Mary Peltola in an official congressional photograph from 2022. Archive photo by Office of U.S. Representative Mary Peltola (Public domain).

TAKEN SOCIAL PROJECTION · NOVEMBER 13, 2026

Taken Social projects that Democrat Mary Peltola will win Alaska’s U.S. Senate race, defeating Republican incumbent Dan Sullivan. The decision rests on three findings: her expanding advantage in the published first-choice returns, a lead that survives the transfer patterns visible in the scanned-ballot data, and a remaining-vote path that requires Sullivan to outperform the evidence now available. The numbers support a Peltola victory. This is Taken Social’s independent editorial projection of the eventual result; Alaska’s count and certification process continue.

The official statewide summary reviewed for this analysis, timestamped Nov. 12 at 6:36 p.m. Alaska time, reports 125,157 first-choice votes for Peltola and 119,357 for Sullivan. Her lead is 5,800 votes. She has 49.08% of the 255,010 valid Senate votes in that report, against Sullivan’s 46.80%. The other candidates and write-ins account for 10,496 votes. Those are the starting figures for the projection, rather than a poll or an estimate of voter enthusiasm. Alaska Division of Elections’ statewide summary.

The decisive analytical question is what happens to that advantage after additional ballots arrive and lower-ranked candidates are eliminated. Simply announcing the current leader would not answer it. Taken Social’s assessment is that Peltola has built enough room to absorb the transfer advantage Sullivan has demonstrated, while the evidence about the unfinished count favors maintaining or enlarging her position. The conclusion is a projected Democratic pickup, with the arithmetic supporting it set out below.

The reported lead has moved beyond its election-night scale

Alaska’s first returns left the Senate contest exceptionally close. Elections Archive’s update log records an initial Peltola lead of 626 votes and a subsequent small increase. Later additions changed the scale: Peltola gained a net 1,880 votes on Sullivan in the Nov. 9 update, another 1,733 in the Nov. 10 update, lost 275 in the Nov. 11 addition, and gained 1,825 in the Nov. 12 update. The resulting 5,800-vote advantage is more than nine times the initial margin. Review the dated vote-update log.

The Republican-favoring Wednesday batch is important to that assessment. Alaska Public Media reported that roughly 5,000 of the approximately 5,600 ballots added that day came from absentee voting in state House districts covering the Matanuska-Susitna Borough, a Republican stronghold. Sullivan narrowed the statewide gap, but only by 275 votes. That batch demonstrated an actual Republican recovery without bringing the race back to its election-night balance. Read the reporting on the Wednesday count.

A series of additions should not be extrapolated as if every later envelope contains an identical electorate. Alaska processes different regions and ballot categories at different times. Nevertheless, the observed direction has evidentiary value: several substantial additions have favored Peltola, and the identified Republican-heavy batch did not erase her accumulated gains. The projection weighs that sequence alongside the rankings, rather than treating a single favorable release as a statewide forecast.

Ranked-choice voting strengthens the case for examining real ballots

Peltola has not yet exceeded 50% of the first-choice vote in the official summary. If that remains true when counting is complete, ranked-choice tabulation will determine the winner. The ballots of eliminated candidates move to their voters’ next available preferences. Ballots without a continuing preference become inactive. Consequently, adding every vote for another candidate to the incumbent’s total would misread what voters actually marked.

FairVote has published an unofficial ranked-choice tally using the state’s preliminary cast vote record. Its Senate visualization shows Peltola beginning with 121,176 votes and Sullivan with 117,765. After the elimination rounds, Peltola has 123,651 to Sullivan’s 121,121, or approximately 50.52% to 49.48% of continuing ballots. She leads by 2,530 votes in that particular scanned-ballot subset. The transfer process narrows her advantage but leaves her ahead. Inspect the preliminary round-by-round tally.

The subset and the statewide summary must be kept distinct. The visualization contains 249,036 valid first-choice Senate votes, compared with 255,010 in the statewide report. FairVote explains that the cast vote record does not yet include every ballot represented in the published first-choice results, including hand-counted ballots. Its final-round totals therefore cannot be presented as Alaska’s completed statewide result. They reveal how the rankings in the available subset behave. Read FairVote’s explanation of the data coverage.

Those rankings are more informative than a party-label shortcut. In the visualization, the 10,095 ballots initially assigned to Heikes, Daniel J. Sullivan Jr. or write-ins eventually provide 2,475 votes to Peltola and 3,356 to the incumbent; 4,264 become inactive. Sullivan gains a net 881 votes through those transfers. That is a real advantage to account for, but it is substantially smaller than Peltola’s 5,800-vote first-choice lead in the fuller count.

Our transfer adjustment leaves Peltola ahead by approximately 4,884

Taken Social’s central calculation applies the observed transfer proportions to the 10,496 lower-candidate and write-in votes in the statewide summary. In the available subset, approximately 24.52% ultimately continue to Peltola, 33.24% to Sullivan and 42.24% become inactive. Applying those proportions to the larger pool produces an estimated net transfer gain for Sullivan of approximately 916 votes. Subtracting that estimate from Peltola’s first-choice advantage leaves a projected continuing-ballot lead of approximately 4,884.

