Prediction Market Volume in Q2 2026: $110B and a Reshuffled Leaderboard

Most of the quarterly prediction market volume figures circulating this year came from somebody else's dashboard. Where two of them disagree by multiples, it is almost always because one counts notional and the other counts cash flow. Where they disagree by a few per cent, it is usually the venue list. These are ours. We ranked the venues we track by notional volume for Q1 and Q2 2026 and published the table on 6 August 2026. This is the long version, with the analysis that does not fit in a thread.

The growth rate is the headline. The reshuffle is the story. Two of Q1's top five are gone from Q2's, the fastest-growing venue in the set is one of the newest regulated exchanges in it, and one venue that looked substantial three months ago lost most of its volume, with a smaller one losing nearly all of it. A sector total hides all of that.

Prediction Market Volume Ranking: Q1 vs Q2 2026

Every venue we rank, in notional US dollars, sorted by Q2. This is our data rather than an aggregation of press reports.

The cells marked "derived" are not separate measurements. They fall out of the quarter we published and the growth rate we published, so anyone with a calculator can recover them, and we would rather print them than pretend they are private. Gemini, SX Bet and Myriad each traded less than the venues above them and we have not published dollar figures for those three.

The top five in Q2 add up to $106.66B, roughly 96.7% of the $110.35B total. The churn is underneath them. Opinion and Crypto.com ranked third and fifth in Q1 and finish sixth and seventh in Q2, while Polymarket US, sixth in Q1, and Limitless move up to take their places. Falling out of the top five in a market growing 44.8% takes some doing.

On the question everyone asks first, Kalshi's $65.72B is about 2.45 times Polymarket's global $26.79B. Treat the two Polymarkets as one business and the combined $33.81B still leaves Kalshi roughly 1.9 times larger. Our Polymarket vs Kalshi comparison covers the rest of that head to head, including fees, access and where the two disagree on price.

Quarter on Quarter Growth: Every Venue We Ranked

Ranking by size tells you who is big. Ranking by prediction market volume growth tells you what is moving, and the fourth column above is the one to read twice. Polymarket US, the CFTC-regulated arm of Polymarket, led the whole set at 463.3%. Limitless followed at 252.9%. Kalshi managed 102.8% on much the largest base in the table, which is the harder trick by some distance. Predict.fun and Gemini finished a little over three points apart at 89.4% and 86.1%, SX Bet added 43.3%, and Polymarket's global exchange barely moved at 2.3%. Three venues went backwards: Crypto.com by 32.4%, Opinion by 80.7% and Myriad by 93.8%, all three shedding market share as well as volume.

Why the Regulated Venues Ran

Polymarket US grew off a base that was deliberately throttled: an invite beta from late 2025 and a waitlist through all of Q1, with press reporting putting the removal in mid-May 2026 and the app on iOS only at that point. Platform availability changes without a press release, so check Polymarket US before you repeat that detail. The point for the table is that +463.3% is a gate opening rather than a demand shock. CNBC, citing Dune, reported daily volume on the US exchange rising from around $50 million a day in mid-May to more than $200 million on 20 June, so access opened mid-quarter and the World Cup did the rest. The Q1 base that produces 463.3% is about $1.25B.

Kalshi's +102.8% has two clear drivers and one complication. The World Cup began on 11 June and pushed daily volumes across the sector to records. The Third Circuit ruled on 6 April 2026 that sports event contracts are swaps under exclusive CFTC jurisdiction and that New Jersey's gambling law is field preempted, meaning federal regulation of the area is exclusive and state gambling law cannot reach these contracts at all. That lifted a real overhang from the largest regulated venue's biggest category. The complication is that Kalshi lost an exclusive mid-quarter. In early June, Robinhood began routing part of its World Cup book, including match outcomes, tournament winner and goal totals, to Rothera, the CFTC-regulated exchange it co-owns with Susquehanna; player and combination markets stayed on Kalshi. The diversion was real and, so far, small: CNBC, citing Bank of America, put Rothera at roughly $2B of notional across the whole of June against Kalshi's $31B in the same month. Kalshi doubled while its largest retail partner began to diversify, which is a very different sentence from doubling while losing it. That arrangement can change without notice, so treat it as the position in August 2026. We break the pairing down in Kalshi vs Robinhood, and the wider sports question in prediction markets vs sports betting.

Gemini's +86.1% comes off a base that only began in December 2025, so the percentage flatters a venue that is still small in absolute terms. Then there is Polymarket's global exchange at +2.3%, which is effectively flat in a quarter when the sector grew 44.8% and a World Cup happened. Two things plausibly account for it. The US arm now serves traffic the global site cannot, and Polymarket introduced taker fees across nearly all categories on 30 March 2026, the day before the quarter opened, then rebuilt fee handling in April's CLOB V2 upgrade. A fee regime that landed on the eve of Q2 is the more proximate candidate, though volume data on its own will not tell you how the two split. Merge the two Polymarket rows into a single number and the question stops being visible at all.

