Market cap $2.63 T −1.40 % over 24 hours24h volume $61.72 BBTC dominance 59.3 %ETH 10.6 %Fear & Greed 26 FearCoins 22,751

Polymarket says it investigated ‘dozens’ of midterm bets, with at least two cases sent to the CFTC

Polymarket's investigations chief told NPR it has probed dozens of midterm bets; at least two went to the CFTC. No case, account or outcome was named.

9 min read

Abstract magnifying glass over a ballot box beside an eye and a warning sign
Table of contents
  1. What Bautista told NPR about Polymarket insider trading
  2. How Polymarket’s election surveillance system works
  3. Two platforms under one investigations team
  4. What NPR’s report leaves unanswered
  5. Why thin election markets are easy to move
  6. Kalshi bans candidates, but a staffer got round its block
  7. A regulator that leaves policing to the platforms
  8. What to watch before and after 3 November

Polymarket has investigated “dozens” of bets placed on its markets about the 3 November US midterm elections for possible insider trading and manipulation, the company’s head of global intelligence and investigations, Shana Bautista, told NPR in a report published on Tuesday 27 October. At least two of those cases have been sent to the Commodity Futures Trading Commission (CFTC), NPR reported, citing a person familiar with the probes who was not authorised to speak publicly. That figure did not come from Polymarket on the record.

The disclosure, made by NPR reporter Luke Garrett, lands one week before Election Day. It is also thin. NPR does not say which markets were involved, who placed the bets, how much money was at stake or what happened to the accounts behind them. Nothing in the report points to a charge, a fine or any public action by the CFTC on the two referrals.

What the report does give is a rare look at how the crypto-based prediction market says it watches election trading, from a former FBI official who has run that work since June. Coindaro’s Polymarket project page covers how the platform works and how it is regulated.

What Bautista told NPR about Polymarket insider trading

Bautista, a former FBI official and cryptocurrency analyst, told NPR that a team of 16 people and a new election surveillance system are policing bets ahead of the vote. Her warning to anyone holding confidential campaign information was blunt. “If you have information that you have a duty to protect, don’t trade on it,” she said, according to NPR.

She added that such trading breaks Polymarket’s terms of use on both of its platforms as well as the law. Then: “If you decide that you are smarter and can do it, we will find you.”

NPR frames the investigations as concerning possible election insider trading and manipulation. It does not break the “dozens” down between the two, and Bautista did not describe a single case.

How Polymarket’s election surveillance system works

According to NPR, the system pulls “gigabytes” of data on each user. The inputs Bautista listed include social media, Federal Election Commission (FEC) records, cryptocurrency companies, Polymarket’s banking partners, statistical models and other public information. When a user trips enough internally defined red flags, an alert goes to Bautista’s phone.

“We’re getting hundreds of alerts a day,” she told NPR. She declined to say what the detection criteria or thresholds are, citing the security of the system.

She did describe one pattern her team looks for when someone may be trying to push an election market in one direction: several accounts opened, trades placed in short bursts, and the same device IDs turning up on trades made in quick succession from new wallets across four or five accounts. Once a case of that kind is worked, she said, its findings feed back into the detections. Her broader point to NPR was that insider trading and market manipulation look the same whatever the market, and that prediction markets see old tactics applied in a new place.

Abstract eye and magnifying glass above a server rack and a group of people
Bautista says a 16-person team and a new system generate hundreds of alerts a day on election bets.

Polymarket had flagged this build-up before. In an interview with Reuters at the end of August, as reported by PYMNTS on 31 August, Bautista said the company had “the systems in place to be able to identify anomalous activity when the midterms do come”, and announced a web page describing its market integrity programme.

Two platforms under one investigations team

Bautista oversees information from both versions of the company. Polymarket US is the regulated American platform, where users must give their real identities. Polymarket International, the larger of the two, is described by NPR as essentially unregulated, with anonymous trading in cryptocurrency.

That second platform is where questions about anonymity usually start. Bautista argued that crypto trades are rarely fully anonymous. “There’s never been a case that we work where we haven’t been able to identify, potentially, who these people are,” she told NPR. The word “potentially” matters: she did not say every suspect was identified, only that her team has always found leads.

The company has used outside help for this. Decrypt reported on 27 October that the company partnered with the blockchain analytics firm Chainalysis on market surveillance in April. Days later, an active-duty US Army soldier was charged with using classified information about the raid to capture Nicolás Maduro to bet on Polymarket, staking more than $33,000 and winning nearly $410,000, according to Decrypt. He has pleaded not guilty; Polymarket says it detected the activity and reported it to the authorities before the arrest. Bitcoin.com, in its 27 October write-up of the NPR story, reports that Polymarket has referred more than 100 cases across all its markets to law enforcement.

What NPR’s report leaves unanswered

The report rests on two numbers, and neither is firm. The number of investigations is a range word, “dozens”, not a figure. The two CFTC referrals come from an unnamed source, not from the company or the regulator. There is no word on whether those referrals led to any CFTC inquiry.

