The prediction market boom has an infrastructure problem

When we co-led Pascal's seed round with DBA in August 2025, we did so on a simple conviction, which was that prediction markets were becoming a serious trading category, the venues serving them weren't built for people who actually trade, and Ivo Crnkovic-Rubsamen and Matthew Downey were the right founders to fix that. Now that Pascal has raised a $9 million Series A led by Union Square Ventures and launched their beta, we wanted to share why we backed the team and why we think it wins.

Prediction markets found their product-market fit

For years, prediction markets sat on the margins of finance. That has changed fast. Combined monthly volume across the two leading venues climbed from under $5 billion in September 2025 to roughly $24 billion by April 2026, more than Americans wagered through legal sportsbooks in an average month last year. The 2024 US election was the proof of concept. Sports, macro, and crypto markets turned a one-off spike into recurring volume. Institutional validation followed, with the NYSE's parent company committing up to $2 billion to Polymarket as US regulation broadly cleared rather than tightened.

The 2026 World Cup has since erased any doubt about scale. Combined volume hit a record $44.8 billion in June, up 75 percent in a single month, with Kalshi clearing over a billion dollars a day since kickoff on June 11. It is the clearest test yet of whether these venues hold up when real volume arrives.

But the infrastructure was never built for traders

Look beneath the headline numbers and the market structure is surprisingly thin. Two platforms account for the vast majority of all volume. That concentration is not a sign the category has been solved, but a sign it outgrew its infrastructure before anyone built specifically for traders. Order books are often thinner than the totals suggest, fees are high, and the retail experience is designed for browsing rather than trading, down to having to sign every order. The API that professional desks depend on tends to be an afterthought, when it is the surface that matters most in every mature market. And settlement still leans on mechanisms many traders do not fully trust.

Trading teams build the best exchanges

The best products are built by the people who use them all the time. Pascal's founders come from HFT and quant backgrounds, so they have spent their careers on both sides of the order book, building trading venues and trading against them. That shows up most in the part every user touches, the experience itself. People who trade all day know which numbers have to be right, where the friction hides, and how fast a screen has to respond before it becomes unusable. Most of that work is unglamorous, sitting in the API and the infrastructure beneath the interface, and it is the reason early users describe the speed and functionality of Pascal's front end as “night and day” against existing platforms. It is the same pattern that keeps repeating in crypto, where small, trading-native teams have outbuilt first-generation incumbents. FTX overtook BitMEX, and Hyperliquid overtook dYdX. Pascal is applying that playbook to prediction markets.

Better liquidity, better fees, built for traders

That heritage is visible in the product. Fees across the prediction markets category work in a similar way, as a constant scaled by the contract price, Pascal’s constant is 0.01 against 0.03 to 0.07 at Polymarket and 0.07 at Kalshi, so traders pay a fraction of the fee for the same trade on Pascal, with a 50% maker rebate on top. Pascal’s liquidity already matches Polymarket on most contracts. This means that anything you can trade on Polymarket you can typically trade on Pascal, at the same size, for less. The experience is built for people who trade, with fast execution, reliable data, visible order books, and the ability to watch several contracts at once, all sitting on an API treated as a first-class surface rather than an afterthought. It has also pushed past simple yes-or-no bets, listing markets on companies like Anthropic and OpenAI that trade more like futures, aimed at more sophisticated participants.

Why we backed Pascal

Wintermute Ventures is the Ventures arm of Wintermute, a global algorithmic trading group and one of the largest players in digital assets. That vantage point gives us an unusually direct view of the parts of an exchange that Pascal has chosen to prioritize, from liquidity depth to fee structure to the quality of the API. We know how much these fundamentals matter, and how rarely they are executed well, which is why we backed Pascal at seed and why we are glad to be part of Pascal’s journey into its next phase.

Everyone is starting to trade prediction markets. The only question left is whether the infrastructure can keep up. Pascal is being built so that it can, and it is worth seeing for yourself in their private beta at https://www.pascal.trade/.