Prediction markets moved more than $15 billion in trading volume in the first half of 2026, up from roughly $3 billion in all of 2023. That explosive growth has attracted a new contender: Pascal, a New York-based startup that just raised $9 million in seed funding from Wintermute and Union Square Ventures to challenge the two dominant players, Kalshi and Polymarket. The bet is that the prediction market category is still in its infancy, and the real opportunity lies not in copying what exists, but in building for the next wave of AI-driven event resolution and institutional demand.
Why Wintermute and USV Are Betting on a Third Player
Wintermute, one of the largest crypto trading firms in the world, and USV, a venture firm that backed the earliest iterations of decentralized finance, are not late to this party by accident. They are betting that the prediction market category is about to fragment. Kalshi dominates the CFTC-regulated segment, offering event contracts on everything from Fed interest rate decisions to hurricane landfalls, with over $2 billion in notional volume since its launch. Polymarket leads the decentralized, blockchain-based segment, processing more than $500 million in monthly volume during the 2024 election cycle alone. But both platforms were built for different eras of the market. Pascal's thesis is that a generation of new financial instruments, powered by AI agents that can verify and resolve event outcomes in real time, requires a fundamentally different architecture. Think of it as the difference between building a marketplace for stocks in 1995 versus 2025. The underlying concept is the same, but the technology stack, user experience, and target audience are completely different. Pascal is building for the latter.
How AI Changes the Prediction Market Equation
The biggest bottleneck in prediction markets today is resolution. Every contract requires someone, or some process, to determine whether an event occurred. Kalshi relies on designated reporting sources and manual verification for complex events. Polymarket uses decentralized oracles and community voting through the UMA protocol. Both approaches work, but they introduce latency and limit the types of events that can be traded. Pascal is betting that AI agents, specifically large language models trained on real-time data feeds, can resolve events faster and across a broader range of categories. If an AI agent can ingest news feeds, government data releases, satellite imagery, and social media signals simultaneously, it can resolve a contract on something as granular as whether a specific company will announce a product launch within a quarter, or whether a particular bill will pass a committee vote, within minutes of the event occurring. This reduces the operational overhead of running a prediction market and opens up thousands of new event categories that were previously too costly or too slow to resolve manually. For founders building in adjacent spaces, this is the same pattern that transformed ad tech when programmatic bidding replaced manual insertion orders, and it is happening again in financial infrastructure.
What the Prediction Market Landscape Looks Like in 2026
The prediction market space has matured significantly since the Polymarket boom of 2024 and the Kalshi regulatory breakthrough in 2023. Kalshi is now the only CFTC-regulated exchange offering event contracts directly to retail users, and it has expanded beyond political betting into finance, weather, sports, and entertainment. Polymarket, despite regulatory headwinds, remains the global volume leader on blockchain rails, processing over $6 billion in annualized volume. New entrants have tried to break in. Metaculus focuses on long-term forecasting for science and technology. Manifold Markets operates as a play-money platform focused on community-driven questions. But none have seriously challenged the top two. Pascal is different because it has both regulatory intent and crypto-native backers. The company has publicly stated its intention to work within the CFTC framework, similar to Kalshi, while leveraging Wintermute's deep liquidity infrastructure and USV's regulatory playbook from their Coinbase and Filecoin investments. If Pascal can carve out a niche in AI-verified event contracts, it could capture the fastest-growing segment of the market without directly competing on Kalshi's core political and financial events or Polymarket's crypto-native user base.
What This Means for Founders
The $9 million seed round for Pascal signals something larger than one company's funding. Prediction markets are becoming a legitimate category for venture investment, and the infrastructure layer is commoditizing. For founders, three takeaways stand out. First, if you are building any product that relies on real-world data verification, from insurance to supply chain finance, the AI resolution technology Pascal is developing will become available as an API. Watch for infrastructure spins from this space in the next 12 months. Second, the prediction market model is spreading beyond finance. The same concept of crowdsourced probability estimates is being applied to corporate forecasting, scientific consensus tracking, and even project management inside enterprises. Companies like Gnosis and Azuro are already providing the underlying contract infrastructure for these use cases. Third, the Wintermute and USV combination signals that institutional capital sees prediction markets as a natural extension of both crypto infrastructure and traditional finance. If Pascal succeeds, it will open the door for more regulated, institution-friendly prediction market platforms. If it struggles, the technology it builds for AI-driven resolution will still find a home elsewhere. Either way, the prediction market category is not a two-player game anymore.