Kalshi has reported more than $2.5 billion in trading volume on its 2026 U.S. midterm election markets, nearly double the $1.3 billion recorded during the 2024 election cycle.
Summary
- Trading volume during the 2026 midterms exceeded the 2024 total by more than $1.2 billion.
- Kalshi said its favored candidates won 479 out of 487 races counted by early November 4.
- The platform reported 98.2% accuracy in predicting winners when polls closed, compared with 93.8% three months earlier.
- Election-night markets showed traders predicting Democratic victories in Texas and South Carolina as results came in.
According to Kalshi’s November 4 report, traders expected Democrats to take control of both chambers of Congress even before Election Day. However, they did not fully expect the size of the party’s victories.
The reported $2.5 billion in trading volume represents an increase of more than 92% compared with the 2024 election cycle. However, the two elections were different, as 2024 included a presidential race, while the 2026 cycle focused on congressional, state, and local elections.
Kalshi Reports Nearly 98% Accuracy in Election Predictions
Kalshi said its favored candidates won 479 of the 487 races counted by 6 a.m. ET on November 4. The results included House, Senate, and governor races, along with elections for attorney general, lieutenant governor, and secretary of state.
The platform reported that its favored candidates won 356 out of 360 House races, 32 out of 33 Senate races, and 33 out of 34 governor races. Across all the contests, only eight market favorites lost, giving Kalshi an overall success rate of around 98.4%.
Some races were still undecided when the report was released, including Senate contests in Alaska and Mississippi. Therefore, the accuracy figure only covered races with results available at that time.
Kalshi also reported that its market predicting Democratic control of the House reached a 99% probability at 8:10 p.m. ET on election night. The market for Senate control reached the same level at 9 p.m.
Election-Night Predictions Shifted as Results Came In
Kalshi’s live updates on November 3 showed how trading predictions changed as election results started coming in.
At 6:55 p.m. ET, the platform gave Democrats a 97.2% chance of controlling the House. Later, at 8:06 p.m., Republican incumbent Greg Abbott was no longer the favorite in the Texas governor’s race.
By 9:35 p.m., Democrat James Talarico had reached a 99% probability of winning the Texas Senate race. At 10:10 p.m., Annie Andrews received the same probability in South Carolina. Texas governor candidate Gina Hinojosa also reached a 99% probability at 11:35 p.m.
Kalshi clarified that these figures represented market predictions and were not official election results or confirmed race calls.
Reuters also reported that prediction markets expected Democrats to take control of Congress but underestimated how strongly some candidates would perform. Research from analytics firm BETINT found that 14 Democratic winners had been considered underdogs by prediction markets just one week before the election.
Kalshi Research Shows Improving Prediction Accuracy
A separate study by Kalshi Research examined how accurately its markets predicted election winners during the three months leading up to the November 4 results.
The research found that prediction accuracy increased from 93.8% three months before Election Day to 98.2% when polls closed. This represents an improvement of 4.4 percentage points.
The study also found that Brier scores improved from 0.040 to 0.016. This measurement compares predicted probabilities with actual results, with lower scores showing better performance.
However, the researchers explained that simply counting winning predictions does not tell the whole story. For example, a candidate given a 51% chance of winning and another given a 95% chance both count as correct if they eventually win.
The study was written by Nicole Kagan and Rubens Baiocchi. It compared market predictions with election results using specific time points before and during Election Day.
U.S. Election Trading Faces Growing Attention
Before the midterm elections, prediction markets also faced increased attention from U.S. regulators over concerns about possible market manipulation.
Earlier reports highlighted comments from Commodity Futures Trading Commission Chairman Michael Selig, who said regulators were monitoring election markets.
Research from Dune showed that Kalshi and Polymarket had a combined $250 million in open interest on midterm election markets as of October 18. Kalshi accounted for $210 million, while Polymarket held around $40 million.
There were also calls for stricter rules on election trading. In October, Representative Don Davis introduced the No Betting on Your Own Race Act, which would prevent federal candidates, their spouses, dependent children, and authorized campaign committees from trading contracts connected to their own elections.
The proposed legislation includes penalties of $10,000 or three times the net financial gain from the trade, whichever amount is higher.