Lawsuits did not end when voting ended on November 3, 2026. In American elections, the period after election day is often the period when courts matter most, because that is when counting rules, ballot eligibility and certification duties are tested against real ballots and real margins rather than against hypothetical ones. Some of the most consequential disputes of the 2026 midterms were filed before election day and are still moving through the courts afterward. Others arise only after the count, when a close margin triggers a recount or a losing candidate files an election contest. Understanding which kind of case is which, and what remedy a court can actually grant, is essential to judging claims about the outcome.
This article does not treat a lawsuit as a verdict, an allegation as a finding, or a request for emergency relief as a ruling on the merits. It separates cases that courts have already decided from cases that remain pending, explains the legal standards judges apply, and sets out the arguments on both sides where a dispute is contested. Where a claim is unproven, it is labelled as an allegation. Where a post-election filing has not been confirmed in a reliable source, it is not invented here. The picture that emerges is narrower, and more legally grounded, than the sweeping claims that often circulate in the first days after an election.
2026 midterm election lawsuits: decided, pending and merely filed
The first distinction to draw is between a filing, a temporary order and a final judgment. A filing is a request. It alleges facts, cites law and asks for relief, but it proves nothing by itself. A temporary restraining order or preliminary injunction is a short-term measure, granted when a judge thinks a party may suffer irreparable harm before a full hearing, and it can be reversed after fuller argument. A final judgment, after briefing and often after an appeal, decides the legal question presented, subject to further appeal. News coverage that treats all three as equivalent will inevitably overstate what a court has done.
A second distinction is between federal and state courts. The Constitution assigns the administration of elections primarily to the states, subject to federal law. That division shapes almost every post-election case. A federal court can decide whether a state practice violates federal statutes such as the Voting Rights Act, the Civil Rights Act or the National Voter Registration Act, or whether a federal agency has authority to intervene in a state count. A state court, by contrast, is usually the forum for recounts, election contests and disputes about state election statutes and deadlines. A party that loses in one system cannot simply refile the same claim in the other and expect a different result.
A third distinction is between a challenge to a rule and a challenge to a result. A rule challenge asks whether ballots of a certain kind should be counted, whether a deadline should be extended, or whether an official has a duty to certify. A result challenge, often called an election contest, asks whether a certified outcome should be set aside because of an error, irregularity or illegal vote that could have changed the outcome. Courts treat those questions differently, apply different burdens of proof, and grant different remedies. A court that refuses to block the counting of a category of ballots has not thereby declared a winner, and a court that orders a recount has not thereby found fraud.
California: the ballot challenge courts rejected before and after voting
The largest pre-election ballot dispute to carry into the post-election period concerned California. The United States Department of Justice sued to require California officials to set aside more than 20,000 ballots associated with voters the department identified as potentially lacking United States citizenship. The department argued that eligibility should be verified before counting, and that counting ballots from ineligible voters would dilute lawful votes. Supporters of the federal position described the action as a safeguard for election integrity and a necessary check on voter rolls.
Critics, including state officials and voting-rights advocates, argued that the federal list was unreliable and that segregating the ballots risked excluding votes cast by eligible citizens. An analysis cited in litigation, prepared by researchers associated with a California Democratic Party-affiliated group, found that thousands of names on the federal list could not be matched to entries in the state’s voter rolls, and that the list included at least some confirmed citizens. Those findings were presented as evidence that the list could not safely be used to hold ballots aside, although the department disputed the weight to be given to that analysis and maintained that verification was required by federal law.
Courts sided, at the emergency stage, with the state position. A federal district judge in California declined to order the ballots set aside, reasoning that the proposed segregation would probably violate federal voting statutes. A panel of the Ninth Circuit Court of Appeals upheld that ruling, holding in substance that the department had not identified federal authority allowing it to take charge of, or intervene in, a state’s administration of an ongoing election. A separate federal judge in San Francisco issued an emergency order barring federal demands that California election workers remove named voters from the rolls or refuse to count specific ballots, reasoning that workers who administer the election as state and federal law require do not violate federal criminal law merely because a particular ballot is alleged by the administration to belong to an ineligible voter.
The administration then sought emergency relief from the Supreme Court, asking the justices to compel California to set aside the disputed ballots. The request arrived on November 2, hours before election day. The Supreme Court rejected it within hours in a brief, unsigned order, recording no dissent and giving no explanation, as is common for emergency applications. That denial did not decide every underlying legal question about voter-roll maintenance or eligibility verification, and it did not prevent the department from pursuing other theories in other cases. What it did decide, for the 2026 midterms, was that California would not be required to segregate those ballots on the department’s emergency timetable. Counting in California therefore proceeded under state procedures, subject to the state’s own certification deadline of December 11.
