Five weeks of the online debate over AI data centers, set beside the latest polls, ahead of 41 local votes on Nov 3.
Key finding: Polls show about 7 in 10 Americans oppose a data center nearby. Online, that opposition is moving to the town hall: engagement with local fights nearly doubled in five weeks, while power bills, voters’ top worry, faded.
About this series
Ballot Signals is a daily narrative intelligence series on 2026’s biggest ballot fights. Each day until Election Day, we publish a Rolli IQ report on one measure and set it beside the latest public polling. On Nov 4, we check the signals against the results. This is Report 1.
| Date | Measure and what it does | Our question |
|---|---|---|
| Thu Oct 29 | AI data centers Local votes to ban, pause or require approval for new data centers. |
Where is the opposition going? |
| Fri Oct 30 | California Prop 40 A one-time 5% tax on the wealth of Californians worth over $1 billion, mostly for health care. |
Support flipped from 52–46 to 46–52 in PPIC polls. Which arguments moved it? |
| Sat Oct 31 | Missouri Amendment 3 Would repeal the abortion-rights amendment voters passed in 2024, ban most abortions with exceptions, and ban gender-transition procedures for minors. |
Polls show a 43 to 43 tie. Which side’s message is landing? |
| Sun Nov 1 | Montana I-194 Would bar corporations, unions and other groups from spending to influence Montana elections, a test of the Citizens United ruling. |
There is no public poll. What does the conversation show? |
| Mon Nov 2 | California Prop 39 Would require voters to show ID, add ID numbers to mail ballots, and check voters’ citizenship. |
Voter ID polls well, yet the measure trails 42 to 51. Why the gap? |
| Wed Nov 4 | The results What the online signals showed, next to what voters decided. |
Did the signals hold? |
In one page
Polls tell you how many people oppose something. The conversation shows what they argue about, how they feel, and where the fight goes next. Here are five things it adds.
The poll and the conversation rank the worries differently
An Embold Research poll for Heatmap, fielded Oct 3 to 15 among 2,537 registered voters, asked people why they would oppose a data center nearby. Electricity bills led at 40%, followed by water bills and little transparency at 31% each, water pollution at 27%, and tax breaks at 19%. Online, over the same five weeks, water drew 43% of the engagement, local fights 25%, power bills 20% and tax breaks 12%. The two agree on water. The poll sees a bill worry the feed underplays. The feed shows where opposition goes next: into local meetings.
Data centers are on 41 local ballots
Ballotpedia counts 41 local data center measures on the Nov 3 ballot in 10 states, from outright bans and moratoria to rules requiring voter approval. Ohio has 21 of them. All six decided earlier in 2026 went the restrictive way. Nationally, 55% of voters want new AI data centers paused (19%) or prevented in most cases (36%), according to the same Heatmap poll.
From the power bill to the town hall
We split the conversation into four storylines: electricity bills, water, tax breaks, and local fights over specific projects. Water was the largest storyline in four of five weeks. Power bills spiked the week of Sep 29 on two posts, David Sacks hailing Trump’s AI summit and Sen. John Kennedy on a failed bill to make data centers pay their own power, then settled. Local fights grew steadily and had their biggest week of the five in the week of Oct 20.
Gallup found 48% of Americans strongly oppose a data center nearby. Online, that intensity now shows up in person. In Utah County, about 100 residents turned out against a data center ordinance. After the vote, a protester’s arrest went viral, and two posts about it drew 47,000 interactions. The story spread to Reddit and YouTube. The sheriff and the activists tell it differently.
Every storyline has its own emotional fingerprint
A sentiment score would call all four storylines negative. The emotion labels tell a more useful story. Tax breaks are almost pure grievance, even though only 19% of voters name them as a reason to oppose. Local fights are angry with a hopeful edge, as communities celebrate moratorium wins. Water carries the widest range. Power bills are the most positive storyline, because boosters and investors are part of that conversation.
Same fight, different rooms
Every platform shows a different side of the fight: senators and candidates on X and Bluesky, neighbors trading wins and losses on Reddit, local news and confrontation videos on YouTube, tribes and politicians on Threads, residents and city halls on Instagram, and big explainer pages on Facebook.
Before, during and after the poll
The Heatmap poll was in the field from Oct 3 to 15. It came after the power-bill spike, during the peak in tax-break talk, and before Utah County. Each moment raises a question for the next poll.
Five moves before Nov 3
For campaigns, communicators and anyone in a data center fight, on either side: answer the bill first, plan for the room and not just the feed, treat tax breaks as an accountability story, be specific on water, and watch Ohio.
Coming next
How to read a signal
A signal is not a poll. Online share is not vote share, and we set every finding beside public polling. Ballot Signals takes no side on any measure. What a signal can show is which arguments are gaining, how people feel about them, and who is driving them.
Want this read on your issue? Send us a topic and we’ll show you what Rolli IQ finds. Talk to Rolli →
Method and how sure we are. Rolli IQ, public posts from Sep 22 to Oct 26, 2026 across X, Bluesky, YouTube, Reddit, Facebook, Instagram and Threads. Weekly storyline trends are measured on X and Bluesky, where every post can be dated to its week: 20 weekly searches and about 1.8 million interactions on the top posts collected, with posts dated outside the week removed. Four keyword-defined storylines that can overlap; off-topic posts removed. The rise in local fights holds on X and on Bluesky separately. Emotion labels are Rolli IQ’s own, on the 235 most-engaged on-topic posts; tax-break anger is 89% (95% range 78 to 94%, n = 62). These are the most-engaged public posts each search returned, not a sample of voters. Polls: Embold Research for Heatmap (Oct 3 to 15, 2026, 2,537 registered voters), Gallup (March 2026), PPIC, Missouri Scout/Remington, Citrin Center/POLITICO. Ballot counts: Ballotpedia. Private accounts are described, not named.
