September’s record volume and a wave of institutional liquidity providers point to the same problem: moving capital across a fast-splintering map of venues.
In the last week of September, U.S.-facing prediction markets traded $20.4 billion in notional volume. It was the first week the category had ever cleared $20 billion. Kalshi alone logged its first $3 billion day on September 20.
For the month, volume across seven tracked venues reached $77.7 billion, up 53% from August on a like-for-like basis, according to DeFi Rate data compiled in Shift Markets’ September 2026 Prediction Markets Brief.
The headline number deserves the scrutiny it is getting. What it says about who is trading, and what those traders now need, deserves more. Read closely, September’s data shows a market whose growth is limited less by demand than by the speed at which serious capital can move between the places where that demand lives.
How prediction market volume is measured, and what notional volume doesn’t tell you
Start with the caveat that most coverage skips. These are notional figures: the face value of contracts traded, not revenue, not deposits and not capital at risk. Low-priced combination contracts inflate notional totals. Shift Markets notes that combos made up more than 62% of Kalshi’s volume in the week of September 14–20.
Even with that discount applied, the trajectory is hard to argue with. Pew Research Center reported that combined Kalshi and Polymarket volume roughly doubled between May and July, from $26 billion to $53 billion. September then beat July by a wide margin. Separately, a FinanceFeeds tally cited by law firm Troutman Pepper Locke put year-to-date volume at $410 billion as of mid-October.
| Venue (September 2026) | Notional volume | Share | Month-over-month note |
|---|---|---|---|
| Kalshi | $60.7B | 78% | Peak week $15.66B (Sept 21–27) |
| Polymarket US | $8.8B | 11% | Volume more than doubled |
| Polymarket (international) | $4.6B | 6% | Down 6% |
| Novig, Rothera, Underdog, ProphetX (combined) | ~$3.6B | 5% | — |
| Total, seven venues | $77.7B | 100% | +53% vs. August |
Source: DeFi Rate data as compiled in Shift Markets, “The Prediction Markets Brief – September 2026.” Notional, not revenue.
Kalshi vs. Polymarket: market share and the venues behind them
The table shows a market that is concentrated and fragmenting at the same time. Kalshi holds nearly four of every five dollars traded. Yet the venue count keeps rising. Polymarket’s U.S. exchange doubled in a month, and a second tier of exchanges together cleared billions.
How many prediction market platforms are there?
The pipeline is wider than the volume table suggests. In March, DeFi Rate counted 20 new Designated Contract Market (DCM) filings or designations since the start of 2025, from crypto-native exchanges to units owned by DraftKings, Underdog and a Robinhood–Susquehanna joint venture. New applicants keep arriving. CNBC reported on October 9 that Blockchain.com is the latest firm seeking CFTC approval to operate a prediction market.
Many of these venues list overlapping contracts. For a trader who wants the best price on an economic release or an election outcome, each new exchange is another order book worth watching. For a firm that provides liquidity, each one is another balance to fund.
Who are the market makers in prediction markets?
The more consequential shift in 2026 is less visible in volume tables: the arrival of professional liquidity.
Jump Trading, Susquehanna and Galaxy Digital were already making markets in event contracts when Wintermute said in late May that it had begun quoting two-sided markets on Kalshi and Polymarket, Finance Magnates reported. Prime brokers including Clear Street and Marex have followed with clearing on-ramps for hedge-fund clients, the outlet added. Galaxy has since launched an institutional over-the-counter desk for event contracts. BitGo Prime partnered with Susquehanna Crypto to bring prediction-market liquidity to institutional clients. Markets Media also reported that Kalshi’s institutional trading volume rose 800% over the six months before its May funding round.
Wintermute’s head of OTC trading, Jake Ostrovskis, summed up where that leaves the category: “Prediction markets have the demand profile of a major asset class but the liquidity profile of an early-stage one.”
Why prediction market liquidity lags volume
Market observers keep describing the same symptoms: wide spreads, thin depth and meaningful slippage on larger orders. Finance Magnates attributed the gap to liquidity failing to keep pace with volume growth. The structure of these markets explains a good part of why.
