 2,965 yes / 909 no (76.5%). I did not retrieve a town-clerk page, so treat that count as single-source and unconfirmed here.
• Sabattus, Maine. Same tracker: 1,565 yes / 791 no (66.4%). Same caveat.
• Brownfield and Freedom, Maine. The Maine Monitor and Common Dreams listed both among towns with moratorium questions on November 3. Underclassic still had both as “results pending” early on November 4. Results not confirmed in sources retrieved for this report.
• The Maine Monitor’s pre-election list of November 3 town questions was Brownfield, Eliot, Freedom, North Berwick, Peru, Sabattus, Wells, and Wiscasset—eight towns, versus Ballotpedia’s four. North Berwick, Peru, Wells, and Wiscasset are in the table above.

If the goal is “every” municipal question, the honest status is: Ballotpedia’s 41 are documented with vote totals; Americus is documented locally; Eliot and Sabattus appear only in a secondary tracker in the sources I opened; Brownfield and Freedom were on warrants and their counts were not in hand as of this search.

Who paid, and who did not have to

Organized money showed up in a few places and was not decisive where it did.

Grove City is the cleanest committee comparison. Before Election Day, WOSU reported preliminary filings: the yes side (Protect Grove City) had raised a little over $3,500 from 34 donors; the no side (Save Grove City / Ohio Chamber) had raised $145,000 from eight donors, including $50,000 from the Ohio Chamber and nearly $100,000 from construction-contractor PACs. The charter change still passed, by the narrowest margin among the large Ohio cities. That is the mechanism: a chamber-and-trades committee can outspend a petition committee roughly 40 to 1 and still lose when the question is “do you want a vote before the next 300-acre campus.”

Oregon, Ohio, is the other sourced local disparity. United Neighbors of Oregon qualified the 25 MW charter ban against a Vantage proposal. Conserve Ohio’s Lyn Cox told The Blade the yes side was outspent $4,600 to $335,000 and still won about 61–39. I did not retrieve the underlying campaign-finance report, so that dollar pair should be read as a quoted organizer claim, not a verified filing.

Industry money that looked like ballot spending mostly was not. DeSmog found Build American AI—linked to the Leading the Future super PAC—ordered at least $2.54 million in Ohio TV ads, $1.5 million in Wisconsin, and $1.1 million in Kansas between September 7 and October 24. Those ads sold the economic case for data centers during Senate and statewide races. They were not disclosed as committees for Conneaut or Emporia. The Data Center Coalition’s state-branded ad network (Connected Ohio and peers) is the same category: industry persuasion, not a line item on these local questions. CNBC’s figure of more than $377 million from crypto, AI, and online-gambling interests, and Elon Musk’s cycle spending, belongs to candidate races, not these municipal measures.

Harwood is the mirror image. Voices of Harwood qualified a charter amendment with 91 signatures. The chamber mailed against it. The organizer said he paid out of pocket and could not match that. The review requirement lost 105–325. In a town of roughly 800 to 940 people, with an Applied Digital project already under construction, a long charter amendment described on the ballot in language even its author said he might have voted against was a bad instrument. That is why it is a weak “industry win”: the project was already moving, and the question was process, not permission.

For the other three dozen measures, committee totals were not in the sources retrieved. Many were town-meeting warrants or advisory questions with no paid campaign. Alhambra’s organizer said Measure DB won with little or no campaigning, and no opposing argument was even filed. Absence of a reported committee is the normal case here, not evidence of a dark-money total.

What the categories mean for the next project

Bans and caps are now in city charters, which a council cannot quietly undo. That is the point of the Conserve Ohio model and of Alhambra’s Measure DB: the June ordinance already banned data centers; the ballot measure makes a later repeal a voter question. Edgewater did the same move from ordinance to charter.

Moratoriums buy six months, not a permanent no. Wells is the sharp version because the text reaches back 45 days and voids recent approvals. A developer who pulled a permit in the window before Election Day is the party that has to litigate, not the party that won a policy argument.

Referendums killed specific land-use decisions. Wilmington’s four light-industrial rezonings each failed about 80–20. Ashville’s EdgeConneX term summary failed about 82–18. Those are project deaths or forced renegotiations, not zoning philosophy. Janesville’s $450 million trigger survived even after Viridian walked away, so the next applicant at the GM site—data center or not—needs an election if the cost clears that line.

Advisory questions in Illinois and in Monroeville and Downe do not bind. Councils cite them anyway. A 88–12 instruction in Morgan County, or a 17–83 rejection of “permit data centers” in Chouteau Township, is political cover for a denial and a litigation fact for a developer who later claims community support.

Tax and utility structure was barely on this ballot. Perry’s guarantee requirement is the one direct hit on abatements. Grove City and the Kansas “high impact” definitions use water and power as triggers, which is how residents proxy for rates: they cannot set the tariff, so they ban the load that would show up on the bill. The bond-and-PILOT votes that actually authorize tax-advantaged financing—Clarksville, Arkansas, up to $50 billion for Hatchbo LLC and up to $5 billion for SF ARK1 LLC—are December 1, not November 3. Mandan, North Dakota’s 25 MW ban is December 15. Those three are the “still pending” in the year-long 50-measure count.

Recalls ran on the same night and are not in the table because they are candidate elections, not policy measures. Duane Morris reported removals in Piqua and Urbana, Ohio, Mason, Michigan, and Yukon, Oklahoma, tied to data-center decisions. A developer who won a council vote in 2025 can still lose the council in 2026.

