T+1 outcome / Ohio Conserve Ohio ballot initiative failed 83361 of 413488 signatures 20 percent / no Nov 2026 statewide vote / ballot lane requires paid signature firms 2-4M budget / grassroots single cycle proven non-viable / Linn County IA 18-month moratorium 2-1 vote Jul 1 despite most restrictive ordinance / Washington County MD 12-month 4-1 doubled from 6 / Prince George's County MD expired Jul 1 2-year council vote next week / county commissioner lane accelerating / structural shift toward local friction
T+1 outcome: Ohio statewide ballot ban on data centers >25 MW MISSED signature deadline by ~80% (Conserve Ohio logged ~83,361 signatures vs 413,488 required by June 30 2026), so NO November 2026 statewide vote on data centers in the largest PJM state — but Linn County IA passed 18-month moratorium 2-1 the same morning DESPITE having adopted 'one of the most restrictive' data-center ordinances in Iowa months ago, Washington County MD approved 12-month freeze 4-1, and Prince George's County MD's expired moratorium goes to 2-year council replacement vote next week. Statewide ballot lane = collapsed. Local county-commissioner lane = accelerating.
Plain-English background on ballot initiatives. A statewide ballot initiative is a mechanism in about half of US states where citizens can put a proposed law or constitutional amendment directly on the general-election ballot, bypassing the state legislature. To qualify, organizers collect signatures from registered voters — the threshold in Ohio is 10% of the last gubernatorial vote total, which for the November 2026 ballot works out to 413,488 valid signatures. Ohio also has a distribution requirement: signatures must come from at least 44 of 88 counties, with a per-county minimum. In practice this means a serious statewide ballot campaign in Ohio needs to pay professional signature-gathering firms (running $3-8 per valid signature), which means a realistic budget is $2-4 million for a single ballot cycle. The Ohio Prohibition of Data Center Construction Amendment was the highest-profile citizen-led attempt to ban data centers in the country. If it had qualified, Ohio voters would have decided this November whether to constitutionally prohibit any data center with aggregate power demand exceeding 25 MW anywhere in the state. That would have been the first US statewide vote directly on data-center-scale limits. What happened. As of yesterday's deadline (June 30, 2026), the sponsoring group Conserve Ohio had collected roughly 83,361 signatures — about 20% of the required total. They publicly acknowledged missing their internal target. There is no November 2026 statewide vote on data centers in Ohio. The structural read. Statewide ballot initiatives against data centers require paid signature-gathering firms and a $2-4M budget. Volunteer-led grassroots efforts do not clear the signature threshold in a single 8-month gathering window. That's a durable structural fact — it applies to every future citizen-led statewide data-center ban attempt in every state with a similar threshold (Ohio, Michigan, Colorado, Arizona, Missouri, and roughly ten others). The petition-to-ballot lane is effectively closed to grassroots opposition without institutional backing. But the local-moratorium lane fired the same morning the ballot lane collapsed. **Linn County IA — 18-month moratorium passed 2-1 on July 1, 2026.** The Linn County Board of Supervisors voted 2-1 (Meisheid and Running-Marquardt for; Scheetz against) to approve an 18-month moratorium on new applications to rezone unincorporated county property to the new EU-3 large-scale data-center zoning district. Effective immediately, in place until 11:59 PM Central Time on January 1, 2028. The key detail — Linn County had ALREADY adopted 'one of the most comprehensive and restrictive' data-center ordinances in Iowa just months ago. The supervisors moved to a moratorium anyway because 'a growing number of proposed projects across the county changed the conversation and showed more study is needed.' Read that literally: even a comprehensive ordinance regime does not slow developer proposal volume, which triggers a moratorium as a second-layer defense. **Washington County MD — 12-month moratorium approved 4-1 on July 1, 2026.** The Washington County commissioners (western Maryland, along I-70) voted 4-1 to approve a yearlong moratorium. Commissioners Wagner, Leatherman, Cline, and Barr for; Parrott against. Originally on the agenda as a six-month moratorium; commissioners doubled it during the meeting after hearing residential concerns about water resources and environmental impacts. **Prince George's County MD — old moratorium expired July 1, new 2-year replacement queued for next-week council vote.** The prior Prince George's County AI-data-center moratorium expired at midnight on Tuesday July 1. Council Member Wala Blegay confirmed the votes are there to pass a two-year replacement. Council Chair Krystal Oriadha confirmed the vote is scheduled for next week. A rally against data centers happened outside the Curry Administration Building in Largo yesterday, met with counter-protesters. Residents pushing for a permanent ban call the 2-year moratorium a 'stalling tactic.' The pattern. When the highest-profile statewide ballot-democracy attempt collapses on the exact same day that three separate county-level moratorium actions land, that's the market telling you where the durable friction lives. County commissioners are cheap to influence, they don't need to pass a signature threshold, and they have the political incentive to be seen as protecting residential ratepayers. Statewide ballots are expensive, slow, and structurally biased toward institutional actors. From a Cliff-relevant perspective: any regulatory-tracking product should weight county-commissioner ordinance dockets much more heavily than statewide ballot filings.
Why it matters
Three implications. (1) The statewide-ballot-vs-county-moratorium divergence is now the single most important structural signal in the 2026 US data-center policy landscape, and Cliff should productize it directly. The regulatory-risk score for a proposed site should have (a) statewide-legislative-action risk (medium probability, high impact — VA / IL / NY / OR-like actions), (b) statewide-ballot-action risk (empirically near-zero in 2026 without paid signature firms — Conserve Ohio just proved it), (c) county-commissioner-moratorium risk (HIGH and rising, per Linn County IA and Washington County MD landing the same day). Cliff should build the composite scoring model against those three tracks separately. Right now most site-selection consultants collapse them into a single 'regulatory risk' number, which materially misprices sites in states like Ohio (low ballot risk, low state-legislative risk short-term, but rising county-commissioner risk in Licking / Delaware / Union counties). (2) The Linn County IA moratorium-DESPITE-comprehensive-ordinance move is the most Cliff-relevant new precedent this week. Every county that has already adopted a data-center ordinance in 2025-2026 can now cite Linn County to justify a second-layer moratorium if developer proposal volume increases. That's a specific new failure mode for the 'we already have an ordinance, we're protected from friction' argument that developer-side counsel has been using. Cliff should update its site-diligence template to explicitly flag jurisdictions with existing ordinances AND high recent proposal volume as elevated moratorium risk — the two together are now the leading indicator, not either one alone. (3) The Ohio ballot failure is a durable structural fact worth compounding into the corpus. Every future 2027/2028 citizen-led statewide ballot ban attempt should be evaluated against the Conserve Ohio benchmark: 8 months of grassroots signature-gathering delivered ~20% of the required threshold in a large, politically motivated state. That's the baseline. Any future attempt needs to clear it via paid signature firms or fail similarly. Cliff should log this as the empirical ceiling for grassroots statewide ballot campaigns and use it to systematically deprioritize ballot-risk in favor of ordinance-and-moratorium risk in state-level modeling for the next 12 months.