U.S.-China Relations
NextFin News - The Justice Department has escalated a criminal investigation into Alaska's Senate race, issuing more than 40 grand jury subpoenas and record preservation notices to key targets a day after a major national newspaper published a report alleging that associates of Democratic candidate Mary Peltola helped recruit a look-alike candidate to siphon votes from Republican incumbent Dan S. Sullivan. The move, less than a week before Election Day, puts federal prosecutors at the center of one of the cycle's tightest contests, where polling averages show Peltola ahead by just over a point in a state Donald Trump carried by double digits two years earlier.
The stakes are immediate and concrete. Alaska's ranked-choice system means the general election could take weeks to certify, and the Justice Department has already filed suit to block certification if Sullivan loses and the second Dan Sullivan's votes would have changed the outcome. Attorney General Todd Blanche says the department is only seeking to preserve evidence; critics say it is laying the groundwork to overturn a result it does not like. Either way, the department that would have to prosecute any election crime is now a party to the dispute over whether the election was fair.
The Two Sullivans and a Ballot the State Tried to Block
The investigation centers on Daniel J. Sullivan, a retired schoolteacher and former U.S. Forest Service employee from Petersburg, Alaska, who is not related to the incumbent senator. In May 2026, shortly before the June 1 filing deadline, he submitted a declaration of candidacy to run as a Republican for the Senate seat. He had previously registered to vote as Daniel J. Sullivan Jr. and changed his party affiliation to Republican in connection with the run. He sought to appear on the ballot under a name matching the incumbent's, requesting to be listed as "Dan Sullivan" and then "Dan S. Sullivan," adopting the middle initial of the senator he was challenging.
On June 15, Alaska Division of Elections Director Carol Beecher determined that his declaration "was not filed in good faith for the purpose of genuinely pursuing election" but was instead filed "with a purpose to confuse or mislead and to thereby compromise the ballot's fairness or neutrality." Beecher called the circumstances "unique, and to my knowledge utterly unprecedented facts," citing his voter registration, recent party change, ballot-name request, campaign website design, and his engagement of a political consultant whose prior clients included Democratic candidates. Her determination did not cite evidence of direct coordination with Democrats.
The state's disqualification did not stick. On June 26, Alaska Superior Court Judge Thomas Matthews ruled that the Division of Elections had abused its discretion by applying a previously unstated "good-faith" criterion not grounded in the Constitution, Alaska statutes, or the division's own regulations, and ordered Dan J. Sullivan placed on the primary ballot. He survived the August 18 top-four primary with just over 3,000 votes, about 2.5% of the total, behind Peltola's 49.5% and the incumbent's 41.4%. When fourth-place finisher David Leslie withdrew on August 31, fifth-place Republican Gerald L. Heikes advanced instead, leaving four candidates on the November 3 ballot: Peltola, Dan S. Sullivan, Dan J. Sullivan listed with no party affiliation, and Heikes.
Why does 2.5% matter? Because Alaska's general election uses ranked-choice voting. If no candidate wins a majority of first-choice votes, the lowest candidate is eliminated and those ballots redistribute to the next-ranked contender still in the race. In a contest polling within the margin of error, the second Sullivan's share of the vote, and where his supporters' second choices flow, can decide the seat.
A Slow-Moving Probe Suddenly Accelerates
Federal and state investigators began examining the candidacy around the same period as the court fight. The FBI, the U.S. Attorney's Office for Alaska, and the Alaska Attorney General's office looked into whether individuals conspired to create the campaign to confuse voters, harm the incumbent, and assist Peltola. The federal inquiry has focused on potential wire fraud or a conspiracy to deprive Alaska voters of a free and fair election process, which could constitute a civil rights violation. The probe had been moving slowly, according to people familiar with the matter.
That changed on October 28, when a major national newspaper published a report alleging that Peltola campaign aides and lawyers, along with staff at the Democratic Senatorial Campaign Committee, were involved in pushing other candidates, including Daniel J. Sullivan, to run. The report said its findings rested on 12 sources, text messages, and documents. It did not directly establish that Peltola personally participated in or directed the recruitment.
A key source was Elisa Rios, a senior campaign adviser and former campaign manager fired by Peltola in August during a campaign shake-up. Rios described the Peltola campaign's involvement to the newspaper, then contacted it the following week to retract her statements. "As discussed on our call earlier today, I am confirming that I have retracted all statements made regarding the 2026 Alaska Senate Race in connection with any article you may or may not be writing," Rios wrote. "I am not willing to be a source for any story in any capacity." The newspaper published its report anyway.