What the California cases cannot do is also important. They do not, by themselves, determine control of the House or Senate, and they do not establish that any particular ballot was cast unlawfully. The allegation that noncitizens voted remains an allegation where it has not been proved ballot by ballot in court, and the counter-allegation that the federal list was inaccurate remains a contested claim about a dataset, not a finding about every name on it. Readers should be sceptical of any summary that converts either side’s argument into a proven fact.
Alaska: a certification lawsuit built around two similar names
A second dispute that will shape post-election discussion concerns Alaska’s Senate race. The Department of Justice sued to prevent state officials from certifying the Senate result if the Democratic candidate, Mary Peltola, won, arguing that voter confusion could arise because another candidate on the ballot, Dan J. Sullivan, bears a name closely resembling that of the Republican incumbent senator, Dan Sullivan. The department’s argument, as reported, was that the similarity created a risk that voters intended one candidate and marked the other, and that certification should therefore be held back pending a resolution of that risk.
State officials and critics of the suit argued that certification is a mandatory, ministerial duty once the canvass is complete, and that a speculative risk of confusion, without proof that a sufficient number of voters were actually misled to change the outcome, is not a lawful basis to withhold a certificate. They also pointed to Alaska’s ranked-choice system, under which voters rank candidates and ballots are reallocated in rounds, as a further reason why a simple claim of name confusion does not map neatly onto the final result. Supporters of the federal position countered that the unusual similarity of the names justified judicial scrutiny before a result became final, and that certification should not be treated as automatic where a genuine question about voter intent exists.
As of the time of writing, no court ruling confirming that Alaska’s certification has been blocked, or that it has been allowed to proceed on a particular date, has been verified for this article, and no certified Alaska Senate outcome is claimed here. Alaska’s state certification date is November 25, and its ranked-choice count, together with mail ballots postmarked by election day and received within ten days afterward, means the final count itself takes time. Any reporting that Alaska’s result has already been finally decided by a court should be checked against a court order or a statement from the Alaska Division of Elections, not against a campaign statement or a social-media post.
The Alaska case illustrates a broader principle that courts apply in certification disputes. Judges are reluctant to enjoin certification on the basis of speculation alone, because certification is the mechanism by which voters’ choices become legally effective. A party seeking to delay or block certification typically must show a concrete legal violation, a realistic prospect that the violation affected the outcome, and a remedy that a court can grant without disenfranchising lawful voters. Allegations about confusion, eligibility or procedure, without that evidentiary link, rarely meet that standard at the emergency stage, although they may still be litigated on a fuller record.
Recounts and election contests: the cases that arise from close margins
Separate from the California and Alaska disputes are the recounts and contests that arise in individual races because the margin is small. Recount law varies sharply by state, and that variation determines what happens next in a close contest. In about half the states, a recount is triggered automatically when the margin falls inside a statutory threshold. In most states, a trailing candidate may also request a recount, sometimes only within a short window and sometimes at the candidate’s expense if the margin is outside the automatic threshold.
The most common automatic trigger is a margin of half a percentage point or less, used for relevant races in states including Arizona, Florida, Pennsylvania and Alabama. Ohio uses a quarter of a point for statewide offices and half a point for district races. Oregon uses two-tenths of a point or a tie. Michigan uses two thousand votes or fewer in a statewide race. Georgia, under its new 2026 law, automatically orders a manual recount for the first or second statewide executive races on the ballot when the margin is half a point or less, while Senate, House and most other races in Georgia require a written request within two business days after certification. Wisconsin allows a requested recount within one point, free to the requester at a quarter of a point or less. North Carolina allows a requested recount for statewide races at half a point or ten thousand votes, whichever is less.
A recount, whether automatic or requested, is not an investigation into fraud. It is a re-tabulation, by machine or by hand, of ballots already cast. It can correct a tabulation error, resolve a ballot that a machine could not read, or confirm that the original count was accurate. It cannot add votes that were never cast, and it cannot change the eligibility rules that governed which ballots were counted in the first place. That is why recounts typically move totals by small amounts, and why a recount is most consequential only when the original margin is already tiny.
An election contest goes further than a recount. Filed after certification in most states, a contest asks a court to set aside the certified result, order a new count under different rules, or in extreme and rarely granted circumstances, order a new election. The contestant usually must prove not merely that an error occurred, but that the error was material: that it affected enough votes to cast doubt on, or change, the outcome. Courts apply that materiality requirement strictly, because setting aside a certified election disenfranchises voters whose ballots were lawfully cast. A contest that alleges irregularities in general terms, without tying them to specific ballots or a specific margin, is therefore likely to be dismissed, while a contest that identifies a defined set of ballots and a defined legal error stands a better chance of a hearing.
Deadlines for contests are short and unforgiving. In many states, a contest must be filed within days of certification. A candidate who waits until the political debate has moved on may find the courthouse door closed, however strong the underlying complaint. That is one reason campaigns retain election lawyers before election day: the post-election calendar is not an extension of the campaign calendar, and a missed deadline is usually fatal to the claim.
