AI Data Centers Are Already Flipping Elections

A major outlet will name data-center electricity bills as the deciding factor in a flipped November 2026 general-election race — after the same backlash already ousted a 20-year Utah Senate incumbent in his primary.

Your electricity bill, your state legislature, and the nearest data-center construction site are all more connected than you think — and 2026 is the year that connection became visible.

Target: Dec 2027(531 days until resolution)

Assessed Probability

50%

More likely than not

Based on 3 expert predictions, 3 evidence items

Community Forecast

Cast your vote

Be the first to weigh in below

Your Prediction

Where do you think this lands?

Join others who've weighed in

5%95%

50% — More likely than not

Share on LinkedIn Share on X

Everyone's watching AI job-loss numbers. Almost nobody's watching the receipt showing up in mailboxes: data centers have already caused $23 billion in cumulative customer electricity price increases through 2028, with prices near some facilities up 267% over five years, and 78% of Americans now tell Gallup they're worried about their bill. That anger has already ended a political career — Utah State Senate President Stuart Adams, a 20-year incumbent, lost his June 2026 primary by 8-plus points over data-center opposition, and Festus, Missouri voters ousted half their city council after it backed a $6 billion proposal. Those are real, verified results — and they're also both lower-turnout, non-general elections, a genuinely easier bar for single-issue backlash to clear than a competitive November general. PJM's capacity auction tells the same story with harder numbers: it has now hit its $325/MW-day price cap for three consecutive years, with the reserve shortfall growing each time and data centers driving 63% of the increase. And Constellation is racing to restart Three Mile Island's Crane reactor by late 2027, the nearest-term test of whether nuclear can actually relieve the pressure. Three different mechanisms, one underlying story: AI's power bill is now a live political and economic fact, not a distant forecast.

Scenarios

Current value: 3-for-3 precedent at the primary/local level (Utah Senate primary, Festus MO city council, Georgia PSC), zero general-election precedent yet; PJM auction hit its price cap for the 3rd straight year (July 2026) with a 6,831 MW shortfall; Three Mile Island's Crane restart targets H2 2027

S-curve position: Early — real precedent exists at the primary/local level, but the harder general-election bar hasn't been cleared yet

Bear Case

The primary-level anger doesn't generalize to November turnout, PJM's shortfall narrows enough to avoid a 4th cap hit, and Crane restarts on schedule — the issue stays a local/primary phenomenon

Base Case

No named general-election flip yet, but PJM hits the cap a 4th time and Crane slips toward 2028 — the economic and infrastructure signals confirm the story even without an electoral trophy case

Bull Case

A clear general-election flip gets named by multiple major outlets, PJM hits the cap a 4th time, and Crane slips — the full trifecta of backlash-tax-nuclear-can't-save-us lands in the same window

How We'll Know

What we measure: Whether a November 2026 general-election result is publicly attributed by a major outlet to data-center/electricity-bill backlash, alongside PJM capacity-auction pricing and the Three Mile Island/Crane restart timeline as supporting economic and infrastructure signals
Confirmed if: A flipped November 2026 state legislative or Congressional seat is named by a major outlet as decided primarily by data-center/electricity-bill backlash
Refuted if: No such general-election race is reported by major outlets by mid-December 2026 — data centers remain a secondary talking point, not a decisive factor, even as the primary/local precedent (Utah, Festus, Georgia) continues
Data sources: AP / NYT / WaPo / Politico / Stateline post-election coverage
Ballotpedia race summaries
PJM Inside Lines auction results
Constellation Energy quarterly earnings and FERC filings

Evidence Trail

Evidence For

Jun 15, 2026
Newsweek: Utah State Senate President Stuart Adams (20+ year incumbent) lost his June 2026 Republican primary by 8+ points, driven by data-center opposition; Festus, MO voters ousted half the city council in an April 2026 off-cycle local election over a $6B data-center proposal.→ Probability: 45%
Jul 14, 2026
Fortune: $23B in cumulative customer electricity price increases through 2028, prices near some data centers up 267% over 5 years, 78% of Americans (Gallup) worried about their bill; the same week, PJM's 2028/2029 auction cleared at the $325/MW-day cap for a third consecutive time with a 6,831 MW shortfall.→ Probability: 50%

Evidence Against

Jun 1, 2026
The Hill: data centers are a live flashpoint in multiple 2026 gubernatorial races on both sides of the aisle, but no general-election result has yet been attributed to this factor by a major outlet — the claim requires a genuine first.

How Our View Evolved

Jul 18, 2026Initial assessment: 50%
Baseline — topic launched with the July 2026 second-iteration research sweep. Anchored below the original 0.70 pitch because every current precedent (Utah, Festus, Georgia) is a primary or off-cycle local race, not the November general the claim requires.

What Experts Say

What Could Go Wrong

Data-center backlash proves to be real but diffuse — present in every 2026 race as a talking point, decisive in none, because rising electricity costs get blended with broader inflation anger that voters and reporters alike attribute to the economy generally, not to data centers specifically.