Prediction markets near certainty on Democratic congressional sweep

Prediction markets swung sharply toward a Democratic sweep of Congress after polls began closing in the U.S. midterm elections, with aggregated prices putting the likelihood of Democrats winning both chambers close to 100% by 9 p.m. Eastern time. More than $1.48 billion had been wagered on the midterms overall.

The move followed a late rush of money once the first polls closed at 6 p.m. Eastern. Traders placed more than $4 million in bets within minutes, mostly favoring Democrats, including a $951,000 Polymarket wager on the party taking the House. Nearly $934,000 in election bets entered Kalshi in a single second.

The confidence of the markets ran ahead of television-network projections. NBC News and CNN projected a Democratic House win, but said it remained too early to determine whether the party would win enough Senate seats for control. By 10 p.m., Democrats had flipped Republican-held Senate seats in North Carolina and South Carolina, while showing unexpected strength in several Republican-leaning Florida House districts.

Earlier on Election Day, Kalshi’s own prices had been markedly less definitive. Its Democratic-sweep contract rose from 62.4% the previous week to a 75.6% high, before settling at 73% at 4:50 p.m. Its separate markets put Democratic chances at 93.6% for the House and 74% for the Senate. The difference underscores how rapidly prices changed as voting ended and results emerged.

Trading had already reached exceptional levels before election night. Kalshi reported $1.33 billion in midterm volume on Monday, and its total exceeded $1.4 billion on Tuesday, according to Yahoo Finance. Its Congress-control contract alone drew $251 million, while the House-control contract exceeded $181 million. About $145 million had been wagered in Polymarket’s congressional markets, according to the Anti-Corruption Data Collective.

As this publication covered on Nov. 1, polling and prediction markets had both pointed to Democratic gains, while raising questions over whether markets offered independent evidence or largely tracked polls. Their performance remains contested: a 2025 Vanderbilt University study of nearly 2,500 markets found them generally well calibrated, but researchers cautioned that broad measures can obscure accuracy in close races. Markets have also missed high-profile contests, including Brazil’s presidential election and Wisconsin’s Democratic gubernatorial primary.

The midterms therefore represented a consequential test for platforms that gained prominence after traders favored Donald Trump in 2024 while traditional polls largely treated the presidential race as a toss-up. More than 28,000 midterm markets had opened, although a relatively small share of contests accounted for much of the trading activity.