The effect inside the Justice Department was immediate. Federal prosecutors issued more than 40 grand jury subpoenas and record preservation notices to key targets by that evening, with additional subpoenas expected. Michael Heyman, the U.S. Attorney in Alaska, who had recused himself early in the criminal investigation in part because he was in discussions with Senators Sullivan and Lisa Murkowski for a judgeship nomination, un-recused Wednesday afternoon at the request of department leadership. "Suddenly the options in the criminal case became much wider," a department official said Thursday.
Prosecutors are now examining potential civil rights charges, including a deprivation-of-rights count more typically brought against law enforcement officials for wrongdoing. The investigation has reportedly roiled the U.S. Attorney's office in Alaska, where some career prosecutors have balked at aspects of the probe and whether it broke federal law or simply represents electioneering dirty tricks. A department official denied the alleged backlash.
Peltola's campaign and the DSCC deny involvement. "The Alaskans for Mary campaign has no involvement with either Sullivan campaign," a campaign spokesperson said. "Like Mary, our campaign is focused on lowering costs and fixing the rigged system in DC." Legal experts note that recruiting candidates to run is not, by itself, illegal.
The Lawsuit That Came First
The criminal escalation followed a separate, already-filed lawsuit. On October 22, the Justice Department sued Alaska in federal court, asking a judge to declare any certification of the Senate race "unlawful" and to bar the state from certifying if Senator Sullivan loses and the second Dan Sullivan's votes under ranked choice would have been enough to change the outcome. The complaint argues any result would be "tainted by the inclusion on the ballot of Daniel J Sullivan."
On Sunday, October 25, Blanche defended the timing on network television. The lawsuit, he said, is "not saying, 'don't certify.'" Its purpose is to "put guardrails in place" so that if a court later determines the race was decided "because of confusion, and not the will of the voters," the state would not certify. The relief sought, he added, "isn't to determine who the senator is... it's just to make sure that the ballots that come in are kept in a way that may allow a court to investigate further."
The department's stated purpose, Blanche said, is to ensure election officials "have the information necessary to make a determination as to whether Alaskan voters were confused or mislead."
President Trump has said he directed the lawsuit, telling reporters "I'm the one who said do it." Blanche declined to discuss his conversations with the president. Alaska's Republican-led elections office is pressing the court to dismiss the case quickly, arguing it seeks to sow "doubt" and will "cast a cloud over" the contest. Peltola has asked a federal judge to dismiss the lawsuit, with her lawyers calling the complaint "wholly baseless."
Alaska Governor Mike Dunleavy, a Republican, announced late Wednesday that the state would open its own probe into whether the alleged recruitment efforts violate state law. Alaska State Troopers said they are committed to "impartial enforcement of Alaska's criminal laws" as the investigation begins.
What the Polls - and the Models - Are Pricing In
The political stakes are unusually high because the race is genuinely close in a state Trump won in 2024 by double digits. A polling average of 26 surveys puts Peltola at 46.9% and Sullivan at 45.8%, a Democratic edge of 1.1 points. The most recent survey, a poll of 694 likely voters conducted October 16 by a Democratic-aligned firm, has Peltola ahead 47% to 44%. A separate survey by a national newspaper and a university, fielded from late September into early October among 504 likely voters, showed a wider Peltola lead, 50.1% to 43.1%.
Forecast models disagree sharply on what those polls mean. One model, updated October 24, favors Republicans, estimating a Sullivan win in 76 of 100,000 simulated elections with an estimated margin of R+6.1. An election-data firm, by contrast, gives Peltola a 61% chance of winning. A nonpartisan election analyst rates the race a toss-up. One reason for the divergence: ranked-choice voting makes the terminal outcome depend not just on first-choice support but on the elimination order and the second-choice preferences of Heikes and the second Sullivan's voters.
The Senate math amplifies everything. Republicans hold 53 of 100 seats, with 35 up in the 2026 midterms. Roughly seven Senate elections are considered toss-ups. A single-seat flip in Alaska could matter if the chamber ends up narrowly divided.
The Second-Order Question: Who Decides When an Election Is Legitimate?
The first-order story is straightforward: federal prosecutors are investigating whether Democrats ran a dirty trick, and they have sued to hold certification hostage if the trick worked. The second-order story is what that does to the institution that must later certify the result.