What courts have decided about federal power over state elections
Taken together, the California rulings and the broader litigation over voter rolls point to a legal theme that will outlast the 2026 midterms: the limits of federal power over state election administration. The Constitution gives states the primary role in prescribing the times, places and manner of congressional elections, while Congress may make or alter regulations for federal elections. Federal statutes add further protections for voters, including prohibitions on denying the right to vote on account of race and requirements for voter registration procedures. What federal law does not generally do, as the Ninth Circuit panel reasoned in the California case, is authorise a federal agency to take over a state’s ongoing count or to direct state workers to segregate or discard categories of ballots on the basis of a federal list alone.
The administration’s broader voter-roll litigation provides context. The department sued dozens of states and the District of Columbia seeking voter-roll data, including sensitive information such as partial Social Security numbers, and most states declined to provide it in the form requested. Of the cases decided before the midterms, the department lost the overwhelming majority. Those losses do not prove that voter rolls are flawless, and they do not resolve the policy debate about how rolls should be maintained. They do show that courts have been sceptical of federal demands that would override state procedures without a clear statutory basis, and that scepticism has carried into the post-election cases.
That scepticism cuts both ways. State officials who refuse to certify on the basis of a general distrust of the process, without a specific legal violation, have also faced court orders compelling certification in recent cycles. Election-law researchers have documented dozens of instances in which local officials refused or threatened to refuse to certify results. Courts have generally treated certification as a mandatory duty, not a discretionary judgment about whether officials personally trust the outcome. A local board that finds a genuine, documented error in its canvass may correct it; a board that withholds certification as a protest against a result it dislikes is likely to be ordered to certify. That symmetry is important: the rule that protects a result from federal interference also protects it from local refusal.
What litigation can and cannot change before January 2027
Before the new Congress is seated on January 3, 2027, litigation can change three things. First, it can change which ballots are counted, if a court holds that a category of ballots was wrongly included or excluded and orders a remedy that can be applied to identifiable ballots. Second, it can change the timing of certification for an affected race, if a court orders a recount, a further canvass step, or a pause while a legal question is resolved. Third, in a certified election contest, it can change the declared winner of a specific race, if the contestant proves a material error that affected the outcome.
Litigation cannot, on its own, change the overall balance of power in Congress unless it changes enough individual races to move a majority, and that would require successful contests or recounts in multiple seats, each proved on its own record. It cannot create votes, extend voting beyond the deadlines set by law, or retroactively rewrite eligibility rules for ballots already cast without a court finding that the rules as applied violated the law. It cannot delay the constitutional start of the new Congress’s term, which is fixed by the Twentieth Amendment at noon on January 3, although it can leave a particular seat vacant or contested when that term begins, with the House or Senate itself then judging the qualifications and election of its own members.
That last point is often missed. Under the Constitution, each house is the judge of the elections, returns and qualifications of its own members. A court decision is therefore not always the final word on who sits in Congress. The House or Senate may conduct its own review of a contested election, although it typically gives great weight to a state certification and to court judgments. A seat that remains in litigation in January may therefore be seated provisionally, held open, or decided by the chamber itself, depending on the posture of the case and the rules of the chamber. Voters should not assume that a lawsuit filed in November automatically prevents a member-elect from taking the oath in January, nor that taking the oath ends all legal challenges to the election.
How to read the next round of headlines
As recount deadlines pass and contest filings appear, readers will see a new wave of headlines claiming that an election has been stolen, saved, overturned or confirmed. The most reliable test for any such headline is to ask four questions. Who filed the case, and what exactly did they ask for? What did the court actually order, as opposed to what a party said it would order? Does the order change any counted vote, or does it merely preserve the status quo while a fuller hearing is held? And is the race in question decided by a margin small enough that the remedy sought could plausibly change the outcome?
A headline that answers none of those questions, but quotes a campaign’s characterisation of a filing, should be treated as advocacy, not reporting. A headline that cites a court order, names the judge, and quotes the operative language of the order is on firmer ground. A headline that claims a statewide result has been reversed should be checked against the state election office’s certified totals, because only that office can issue the certificate that makes a result official.
The post-election legal story of the 2026 midterms is, at its core, a story about boundaries: between projection and certification, between allegation and proof, between federal and state authority, and between a court’s power to correct an error and its reluctance to override the expressed will of voters. Those boundaries are not technicalities. They are the reason a close election can be recounted without being relitigated from scratch, and the reason a losing candidate can go to court without being able to overturn a result by assertion alone. As the remaining cases move through state and federal courts in November and December, the outcomes that matter will not be the filings that generate the most attention, but the orders that survive appeal and the certificates that states ultimately sign.
Sources
USN reviewed reporting from multiple outlets for this story. Read more from Ballotpedia, Brennan Center and The Korea Herald.