Prediction markets faced their defining test during the 2026 U.S. midterm elections, processing roughly $1.3 billion in election-related contracts. Polymarket and Kalshi correctly indicated a Democratic blue wave before major television networks projected the outcomes, with implied probabilities often above 93 percent for House control and about 75 percent for the Senate. Those chamber calls held, but a New York magazine analysis found the platforms underestimated several Democratic Senate upsets. Republican candidates were favored in South Carolina, Iowa and Kansas, yet Democrats flipped those seats. With Democrats now controlling Congress, the sector also faces an uncertain regulatory path.
Separately, onchain analysis has complicated an insider-trading investigation into a Polymarket account called dubdubdub2. Vanderbilt University researchers had flagged large, profitable positions on Nebraska and Michigan Senate races placed shortly before New York Times/Siena polls were released. The trader earned more than $257,000 across the two Polymarket positions — about $77,000 from a $191,000 Nebraska position backing Senator Pete Ricketts and more than $180,000 from a $622,608 Michigan position backing Democrat Abdul El-Sayed. That Michigan activity represented the largest daily total by a single account across the 16 Senate markets examined.
Public blockchain records and trading data have since revealed a broader trading history. The linked wallets withdrew about $2.2 million from Hyperliquid on October 30, moved roughly $1.39 million toward Zero Hash for Kalshi, and recorded tens of millions of dollars in transfer activity. The account previously wagered about $1.05 million on Zohran Mamdani winning New York City’s mayoral race, operated heavily in decentralized finance through Aave and Pendle, and recorded more than $13 million in stablecoin withdrawals from Kraken since October 2025. Experts said the pattern was consistent with a wealthy trader or small trading firm making concentrated directional bets, rather than a clear attempt to exploit early polling releases.
Still, investigators continue to examine the timing of Kalshi purchases around the Nebraska poll. A $175,000 deposit on October 30 was followed about 31 seconds later by $159,000 in purchases backing Ricketts, while approximately $153,000 in further purchases were recorded around 5:15 a.m. ET on October 31 — roughly 10 minutes after the poll was published. Polymarket confirmed it is investigating the trades, and the New York Times and Siena Research Institute said they would review the allegations. Kalshi declined to comment on specific activity.
The scrutiny arrives as prediction markets face broader regulatory pressure. Polymarket has expanded surveillance through its PolyGraph system and a partnership with Chainalysis. Earlier this year, federal prosecutors charged a U.S. soldier with allegedly using classified information to generate about $409,881 in Polymarket contracts. Lawmakers have also requested more details on customer verification and controls against misuse of material nonpublic information. For now, the unresolved question is not whether the trades were profitable or well-timed, but whether investigators can establish any connection between the trader and nonpublic polling information.
