Where the Money Waits: Polymarket's Open Interest Is 55% Politics, Kalshi's 44% Sports

Kalshi and Polymarket list many of the same questions and are different venues underneath. Kalshi is a CFTC-regulated exchange that trades in dollars. Polymarket's main venue is an on-chain order book settled in USDC on Polygon and closed to U.S. residents, with a separate, sports-first U.S. exchange that opened to the public in May 2026. Both charge takers a rate times p × (1 − p): at 50 cents, 100 contracts cost $1.75 to take on Kalshi and between nothing and $1.75 on Polymarket, depending on the category, and Polymarket makers pay no fee.

What they hold differs as much as how they work. Vultax read every open event on both venues late on 13 September 2026. On Polymarket, $341.3 million was open, and 55.3% of it sat on politics and elections; the economy and the Fed held 16.0%, crypto 12.3% and sports only 7.8%. On Kalshi, 1.04 billion contracts were open, and sports held the largest share, 43.6%, with elections close behind at 34.0%. Holding is not trading: esports was the largest sport by money traded on Polymarket in Vultax's record of 21 August to 12 September, yet it held 0.6% of the venue's open interest, because a match settles within hours.

Their prices sit close but seldom together. In Vultax's July reading of eight matched Fed-rate contracts, the one pair whose terms matched exactly sat a median 574 basis points apart, about 5.7 cents, and the gap did not close at any point in 24 hours. Whether a gap like that can be collected after both venues' fees is a separate question, answered by the arbitrage study.

For the same contracts priced side by side right now, with the gap before and after fees, see the live Kalshi vs Polymarket comparison.

The comparison a trader makes before choosing where to trade. Both venues' taker fees follow the same shape: they peak at 50 cents and shrink toward the ends of the price range.

Download table figures:CSVJSON with sources

Open interest is the value of positions still waiting to settle, and it shows where money stays. Vultax paged through every open event on Polymarket's public events endpoint between 23:22 and 23:24 UTC and on Kalshi's public events endpoint between 23:21 and 23:22 UTC on 13 September 2026. Polymarket reports each event's open interest in dollars; Kalshi reports open contracts for each market, which Vultax summed by event. A Kalshi contract pays $1 if it wins and is bought for less, so its contract count is an upper bound on the dollars at stake and not directly comparable with Polymarket's figure. Shares within each venue are comparable.

Kalshi assigns each event a category. Polymarket tags events with many labels, so Vultax assigned each event to the first matching group in this order: esports, sports, crypto, economy and Fed, politics and elections, geopolitics, culture, tech; an event tagged both sports and esports counts as esports.

Prices come from Vultax's cross-venue record, which keeps a registry of Polymarket markets paired with Kalshi tickers, each pair graded exact, near or excluded by a reviewer according to whether the two contracts resolve on the same condition. In July 2026 it observed eight Fed-funds pairs roughly every 85 seconds for 24 hours from 17 July, 1,018 times each and 8,144 observations in all, recording both prices and the absolute gap between them in basis points.

More than half of Polymarket's open money waits on political outcomes, from the 2028 presidential election to Brazil's. The September Fed decision was the second-largest single event by open interest three days before the meeting.

Download table figures:CSVJSON with sources

Download chart figures:CSVJSON with sources

Kalshi's open contracts divide between sports and elections, and its largest sports positions are not tonight's games but season-long questions such as the next NFL, MLB and NBA champions. Game contracts settle within hours and so hold little open interest even though they dominate Kalshi's daily trading; in a reading of Kalshi on 12 September, sports were 86% of contracts traded in 24 hours.

Download table figures:CSVJSON with sources

Download chart figures:CSVJSON with sources

Each venue's headline trading figures describe fast markets; its open interest describes patient ones. Three examples from Vultax's September measurements show the gap.

• Esports on Polymarket - 21.4% of observed money traded between 21 August and 12 September, 0.6% of open interest on 13 September.
• Crypto up-or-down markets on Polymarket - 59.6% of observed trades in the same window, settling every five minutes to a day, so they barely register in open interest; crypto's 12.3% share is mostly long-dated price questions such as where Bitcoin will trade in 2026.
• Sports on Kalshi - 86% of contracts traded in a day, 43.6% of open contracts, most of them in championships decided months from now.

Open interest is the better guide to what a venue is exposed to and what its users are waiting on. Polymarket's depends heavily on U.S. politics through 2028 and on a handful of events; Kalshi's on the U.S. sports calendar and the same elections. Both would look very different after the midterms settle. For where daily trading goes, see the up-or-down study and Polymarket's sports and esports markets.

