Polymarket zero-fee liquidity scaling

Running at zero trading fees was a deliberate way to buy market depth before trying to harvest revenue. Polymarket used crypto rails, USDC settlement, and an open API model to attract global traders and market makers first, then layered in taker fees only after sports and crypto markets were active enough to absorb them without blowing out spreads. That sequence matters because prediction markets are only useful when orders fill fast and price gaps stay tight.1 2 3 4 5

• The no fee period helped Polymarket solve the cold start problem. In 2024 it grew to about $9B in trading volume, largely on politics, then broadened into sports and other categories. By May 2026 it was doing $8.9B in monthly volume with a blended effective take rate of about 0.3%, still far below Kalshi at about 1.2%.2 3 4
• Polymarket kept maker fees at zero even after monetization began. Official fee docs show taker fees fund maker rebates, which pays traders who post resting orders and helps keep books deep. In plain terms, Polymarket charges the impatient trader, then uses part of that money to subsidize the traders who make prices available.5 6
• This was also a distribution strategy, not just a pricing strategy. Prediction market infrastructure builders describe Polymarket as unusually open, with public on chain trade data and APIs that let analytics apps, copy trading tools, and order routers plug in. That openness helps volume travel to Polymarket even when the user experience sits inside another app.2 4 7

The next phase is a race to turn liquidity into durable margin. As more apps, sportsbooks, and brokerages launch prediction products, the winners are likely to be the venues that stay deepest and cheapest at the core. Polymarket has already moved in that direction by adding fees selectively, preserving free maker liquidity, and turning its market data into an institutional product through ICE.3 5 8