This adjustment is a reproducible scenario, not an official tabulation. It assumes that rankings among the additional lower-candidate ballots resemble the observed subset. It does not count every unseen ranking or claim that all districts have identical transfer patterns. Its value is that it gives Sullivan the benefit demonstrated in the available evidence and tests whether that benefit overturns the current lead. Under that assumption, it does not.

There is also a simpler check on the argument involving the two candidates named Sullivan. Peltola’s 125,157 first-choice votes exceed the combined 123,556 votes of Dan S. Sullivan and Daniel J. Sullivan Jr. by 1,601. That comparison does not dispose of the Heikes ballots or replace ranked-choice tabulation. It does show that combining the two Sullivan totals would no longer put the incumbent ahead, even before considering that many voters for the other Sullivan ranked Peltola or no continuing candidate next.

What Sullivan would need from the remaining vote

The outstanding ballots must still be processed, and estimates of the eventual total can change. To avoid presenting an estimate as a ballot inventory, Taken Social tested several explicit volumes of additional votes that ultimately count for either Peltola or Sullivan. These are continuing votes after transfers, not a claim that precisely that many physical ballots remain. The distinction allows the reversal threshold to be stated clearly.

With an estimated existing continuing-ballot lead of 4,884, Sullivan would need approximately 59.77% of 25,000 additional continuing votes to reach a tie. With 35,000 such votes, he would need 56.98%. With 45,000, he would need 55.43%. Those are substantial Republican advantages across the unfinished vote. They are mathematically possible, but the evidence reviewed does not make them the expected outcome.

The stress test is deliberately demanding for Peltola. A 55–45 Sullivan advantage among 35,000 additional continuing votes would still leave her ahead by about 1,384. An even split would preserve the approximately 4,884-vote advantage. A 55–45 Peltola split would enlarge it to about 8,384. The exercise identifies how much of a change in the remaining electorate is required, rather than assuming a lead must survive merely because it exists.

Larger volumes make a reversal easier: with 60,000 additional continuing votes, Sullivan would need about 54.07% to tie. That scenario is included in our calculations as an adverse test, not a forecast of the outstanding count. The case for the projection depends on the combination of volume, geography and rankings. Taken Social is not assigning a statistically calibrated win probability from these scenarios; the decision is an editorial judgment grounded in their results.

Why the remaining-vote evidence favors the projection

FairVote’s current analysis, drawing on reporting about the unfinished count, says Peltola is expected to receive most first choices among the remaining ballots. Alaska Public Media also reported the Peltola campaign’s expectation of favorable outstanding votes in Anchorage and rural Alaska. The campaign’s statement is an interested party’s assessment. It becomes more useful when considered alongside the independent ballot analysis and the actual pattern of later additions. Read the remaining-ballot analysis and its assumptions.

The incumbent’s campaign has emphasized that uncounted ballots remain and called for a complete, accurate count. That position is consistent with the continuing election process. It does not, by itself, demonstrate the Republican margins required in the sensitivity analysis. A projection asks which outcome the evidence supports before certification; election officials must still count every eligible ballot regardless of any newsroom’s assessment.

Our conclusion therefore does not rest on one campaign’s confidence. Peltola leads in the statewide tally, leads after transfers in the available scanned-ballot subset and retains an estimated advantage when those transfer patterns are applied to the fuller count. Sullivan would need a sufficiently large and substantially more Republican remainder to reverse that position. The reporting about the remaining electorate points in the opposite direction.

The conclusion: Taken Social projects a Peltola victory

Alaska lists Nov. 18 for ranked-choice tabulation in races requiring it and Nov. 25 as its target certification date. The official summary’s 100% of precincts reported does not mean every eligible absentee or questioned ballot has been counted. Those administrative milestones remain important, but they do not prevent an independent assessment of the eventual winner based on the evidence already available. Check the Division of Elections’ published schedule.

Peltola’s projected victory would add another Democratic pickup to a Senate election in which the party has already secured control. NPR’s report on Scott Colom’s Mississippi victory put the incoming Democratic majority at 55 seats with Alaska unresolved; a Peltola win would raise that total to 56. That is the national consequence of a contest whose decisive evidence remains local: counted Alaska votes and the preferences written on Alaska ballots. Read the reported Senate balance after Mississippi.

Taken Social’s call is Mary Peltola to win. The lead has grown beyond its election-night scale, the observed transfer advantage for Sullivan is insufficient to overturn it, and the remaining-vote thresholds favor the Democrat. We will update the report if new official data materially changes those inputs. On the published numbers and the ranked-ballot evidence reviewed Nov. 13, Peltola is the projected winner of Alaska’s 2026 U.S. Senate election.