Matched contracts show how closely the two venues agree, provided the contracts are the same bet. The July window covered eight Fed-funds pairs: one exact match (Polymarket "upper bound 4.50% at end of 2026" against KXFED-26DEC-T4.25), six near matches (Polymarket "reach X or higher before 2027" against Kalshi December thresholds), and one pair excluded by the reviewer.

On the exact pair the median absolute gap was 574 basis points, about 5.7 cents on a dollar contract, with a 90th percentile of 934 bps, and it did not close at any point in the window. Of the 19 gaps Vultax tracked as they opened, 17 were still open when the window ended; the two that closed lasted under a minute. Five of the six near matches sat 128 to 352 bps apart at the median, but a gap between correlated questions is not a gap in price: "reach 4.50% or higher before 2027" and "4.50% at the end of 2026" resolve differently on a path that touches 4.50% in March and ends at 4.25%. The sixth, a mismatch in threshold direction, showed a median gap of 6,430 bps between two different questions.

A gap that persists for a full day against a public order book on each side is not being missed by capital; it is being priced, as the cost of moving USDC on Polygon against USD at a regulated broker, of different resolution sources, and of different counterparties. The midterms show the same pattern across a wider set: on 65 matched contracts the two venues usually sat 1 to 3 points apart, with Polymarket higher on Democrats, as the midterms comparison sets out contract by contract.

Download chart figures:CSVJSON with sources

A price gap is only money if the two contracts resolve on the same facts, both venues' taker fees leave something over, the smaller book holds enough size, and the return justifies holding capital until settlement. Vultax tested those conditions by walking both venues' order books every five minutes for 28 hours on sports, midterm and Fed contracts; the arbitrage study shows where a gap cleared fees, what it returned, and runs the arithmetic on any two prices.

The strip at the top of this page is rewritten every ten minutes from Polymarket's Fed events and Kalshi's Fed-decision series for the September, October and December meetings, nine exactly matched contracts in all, plus Manifold and PredictIt on the September hike. The Vi IQ beneath it scores the six domains the Vultax terminal scores for a pair, computed for this page's subject from published inputs, with a domain that cannot be computed shown as unavailable rather than filled in. The figures in the body of the article are the measurement it was written on and are dated; the strip is what the same measurement says now.

The first figure is the September hike gap, Polymarket minus Kalshi, in probability points. The second is the median absolute gap across all nine matched pairs, the third the widest of them after a two-point fee floor, and the fourth how many venues are currently pricing the September hike. A gap that clears the floor is the exception; the pack records every pair and the day-ahead calibration of both venues over the last week.

The two venues are less alike than a matched contract suggests. Kalshi is a CFTC-designated exchange whose contract volume reached about 90% of the pair's combined total in late August 2026, driven by sports and multi-leg combos; Polymarket's main venue is offshore, while its U.S. exchange, launched in December 2025 on the QCEX designation and opened to the public in May 2026, is sports-first and iOS-only. Since mid-2024 sports has been about 80% of Kalshi's volume and 39% of Polymarket's. Fed contracts are a small corner of both.

Regulation is now the largest source of structural difference. The Ninth Circuit ruled that federal law does not pre-empt state regulation of prediction markets, contradicting other rulings and pointing at the Supreme Court; Kalshi is blocked in Washington, Michigan and Nevada, Minnesota banned prediction markets from 1 August 2026, and the CFTC has proposed rules barring contracts on injuries and officiating. Each of those changes who can trade a given contract on which venue, which is what a gap ultimately measures.

Both venues publish calibration evidence of their own. Kalshi's August 2026 study over more than 2.2 million data points found forecasts improving with time, participation and volume; a separate 2026 academic study found its federal-funds-rate contracts near-perfectly calibrated by decision day. The live pack on this page runs the same test on both venues every week, with the same rule on each: the yes price a day before settlement against the outcome.

Open interest is one reading at one moment. It moves with prices as well as with new positions, and the September Fed event was due to settle on 16 September. Polymarket figures cover its international exchange, not its U.S. exchange. Category groups for Polymarket are Vultax's assignment from Polymarket's tags and would shift slightly under a different order. Kalshi's contract counts overstate dollars at stake by an amount that depends on each contract's price.

The price comparison is one 24-hour window, eight Fed-rate pairs and one exact match: enough to show how matching decides what a gap means, not a market-wide statistic. Prices are the venues' quoted prices at the observation instant, not fills. Polymarket's fee rates were read from its documentation on 5 September 2026; Kalshi's fee schedule was cross-checked from a secondary source on that date because kalshi.com rate-limited automated reading. Nothing here is a recommendation to trade either venue, and nothing here is financial advice. Challenge a figure at [email protected]; material changes are logged in the changelog.