19 Polymarket Accounts Won 41 of 42 Bets on KPMG-Audited Firms
The figures come from a Bubblemaps blockchain analysis reported by The Wall Street Journal. The bets began in November 2025 and continued for several months, covering companies including Wells Fargo, Home Depot, DoorDash and General Mills.
Bubblemaps also identified fund movements between the 19 accounts. The pattern suggested they could have been operated by the same person or group, although the analysis did not establish who controlled them.
KPMG Bets Ran 15.6 Points Above the Comparison Rate
The headline 98% figure is rounded. With 41 wins from 42 bets, the actual success rate on the KPMG-audited companies was 97.6%. By 15M’s calculation, that puts the KPMG-linked hit rate 15.6 percentage points above the 82% recorded on companies with other auditors.
The difference is notable, but it does not prove that confidential information was used. The available reporting also does not provide the full number of non-KPMG bets behind the 82% comparison, so the gap should be treated as a descriptive result rather than a statistical test.
The structure of the contracts makes the pattern relevant to market integrity. Polymarket earnings markets are based on whether a company reports above or below a preset Wall Street consensus figure. A trader does not need to predict how the stock will react after the release; the contract settles on the reported result itself.
No Link to the KPMG Investigation Has Been Established
The account cluster has emerged while US authorities are separately examining a KPMG employee over suspected prediction-market trading involving corporate earnings information. The two matters have not been connected. It remains unknown whether the person or people behind the 19 accounts have any relationship with the employee under investigation, and no charges have been announced against the operators of the cluster.
KPMG told the Journal that it has zero tolerance for trading based on nonpublic client information, including through prediction markets. The firm also said it has strengthened its monitoring capabilities.
Polymarket already prohibits trading on confidential information when its use breaches an existing duty of trust or confidence. Its market integrity policy also bans trading on illegal tips and by people able to influence the outcome of a contract.
The Missing Link Is Account Ownership
The profit itself was relatively small at about $22,000. The unusual part of the dataset is the concentration: 18 companies sharing the same auditor, 42 earnings bets and only one losing trade.
The case also shows both sides of blockchain-based market surveillance. Transaction records allowed Bubblemaps to trace transfers and group the 19 accounts together. The same data cannot establish who controlled the wallets or what information, if any, was available to them.
For prediction markets, that distinction is becoming an important integrity issue. Suspicious trading patterns can be visible on-chain well before the evidence is strong enough to show where the information came from or whether any rule was broken.
The next point to watch is whether US investigators connect the wallet cluster to the separate KPMG inquiry or identify who controlled the accounts. Any such link would turn an unusual trading pattern into a more direct test of prediction-market surveillance and enforcement.