Kalshi's Midterm Odds Say Democrats Are Winning. The Courts Are Still Deciding If That's Legal.

Some coverage is already calling November 2026 the first true "prediction-market election," and it's hard to argue with the framing. Kalshi didn't just add a few political contracts to its board this cycle — it built a dedicated Midterms Hub in July, pairing live event-contract odds with polling averages and FEC fundraising data, and pitched it as the "one-stop-shop" for tracking every competitive Senate and House race. That followed May's launch of Kalshi's "American Power Index," explicitly billed as the S&P 500 of politics. Polymarket is running the same race in parallel. Both platforms are trying to become the place people check instead of a polling average, months before a single vote is counted.

What the odds actually say right now

As of September 26, 2026, Kalshi and Polymarket's averaged odds give Democrats a 62% chance of winning Senate control (Kalshi at 62.2¢, Polymarket at 63¢) and a 92% chance of holding the House — both up roughly 2 points week over week, per DeFi Rate's tracker. A handful of individual races are doing most of the work: Texas' new 51st Senate seat has Democrat James Talarico at 61% against Republican Ken Paxton; Michigan's Aftab Pureval is at 73% against Mike Rogers, up 8 points in a week; an Ohio special election has Democrat Sherrod Brown at 59% over Jon Husted. Maine is the one race moving the other way — incumbent Republican Susan Collins has clawed back 5 points against Democrat Sonia Jackson, with Jackson still favored at 63%. Iowa is one of the few seats where Republicans are favored outright, at 57%.

The more interesting number is the gap between markets and professional forecasters. Piper Sandler's poll-based analysts put Democrats' Senate odds at 45% as of September 22 — up from 40% the week before, but still well below what Kalshi and Polymarket were pricing the same week (61% and 65%, respectively). Real Clear Politics' generic congressional ballot has Democrats up 8.5 points, which is directionally consistent with the market read, but the size of the gap between "markets" and "pollsters" on Senate control specifically is large enough to be its own story. Either the markets are running ahead of what the underlying data supports, or professional handicappers are still anchored to an older read of the race. November settles which.

The legal war underneath the odds

None of this is settled as a matter of law, and the fight is escalating in real time. On April 2, 2026, the CFTC and Department of Justice filed federal lawsuits against Arizona, Connecticut, and Illinois, arguing that event contracts traded on Kalshi's registered exchange are federally regulated derivatives — "swaps," in the Commodity Exchange Act's language — and that state gambling laws are preempted. The CFTC called the move an "unprecedented measure" necessary to protect its exclusive jurisdiction. Arizona's case is the furthest along and the most dramatic: Attorney General Kris Mayes filed a 20-count criminal information against Kalshi in Maricopa County Superior Court in March, the first criminal case ever brought against a CFTC registrant. A federal judge, Michael Liburdi, granted a temporary restraining order blocking that prosecution, ruling that the Commodity Exchange Act likely "field preempts" Arizona law and that the state's enforcement "violates the Supremacy Clause of the United States Constitution." Arizona's AG office says it disagrees and is weighing an appeal.

Kalshi has also been winning outside Arizona: the Third Circuit ruled 2-1 in Kalshi's favor against New Jersey in April, upholding an injunction on the same preemption theory. But the picture isn't uniformly one-sided — separate reporting describes a Ninth Circuit panel going the other way, affirming that states retain authority to regulate prediction markets as gambling. Taken together, that's the shape of a genuine circuit split forming on the exact question of whether federal derivatives law overrides state gambling law for election contracts — the kind of disagreement that typically ends up in front of the Supreme Court, or in limbo until it does. None of it is likely to resolve before the midterms are decided, which means the market you're reading odds from in November may be operating in a different legal environment than the one it's operating in today.

Why election officials are nervous

The concerns from people who actually run elections are specific, not generic anti-gambling sentiment. Maryland election administrator Jared DeMarinis has called the trend "a troubling trend that election administrators across the nation must deal with." Los Angeles County registrar-recorder Dean Logan pointed to a concrete incident from a mayoral primary, where early returns diverging from what the markets had priced "resulted in suspicion and questioning of normal ballot processing and canvassing procedures" — a real-world case of a market number becoming the public's baseline expectation, and the actual count then getting treated as the anomaly. Delaware County, Pennsylvania went as far as banning its own election workers from trading on these markets, comparing it to "referees betting on the outcome of a basketball game."

The mechanical worry underneath all of this is one PumpReport readers will recognize from a very different context: herding. A large enough bet can move a market's implied odds, and other traders tend to follow a move they can't otherwise explain, on the assumption that someone else knows something. That's the same feedback loop covered in our Anatomy of a Pump piece — thin liquidity plus a triggering order plus follow-on momentum — just running on a market where the "asset" is a Senate seat instead of a token, and where the downstream effect isn't a bad trade but public confidence in an election outcome.

Why Kalshi is pushing this so hard

The business logic is straightforward even if the politics are messy. Kalshi posted $31 billion in total notional volume in June 2026, up 70% month over month — but sports contracts made up roughly 85% of that activity. Political contracts are a smaller slice of the business today, and a faster-growing, stickier, higher-attention one: an election cycle generates sustained daily engagement for months in a way a single game never does, and being the reference odds source for a competitive midterm is a brand-defining moment regardless of which party actually wins. The Midterms Hub and the American Power Index are both bets that "Kalshi" becomes a word people use the way they currently use "the polls" — a bet Polymarket is making in parallel, and one that gets more valuable the more often either platform's number is the one that shows up in a headline.

What to actually watch from here

Three threads are worth tracking independently of who you want to win. First, whether the market-implied odds (Democrats clearly favored in both chambers) or the more cautious poll-based read (a closer race, especially in the Senate) turns out to be the better forecast — a real, public test of the "wisdom of markets" thesis at a scale it hasn't been tested at before. Second, whether the Arizona case, the Third Circuit ruling, and the conflicting Ninth Circuit read get resolved — by an en banc rehearing, a circuit split heading to the Supreme Court, or simply running out the clock into November with the legal question still open. Third, whether any swing-state race sees a Dean Logan-style moment, where a market number gets treated as ground truth and an actual, correctly-tabulated result gets treated with suspicion because it didn't match. That last one is the scenario election officials are actually worried about, and it doesn't require anyone to manipulate anything — it just requires a market and a real result to disagree once, publicly, in a close race.

Sources: DeFi Rate, Fortune, CNBC, Votebeat, Arizona Mirror, Forbes, CoinDesk, The Hill. Figures and legal case statuses as reported at time of publication and subject to change.