NPR also does not say how many of the “hundreds” of daily alerts became investigations, or how many investigations closed with no finding. Polymarket published no statement of its own alongside the interview that Coindaro could find. Finally, NPR does not say whether the cases involved Polymarket US, Polymarket International or both.

Why thin election markets are easy to move

Election markets carry a risk that a bet on a football score does not. Rajiv Sethi, an economics professor at Barnard College at Columbia University who has studied prediction markets for decades, told NPR that election markets “have the possibility of being a self-fulfilling prophecy”. If a candidate looks to be losing momentum, voters and donors may act on that impression.

Much of the money sits in very few hands. An analysis by the Anti-Corruption Data Collective (ACDC), published on 17 August 2026 with data to 10 August, counted the figures below.

Measure (ACDC, data to 10 August 2026)Figure
Wagered on 2026 House and Senate races$133 million
Wagered on congressional races in the whole 2024 cycle$92.4 million
Congressional markets on Kalshi, Polymarket and Polymarket US7,466
Polymarket wallets that bet on congressional races39,820
Volume share of the top 1% of Polymarket Global wallets68%
Volume share of the top 10 wallets17%
Seats in which those 10 wallets bet, out of 470426
Polymarket markets with fewer than 100 wallets80%

Most Polymarket congressional markets have a small crowd of traders, and a handful of wallets are present almost everywhere. That is the setting in which a few modest trades can move the displayed odds of a race. The big markets are deeper: Bitcoin.com put the Polymarket market on control of the House at about $24 million in trading volume and the Senate market at about $9 million as of 27 October, across 506 election markets on the platform.

Kalshi bans candidates, but a staffer got round its block

Polymarket is not alone in this. Its main rival, Kalshi, the CFTC-regulated prediction market, banned and fined a handful of political candidates in April for betting on themselves, NPR reported. In May, NPR found that campaign staffers had made “thousands” by using insider polling information to bet on their own candidates on platforms such as Polymarket and PredictIt, in a report published on 7 May.

Also in May, Kalshi launched its own election surveillance system built on FEC data, meant to block any campaign staffer listed by the FEC from betting on elections. In July it said it had blocked “dozens” of staffers. At least one staffer told NPR they got round the block, and shared records of recent trades with the broadcaster on condition of anonymity. Decrypt adds that Kalshi says it has opened more than 150 investigations in the past year and referred more than 20 cases to law enforcement, figures that come from the company.

Both companies now publish counts of investigations and referrals. Neither count, on its own, says how much misconduct got through.

A regulator that leaves policing to the platforms

The referrals go to an agency that, in NPR’s words, has done little to police prediction markets as they grew over the last two years, largely leaving that work to the companies. That is the context for the “at least two” figure: a referral hands the file to a regulator that has so far relied on the platforms to do most of the watching.

The CFTC has acted on specific risks. On 22 September its Division of Market Oversight issued a staff advisory warning that “mention market” contracts, which pay out on whether someone says a given word, may carry a heightened risk of manipulation, The Block reported. Decrypt reports the agency fined a White House teleprompter operator who traded on speeches he had already read, three weeks before that advisory.

Abstract document and hourglass beside a gavel and a warning sign
Referrals go to a CFTC that NPR says has largely left policing prediction markets to the companies.

The company and the agency have met before. On 3 January 2022 the CFTC ordered Blockratize Inc., which runs Polymarket, to pay a $1.4 million civil penalty for offering event-based binary options without registration, and to wind down markets that did not comply. Four years on, Polymarket is the one sending suspected misconduct to that regulator.

Pressure is rising from other parts of government too. On 23 October the Pentagon barred troops from betting on military operations with nonpublic information, naming Polymarket and Kalshi, and on 26 October the NBA signed a prediction markets pact with the CFTC that names both platforms.

What to watch before and after 3 November

The midterms take place on Tuesday 3 November. Election markets settle on results, and some close races can take days to call, so the window in which confidential information is worth trading on does not end on election night.

Any CFTC statement or enforcement action tied to the referred cases would turn NPR’s report into something firmer. So would Polymarket naming outcomes, such as how many accounts it acted on and in which markets, or a breakdown between the regulated US platform and the anonymous international one. Until then, the record is the word of Polymarket’s investigations chief and one unnamed source, carried by NPR.

Sources

  1. NPR (Luke Garrett), Polymarket has investigated 'dozens' of election-related cases, 27 October 2026, via WFSU news.wfsu.org
  2. Decrypt, Polymarket has investigated dozens of midterm election bets: report, 27 October 2026 decrypt.co
  3. Bitcoin.com News, Polymarket flags suspicious midterm election bets, 27 October 2026 news.bitcoin.com
  4. Anti-Corruption Data Collective, analysis of 2026 congressional prediction markets, 17 August 2026 acdatacollective.org
  5. CFTC press release 8478-22, order against Blockratize (Polymarket), 3 January 2022 cftc.gov
  6. The Block, CFTC advisory on mention markets and manipulation risk, 22 September 2026 theblock.co

Leave a Reply

Your email address will not be published. Required fields are marked *