Most U.S. event contracts are fully collateralized. To hold a position, a trader posts the maximum possible loss up front. That keeps the clearing model simple and the system safe, but it makes market making capital-intensive. A firm quoting both sides of hundreds of contracts must keep real dollars on deposit at every venue where it quotes, sized for the worst case on each one.
The venues know this is a brake. In late September, Kalshi filed with the CFTC to let institutional participants post risk-based margin, rather than the full maximum loss, on some longer-dated contracts. The filing is under review, and it is the clearest sign yet that collateral, not appetite, is what limits depth.
The cost of pre-funding multiple prediction market platforms
Consider a simplified, hypothetical liquidity provider active on four venues with $10 million deposited at each. Over a weekend, flow concentrates on one exchange, and that book needs $18 million to keep quoting at size. The other three balances sit mostly idle.
The firm has three options, and all of them are bad. It can widen its quotes and give up flow. It can stop quoting on the busy venue. Or it can move money, which in most of the banking system means a wire, and Fedwire doesn’t run on Saturdays or Sundays. The cheaper fix of simply pre-funding every venue more heavily raises the cost of idle capital with every exchange added.
Multiply that across 10, 15 or 20 venues and the constraint becomes structural. Every new exchange adds another pool of capital that has to be funded before it can be used.
Prediction market infrastructure: where EDGE Markets fits
That gap has created a category of firms that do not run exchanges at all. Their job is the plumbing between trader capital and the venues. Among them is EDGE Markets, a New York-based fintech that raised a $29.2 million Series A in June led by CoinFund, with participation from Indicator Ventures, Mantis VC, StepStone Group and Bullpen Capital, according to the company’s announcement on PR Newswire.
In a June interview with AlleyWatch, founder and CEO Seni Thomas described EDGE as an infrastructure layer rather than another exchange or data terminal. His argument was that the more fragmented the exchange landscape becomes, the more essential capital-movement infrastructure gets. Its first product, EDGE Boost, a special purpose debit account and Visa debit card, had processed more than $2 billion as of the June funding announcement, according to the company.
On October 6, the company announced programmatic capital allocation and agentic access for EDGE Pro, its forthcoming institutional product. The features let institutions set permissions and capital limits so trading algorithms and AI agents can deploy funds within defined bounds. Execution partners River Markets, OpenMarkets.ai, ParlayX and Pikkit plan to integrate through EDGE’s APIs. “Capital efficiency is a key focus for many of our clients and partners,” River Markets CEO Oscar Levy said in the announcement.
EDGE is not alone in seeing the opportunity, and the market is still early. But the direction of travel is clear. As venues multiply, the competitive question shifts from where to trade to how fast capital can get there, which is why financial infrastructure for prediction markets has become a fundraising story in its own right.
What’s next for prediction markets
- The margin decision. Kalshi’s risk-based margin filing is under a 45-day CFTC review. If it is approved, institutions holding longer-dated positions would see collateral requirements change sharply. Their funding needs would also become more dynamic, since margin can rise ahead of scheduled events.
- Election-season flow. More than $750 million in notional volume had traded on 2026 midterm contracts as of October 15, nearly eight times the total on 2024’s congressional and state races, according to CNN figures cited by iGaming Business.
- 24/7 products. Kalshi added non-expiring gold and silver contracts in September that never stop trading. Products that never close put more weight on funding that never closes.
- Rulemaking. The CFTC issued a proposed rule on October 9 that would broaden the definition of swaps to include certain event contracts, Bloomberg reported, and Troutman Pepper Locke notes that appeals courts are split on jurisdictional questions. Outcomes are unsettled. Troutman Pepper Locke partner Stephen Piepgrass told FinanceFeeds he does not expect a Supreme Court resolution until at least summer 2027.
Demand is no longer the open question for prediction markets. September settled it. What comes next is an infrastructure test: whether capital can move as quickly and as continuously as the markets it is meant to serve.
