Prediction Market Arbitrage Shrinks as AI Traders Enter
In brief
- Kalshi and Polymarket combined for $13.7B in June volume, with July exceeding $11B
- Institutional volume on Kalshi climbed 800% as prop firms and quant shops deploy event contracts
- AI agents and market makers compete in real-time, narrowing retail arbitrage edges
Volumes and institutional inflows
Combined monthly volume across Kalshi and Polymarket reached $13.7 billion in June, with July already registering over $11 billion. Kalshi's annualized volume more than tripled over six months to $178 billion, while institutional volume climbed 800%.
The infrastructure connecting institutions to these markets is now in place. Clear Street connects institutional clients to Kalshi, Marex works across both Kalshi and Polymarket infrastructure, and Jump Trading helps institutions reach event markets directly. AQR, Susquehanna, and OKX have advertised specialist prediction-market roles, signaling serious commitment to the space. Beyond traders, corporate treasuries are testing these same contracts to hedge tariff and regulatory exposure.
The edge-detection problem
Louis Régis, founder of the on-chain prop firm Propr and a former quantitative trader at Credit Suisse, argues that event contracts reward skill differently than conventional markets. "Event contracts make trader selection more rigorous than conventional markets do, as the skill they reward is legible and the risk is bounded," he said.
But legibility cuts both ways. The Foresight Arena benchmark estimates that detecting a real edge of two percentage points with reasonable statistical confidence takes about 350 resolved binary predictions. Confirming a one-point edge takes roughly four times as many predictions. That's a high bar for retail traders operating in isolation.
Propr's model and the AI expansion
Propr plans to extend its evaluation model to Polymarket, letting traders and AI agents qualify for accounts up to $100,000 and hold as much as $300,000 across multiple accounts. Propr offers an 80% profit share to traders once they pass qualification, with payouts settling on-chain in USDC regardless of booking method.
Market makers, quantitative firms, funded-trading shops and AI agents now watch prices, compare related contracts and update probabilities around the clock on prediction markets. That constant surveillance is what's compressing the gaps. When the Fed's July contract showed 87% probability on holding rates at 3.50% to 3.75%, with roughly $29.7 million in volume, it was already priced by algorithms that had run the same calculation milliseconds earlier.
The era of casual prediction-market profits is ending. Retail traders who want to compete now need either a genuine statistical edge, or access to the same infrastructure and capital that institutions are deploying.
