Combos, Sports and Crypto Drove Prediction Markets’ Record September as Politics Lagged

The prediction markets industry experienced unprecedented growth in September 2026, with total nominal trading volume reaching $75.5 billion across four major platforms. This record-breaking month occurred despite the US midterm elections approaching, with political contracts generating only $901 million in total volume.

According to Prediction Frontier's third-quarter report, sports, crypto, and combo contracts accounted for 94% of September's activity. The monthly record coincided with the start of the NFL and college football seasons, highlighting the growing importance of non-political events in prediction markets.

Short Contracts Create Repeat Trading

Sports provide a continuous calendar of matches, while short-duration crypto contracts allow traders to take positions on repeated price movements. Kalshi, for example, offers contracts on whether cryptocurrency prices rise or fall over 15-minute windows, enabling traders to redeploy proceeds after each position settles rather than waiting for distant election results.

Combos work differently by combining several outcomes into one position, similar to a sportsbook parlay. Some combo contracts trade for only a few cents, which affects reported scale. A contract with a maximum $1 payout contributes $1 to nominal volume even if the buyer pays only a few cents, creating substantial headline turnover from relatively small purchases.

Kalshi's Market Dominance

Kalshi captured 71% of third-quarter dollars paid across the platforms, with its share rising from 69% in July to 73% in August and September. The company's combo contracts alone generated $34.2 billion in September nominal volume. However, Kalshi's average purchase price per $1 contract fell from 32 cents to 24 cents over the same period, attributed to the popularity of its combo offerings.

The falling purchase price means each dollar spent generated more nominal volume.

While Polymarket US handled more than half of September's political trading across all platforms, the industry reached record levels without an election-driven surge. This suggests prediction markets are becoming less dependent on political events and more diversified across different categories.