What Happened at the Bottom of the Table

Opinion is a BNB Chain (BSC) venue and Myriad runs across BNB Smart Chain, Abstract and Linea. Both fell hard, which makes "BSC collapsed" a tempting headline, and it does not survive contact with the data. Predict.fun is BNB Chain native, sits in the same top five, and grew 89.4% over the same quarter, partly on the back of its 4 March 2026 acquisition of rival Probable. The chain was not the variable.

Incentives fit better. Both Opinion and Myriad ran weekly points programs rewarding trading volume, order placement and position holding ahead of a token, and Opinion's token event landed on 5 March 2026. Third-party on-chain dashboards recorded a steep drop in BSC prediction market volume in the run-up to it, with daily volume falling from over $94M to $38.3M between 1 and 2 March, before the token event itself, which fits incentive-driven flow unwinding as the event approached rather than a post-launch sell-off. Commentators had already queried Opinion's January figures, given how fast it scaled from an October 2025 launch to roughly a third of sector volume by January. We report those questions rather than endorse them: neither venue has published an explanation and we have not audited their books. What we can say is that volume underwritten by points tends to leave when the points stop mattering.

Apply that caveat evenly or it is just partisanship. Limitless's +252.9% quarter also sat entirely inside an active points season with an airdrop attached, and it earns the same asterisk Opinion's Q1 does. Two honest gaps remain: we found no Q2 event that explains Myriad's 93.8% decline, and nothing at all that explains Crypto.com's 32.4% contraction. Both are on the list to revisit when Q3 closes.

Methodology, and What These Numbers Are Not

The figures cover calendar Q1 and Q2 2026 and were published on 6 August 2026. The thread carried the top five for each quarter plus the ten-venue growth ranking; the table above fills in the four cells that those growth rates make computable. Percentages are calculated on unrounded figures, so recomputing one from the rounded totals in the table will occasionally land a tenth of a point away.

Notional volume is the total dollar value of contracts traded, counted at full contract value rather than at the price paid. Buy 1,000 contracts at 8 cents and you have spent $80 and added $1,000 to notional volume. Cash-flow volume is the other side of that same trade: the dollars actually paid, price times contracts, so $80. That one convention explains most of the confusion in this sector. It is why our own dashboard shows tens of billions of notional against a much smaller cash-flow total for the same window, and why the reports on Kalshi's June 2026 put the month at $31 billion and at $9.4 billion in the same breath. The first is notional, the second is traded value, and a reader who lifts one without the label ends up three times off. For the full tour of who measures what, see prediction market data.

What notional is not: revenue. Sacra put Kalshi's annualized revenue run rate at $3.5 billion in June 2026, up from roughly $735 million in December 2025, against $65.72B of notional in a single quarter. Those are estimates rather than disclosures and they move quickly, so date and attribute any figure you quote. Polymarket was reported by Reuters to have passed $1 billion annualized, on a single unnamed source. It is not profit, which nobody in this table discloses. It is not open interest, which is the value of contracts held open at a point in time rather than traded over a period, and is far smaller. And it is not users: a venue can post record volume on a shrinking base of larger traders, which is precisely the concern raised about Opinion. This is data reporting, not investment, tax or legal advice.

What a Trader or Builder Should Take From This

If you trade, the leaderboard is not a liquidity guide. A venue can rank third overall and still show an empty book in the market you want, and Q2's totals were heavily shaped by one tournament, which makes Q3 comparisons awkward for everyone. Check depth on the contract you are trading, not the logo above it. In practice that means opening the order book on the exact contract and looking at how much size is resting within a couple of cents of the mid price, rather than reading the headline volume figure attached to the event. Both Kalshi and Polymarket expose that book through their APIs if you would rather pull it than eyeball it, which we cover in the best prediction market APIs and SDKs.

If you build, three things follow. Polymarket and Polymarket US are two exchanges with different rails, accounts and fee schedules, and merging their rows produces a number that means nothing. Any venue-level volume series should be read alongside whether an incentive program was running, since the same venue can look like a rocket and a ruin two quarters apart for reasons unrelated to its product. And if you are pulling a prediction market volume chart from a public source, check which metric it plots before it goes in a deck. As of August 2026, The Block publishes a weekly USD notional volume chart covering Polymarket, Kalshi, Hyperliquid, Robinhood and Rothera, DefiLlama publishes per-protocol prediction market volume, and Dune hosts curated tables you can query yourself. Read each dashboard's stated definition before setting one against another, because they do not all measure the same thing. Our live figures sit on the Predictefy analytics page with definitions attached.

Frequently Asked Questions

Conclusion

Q2 2026 was the quarter this stopped being a two-horse race and became a leaderboard that moves. Regulated venues took the growth, incentive-driven venues gave back what they had borrowed, and the largest exchange doubled while its biggest retail partner started sending part of its flow elsewhere. We publish the numbers because the sector is better served by data with a stated methodology than by a fifth unattributed figure in a press cycle. Notional is not revenue, not users and not open interest, so check which of those your source is actually plotting before you quote it. For where any of these figures come from, and why two honest sources can disagree by a factor of three, start with our guide to prediction market data.