Strategy sold Bitcoin in late May for the first time since 2022, specifically 32 BTC for around USD 2.5 million. As a result, a prediction-market contract with more than USD 60 million in trading volume escalated into the largest Polymarket UMA dispute since the Zelenskyy market.
Polymarket is a decentralized prediction-market platform where users trade binary contracts on real-world events. Settlement is handled not by a central counterparty but by an external oracle infrastructure called UMA. The platform runs on Polygon and settles in USDC, with the international Polymarket book operating outside the CFTC-regulated US entity QCX LLC, which has been registered as a Designated Contract Market since December 2025. The disputed contract asked simply: "MicroStrategy sells any Bitcoin by May 31, 2026?", with a cutoff at 11:59 p.m. ET on May 31. Trading volume exceeded USD 60 million, and after the escalation the Yes price collapsed to below 1 cent. Meanwhile, the UMA voting window of 48 to 96 hours has been running ever since.
Strategy sale and the cutoff conflict
Strategy sold a total of 32 BTC between May 26 and 31 at an average net price of USD 77,135, raising around USD 2.5 million. The proceeds were earmarked for dividend payments on the STRC preferred shares, so this was not a strategic exit. Measured against the total holding of 843,706 BTC at a cost basis of USD 75,699, the coins sold furthermore amount to only 0.0038% of the position. However, the sale marks a turning point: the last documented sell-off dated to December 2022, and at the time the company bought back the 704 BTC two days later. The May transaction is therefore the first genuine net sale in years. Michael Saylor had already announced the move beforehand on the Q1 earnings call.
"We will probably sell some Bitcoin to pay a dividend, simply to immunize the market and send the signal that we have done it." - Michael Saylor, Executive Chairman, Strategy
The actual conflict, however, was not sparked by the sale but by its disclosure. Strategy filed the 8-K report with the SEC only on June 1, that is, one day after the Polymarket cutoff. The Yes side argues that the execution window cited in the filing, "as of May 31, 2026, 4:00 p.m. Eastern Time," is decisive, since the sale demonstrably took place beforehand. In its official bulletin, however, Polymarket counters that within the market timeframe neither statements from Strategy nor on-chain data nor a consensus of credible reporting had confirmed the sale. Thus two readings of the same set of facts collide: the timing of the event against the timing of its confirmation.
Subscribe to our newsletter
The best articles of the week, directly delivered into your mailbox.
UMA vote and whale concentration
UMA operates on an optimistic-oracle model, under which a proposed resolution initially applies automatically unless it is challenged within a deadline. In the Strategy market, two "No" resolutions were filed and both challenged, whereupon the case escalated to a token-weighted vote, the so-called Data Verification Mechanism. Voters are presented with three options: P1 for "No" under an announcement-based interpretation, P2 for "Yes" under an event-based interpretation, and P4 for "Too Early" if the rules are deemed too vague. Voting power is directly proportional to token holdings.
This very proportionality makes the model vulnerable. A WSJ investigation from May 2026 showed that in most disputed markets more than half the votes come from the ten largest wallets. Additionally, at least 60% of active voters can be linked to Polymarket accounts, and roughly one in five disputes featured a voter with a financial stake in the contract concerned. As early as March 2025, a single actor used 25% of the voting power to settle a USD 7 million contract in a questionable way. Therefore the issue at hand is less an isolated case than a structural concentration problem.
For the affected traders, considerable sums are at stake. The user "willo2_Poly" puts the loss on the Yes side at around USD 500,000, while "0xDinoCrypto" acquired about 50,000 Yes shares for around USD 35,000 and filed a formal legal claim. Unlike in 2024, when UMA ruled in the dispute over Barron Trump's alleged involvement in the DJT memecoin and Polymarket then overrode the oracle and compensated Yes holders, this fallback option is missing this time. Consequently, the platform is contractually bound by the vote.
Deterministic settlement as a counter-model
While UMA relies on human voting, Hyperliquid pursues a different approach with HIP-4. The Outcome Markets have been live on mainnet since May 2, 2026, and settle via a validator set with automated newsfeed software, entirely without an external oracle and without a token-vote backstop. Each binary contract consequently resolves deterministically to 1 or 0 against a predefined data source, and builders must stake 1 million HYPE per slot, which is slashable in the event of oracle manipulation. On the first day, more than 6 million contracts changed hands between around 4,000 traders.
Kalshi, in turn, resolves disputes through federally regulated exchange mechanisms, since the clearing vehicle Kalshi Klear LLC has been registered as a Derivatives Clearing Organization since August 2024. Moreover, the fee models differ markedly: Polymarket charges around 200 basis points on profits, whereas Kalshi charges around 700. Hyperliquid leaves the opening of positions free of charge and bills only on closing or settlement at the perp rate. Ethereum developer Eric Conner called UMA's model structurally broken and accused whales of exploiting unclear rules to resolve markets incorrectly and protect their own positions. Deterministic settlement is the answer, he said, and that is exactly what Hyperliquid delivers.
Polymarket itself is considering a native POLY token that would internalize the oracle functions. However, concrete plans or a timeline are not yet available. Thus it remains open whether the platform will reform its settlement before competitive pressure from the deterministic alternatives forces it to.
Polymarket UMA dispute reveals a contract-structure problem
The Strategy dispute is not an isolated case but a symptom of a scaling crisis. Polymarket already counted more than 1,150 disputed markets in 2026, more than in all of 2025. The MOOV2 governance update from August 2025, with 37 whitelist addresses, nevertheless did not structurally solve the concentration problem. Consequently, the number of escalated disputes is growing faster than the mechanisms meant to settle them.
One detail is revealing here. The sibling contracts of the same market for June 30 and December 31 already resolved to "Yes" without a dispute. The problem ultimately lies not in the facts but in the wording of the cutoff rule. Polymarket's subsequent bulletin requires proof of information within the market timeframe, whereas the Yes side insists on the timing of the event, and both readings rely on the same contract language. With a stake of more than USD 60 million, such ambiguities pay off for strategically minded actors. The UMA result will therefore show whether token voting remains legitimate at institutional stakes, or whether the pressure toward deterministic settlement increases further.