How to use this, and what is still open

For anyone siting or opposing a project, the operative fact is local and legal, not national. A charter ban above 25 MW in Sunbury or Defiance is in force once certified, unless a court throws out the amendment. An advisory loss in Joliet Township is a political fact a council can ignore. Grove City’s 50-acre / 20 MW / 500,000-gallon trigger will be litigated; the city has already said the law director is reviewing it. Wilmington’s four failed rezonings mean those parcels are not light industrial unless the city starts over and survives another referendum.

Do not treat the 40–1 score as a certified canvass. Ballotpedia’s own Harwood page says the count is unofficial until certified. Grove City’s percentage is stable across sources; the raw vote total moved as ballots were added. Monroe and Sprigg are tiny and were unreported in at least one Wednesday-morning tracker. Ottawa’s percentage differs by a few points across Wednesday write-ups. Eliot, Sabattus, Brownfield, and Freedom need a town-clerk or state page before they belong in the same table as the 41. Campaign-finance completeness is worse: only Grove City has preliminary filings in the sources used here, and Oregon’s $335,000 / $4,600 figure is a quoted claim.

The competitive implication is straightforward. The industry’s statewide ad buys and chamber PACs did not move these local questions, except possibly in Harwood, where the project was already approved and the ballot language was a liability for the petitioners. The next constraint is not a missing power purchase agreement. It is a charter line at 25 megawatts, a water trigger, or a referendum on the parcel—and, where those tools are unavailable, a recall of the officials who said yes.

---

Recent Findings Supplement (November 2026)

Voters overwhelmingly rejected or restricted data center development in 46 of 47 decided local ballot measures across 13 states in 2026 (41 on the Nov. 3 general election ballot in 10 states), with margins often exceeding 70-80% in favor of limits.[1][1]

This represents a broad local backlash focused on issues like power/water use, zoning, tax incentives, and specific projects, driven largely by citizen initiatives. Ohio accounted for 21 measures (over half), followed by Illinois (7) and others. Results as of Nov. 4, 2026, are overwhelmingly unofficial and uncertified; a small number of counts remain incomplete or unreported.[2][3]

The sole exception favoring development was Harwood, ND (detailed below). Six measures were decided earlier in 2026 (all against development), and three more are scheduled for December. No statewide measures appeared.[1]

Sponsors and opponents: Most measures originated as citizen petitions/initiatives (e.g., ~25 of the 41 on Nov. 3) or local referrals. Opponents typically included residents concerned about utility rates, water consumption, noise, and lack of transparency; supporters often included developers, some local officials seeking tax revenue/jobs, and industry groups. Specific named sponsors/opponents and detailed local campaign finance data are not widely reported in immediate post-election coverage.[2]

Spending: Limited public details available for these local measures. One reported example: Build American AI spent at least $2.54 million on TV ads in Ohio supporting development (citing investment and jobs). Broader tech/AI industry spending occurred in related candidate races (e.g., hundreds of millions nationally, including Elon Musk’s contributions), but direct per-measure breakdowns for ballot questions remain sparse as of Nov. 4.[2][4]

Vote margins and examples (selected from Ballotpedia’s Nov. 3 table; all unofficial):[1]

• Alhambra, CA (Referral, Measure DB – Prohibit data centers): 82% Yes (12,496) / 18% No (2,772).
• Edgewater, FL (Referral – Prohibit new data centers): 77% Yes (7,875) / 23% No (2,328).
• Bloomington, IL (Initiative – Advisory moratorium on large-scale centers): 78% Yes (21,851) / 22% No (6,316).
• Morgan County, IL (Initiative – Advisory restrictions): 88% Yes (9,413) / 12% No (1,250).
• El Dorado, KS (Initiative – Prohibit high-impact centers and certain storage): 71% Yes (2,966) / 29% No (1,183).
• Ottawa, KS (Initiative – Prohibit >25 MW centers): 72% Yes (2,882) / 28% No (1,129).
• Wells, ME (Referral – 180-day moratorium): 82% Yes (5,818) / 18% No (1,282).
• Harwood, ND (Initiative – Additional review for industrial projects including data centers): 24% Yes (105) / 76% No (325) — industry-favorable outcome (voters rejected stricter review).[5]
• Ashville, OH (Referendum – Specific EdgeConneX development agreement): 18% Yes (331) / 82% No (1,469).
• Granville, OH (Initiative – Prohibit data centers): 82% Yes (2,015) / 18% No (429).
• Grove City, OH (Initiative – Voter approval for large developments): 55% Yes (11,317) / 45% No (9,411) — closer margin.
• Other Ohio examples (e.g., Conneaut, Defiance, Hubbard) typically 65-77% in favor of prohibitions on large (>25 MW) centers.

Additional patterns include high support (often 70%+) for bans/moratoriums on hyperscale facilities (>25 MW or high water use) and advisory measures against development. Some targeted specific projects (e.g., EdgeConneX in Ohio).[6][7]

Implications for competition/entry: The results signal heightened local approval risk for data center projects, particularly hyperscalers, potentially delaying or blocking builds in affected jurisdictions. Developers may face more referendums, stricter zoning/utility requirements, or shifts toward smaller/enterprise facilities. Industry responses could include greater community engagement, revised incentive structures, or focus on states/locales with fewer measures. Three December measures and any certifications could provide further updates.[3]

Sources for table/details: Primarily Ballotpedia’s comprehensive tracker (updated Nov. 4, 2026), cross-referenced with contemporaneous reporting from CNBC, Washington Post, Business Insider, and others.[1][8][1] Full lists and ongoing updates are available on Ballotpedia and trackers like Data Centers on the Ballot (Redbud Consulting). All Nov. 3 results flagged as unofficial/uncertified pending official canvassing. No comprehensive per-measure spending data located in initial post-election reports.