46 of 47 decided data-center measures favored greater limits. Read the row before you site one.

On November 3, 2026, local ballots asked voters about data centers, which Ballotpedia describes as physical facilities that house the computer systems and related equipment used to store, manage, process, and transmit digital information. Its tracker says 41 of at least 50 local data-center measures in 2026 were scheduled for that general election. As of November 4, the same page says voters had decided 47 measures, and in 46 a majority favored greater limits. A siting choice that uses this count touches public policy, land, and money. The rule for that move is a named human reviewer before release. This post is the new count.

This reading of Ballotpedia’s tracker through the Of One method uses only the tracker’s figures.

What the tracker counted

Ballotpedia is tracking at least 50 local measures in 13 states: Arkansas, California, Florida, Illinois, Kansas, Maine, Michigan, Nevada, New Hampshire, New Jersey, North Dakota, Ohio, and Wisconsin. Ohio accounts for the largest share, 21 of 50 (42%), across 18 distinct jurisdictions. Forty-one of the 50 (82%) were scheduled for November 3, six were earlier in 2026, and three are later.

As of November 4, the scorecard lists one outcome favoring development, labeled 2% of decided measures, and 46 limiting development, labeled 98% of decided measures. Both are shares of measures.

The page states the 46 this way: “In 46 cases, a majority favored greater limits on data center development, either by approving restrictions or requirements or by rejecting a measure that would have facilitated or supported development.” The other decided case is Harwood, North Dakota. Voters there rejected additional review requirements for certain industrial developments, such as data centers. The November 3 row is an initiative to require that review before city approval of industrial developments, including data centers. Ballotpedia marks it defeated, 105 yes (24%) and 325 no (76%).

Yes-no results appear on the November 3 rows and on the six earlier dates. The three later measures have none yet.

Same score, different rows

The page sorts measures into two groups. One would allow or facilitate data centers, or advise officials to support them. The other would limit or prohibit them, advise officials to oppose them, or add requirements.

Alhambra, California, Measure DB would prohibit data centers in the city. Ballotpedia marks it approved, 12,496 yes (82%) and 2,772 no (18%).

Bloomington, Illinois, asked voters to advise the city to impose a moratorium on building large-scale data centers. Ballotpedia marks that advisory measure approved, 21,851 yes (78%) and 6,316 no (22%). Whether the city imposed that moratorium is absent from the row, so it stays a gap.

A decision map keeps the pieces apart. The count is evidence. The sentence you act on is a claim, and it has to match the row. A missing piece is an unknown. Every question you skip becomes a guess treated as a fact.

Four lines for one row

Run this on any row this week. Boulder City, Nevada, Question 1 is the example. Ballotpedia lists it as a November 3 referral to allow data centers within the Eldorado Valley Transfer Area, marked defeated, 1,973 yes (29%) and 4,726 no (71%).

1. Evidence. Copy the place, the date, the measure’s description, the result, and the yes and no counts.
2. Claim. Write one sentence the row supports: voters defeated a measure that would have allowed data centers in the Eldorado Valley Transfer Area.
3. Kill test. Write the later public result that would prove that sentence wrong. A corrected canvass, or a Ballotpedia correction, showing Question 1 passed, or that it did not allow data centers in the Eldorado Valley Transfer Area, would kill it. Every strong claim on an Of One map carries the observation that would falsify it.
4. Gate. Name the person who signs before money is spent, land use changes, or someone takes a step that cannot be undone. Anything you cannot undo waits at a human gate. Public policy requires a named human reviewer. A blank name stays a gap.

The answer a reader sees is a rendering of that map. The rendering states the move the evidence best supports on the criteria you set.

Three votes are still open

On December 15, Mandan, North Dakota, votes on an initiative to prohibit data centers with an anticipated electrical capacity or demand of 25 megawatts or more.

On December 1, Clarksville, Arkansas, votes on two referrals. Ballotpedia describes one as an uphold-or-repeal vote on Ordinance 26-1007, authorizing up to $50 billion in taxable industrial development bonds to finance data center equipment and improvements, leasing the project to Hatchbo LLC, and authorizing a payment in lieu of taxes agreement. It describes Ordinance 26-1008 as an uphold-or-repeal vote authorizing up to $5 billion in taxable industrial development bonds to finance land acquisition for data center operations, leasing the project to SF ARK1, LLC, and authorizing a payment in lieu of taxes agreement. Those dollar figures are the authorization ceilings in that description.

For those cities, December is the timing line. A result then is new evidence, and new evidence patches only the part of the map that depends on it.

What to do this week

• Open Ballotpedia’s 2026 data-center page and copy the row for the place you are weighing: date, description, result, and the yes and no counts.
• Write the one-sentence claim that row supports, and put the kill test beside it.
• Where the measure only advises officials, add a gap for whether the board adopted the advice.
• For Mandan or Clarksville, put December 15 or December 1 on the timing line before you treat 46 of 47 as the local result.
• You can walk one decision through the eight starting questions with the free tool at of1.ai/ask. It runs in the browser and does not use AI.

Consult is time with Utlyze running Ask, Map, Move on your own decision, live. It is $400 per hour, in 5-hour blocks, and you leave with the map and a first move you can undo.