Alaska's election officials cannot declare a winner quickly under ranked choice even in normal years; the state's review and certification process can take weeks. Now the Justice Department has asked a federal court to stand ready to declare that certification unlawful. The practical effect is that the loser has a pre-cleared legal theory to contest the outcome, and the winner may take office under a cloud that lasts well past inauguration day.
That dynamic is not confined to Alaska. Justice Department officials defend the Dan Sullivans investigation, a highly visible search of Representative Ilhan Omar's home, criminal charges against prominent critics of the president, and voter-roll challenges in California and other states as normal, time-sensitive law enforcement. They frame the election-related work as intended "to safeguard the votes, not sway the election." But when the same department that brings charges also publicly argues, days before the vote, that a particular ballot configuration is illegitimate, the line between safeguarding and swaying blurs for the voter trying to decide whether the count will be believed.
Is this cyclical or structural? The dirty-trick itself, if proven, is cyclical: a discrete act of electioneering that courts and prosecutors can remedy case by case, the way Alaska's own elections director tried to do in June. What is structural is the precedent being set for how the federal government intervenes in state election administration inside the final week of a campaign. A lawsuit filed less than two weeks before Election Day, a criminal probe accelerated by a newspaper report, and a recused U.S. Attorney pulled back onto the case are not the machinery of routine enforcement. They are the machinery of a new norm, and norms, once moved, do not move back on their own.
The Strongest Case Against the Alarm
The counter-thesis deserves its full weight. Recruiting candidates is legal. The newspaper's report does not establish that Peltola personally directed anything. The state's own attempt to remove the second Sullivan was struck down by a state judge as lacking legal grounding. And the Justice Department's lawsuit, read narrowly, asks only for evidence preservation, not for a court to pick a winner. From this angle, the department is doing exactly what a law-enforcement agency should do when presented with evidence of possible voter confusion: move quickly, secure records, and let the courts sort out the ballot question after the votes are counted. Blanche's insistence that the relief "isn't to determine who the senator is" is, on its face, a modest ask.
That defense holds only if the sequence reads as neutral. It becomes harder to sustain when the president publicly claims credit for ordering the suit, when the attorney general defends it on television days before the vote, and when the criminal investigation accelerates inside the same week against a campaign whose candidate is currently ahead in the polling average. The strongest evidence for the counter-thesis is also the weakest in court: timing. Prosecutors do not have to prove intent to confuse to bring some charges, but they do have to prove a crime occurred, and so far the public record contains allegations, a retracted source, and a civil determination that a state court already rejected.
The signal that would falsify the structural-alarm view is specific and observable: if the Justice Department files no charges before the election, drops the certification lawsuit without seeking an order, and the ranked-choice count proceeds to a certified result that both parties accept, then this episode reads as aggressive but ultimately conventional enforcement. If instead charges arrive before November 3, or the department wins an order blocking certification, the precedent is set.
What to Watch Next
Three timelines now run in parallel. First, the criminal track: more subpoenas are expected, and prosecutors are weighing civil rights charges. Any indictment before Election Day would roil the race and hand Republicans a powerful closing argument; no charges before the vote would let the story fade but leave the lawsuit alive. Second, the civil track: the federal court's response to Alaska's motion to dismiss the certification suit, and whether the judge grants any relief that delays the count. Third, the count itself: Alaska's ranked-choice tabulation could take weeks, and any result in which the second Sullivan's redistributed ballots change the winner will be litigated regardless of what the Justice Department does.
Short term, the beneficiary is ambiguity. The incumbent's camp gets a closing-week theme of stolen votes; the challenger's camp gets a theme of a weaponized Justice Department. Medium term, whoever wins inherits a contested mandate. Long term, the structural question is whether the 2026 Alaska case becomes a template for federal intervention in close state races elsewhere.
Base case: no pre-election charges, the certification lawsuit is dismissed or narrowed, and Alaska's ranked-choice count eventually certifies a winner that the losing side contests but cannot overturn. Upside case for the alarm thesis: prosecutors secure cooperation from a Peltola aide or DSCC staffer, bring charges before November 3, and a court blocks certification pending trial. Downside case for the alarm thesis: the retracted source and the lack of direct evidence of Peltola's involvement leave the criminal case without a provable conspiracy, and the lawsuit is dismissed as an impermissible federal intrusion into state election administration.
The deeper story is not whether one retiree's candidacy was a prank or a plot. It is that the federal government has inserted itself into the counting of a Senate race before the counting has happened, and the precedent will outlast the verdict.
Explore more exclusive insights at nextfin.ai.
