News

From $100m to $13bn: charting the prediction-market growth curve

Monthly trading volume on prediction markets grew 130× in 24 months. We break down the curve, what drove it, and which parts of it are durable.

← News

Note, November 15, 2026: the volume figures in this article are estimates. We have not been able to trace them to a primary source and are re-checking them.

TL;DR:

Aggregate monthly volume on Polymarket and Kalshi climbed from under $100m in Q1 2024 to north of $13bn by December 2025 — a 130× expansion driven by the 2024 US election cycle, the listing of sports event contracts, and a step-change in retail mobile usability. The pace will not continue at this rate.

Key takeaways
  • Monthly volume grew roughly 130× between Jan 2024 and Dec 2025 (Source: NEXT.io, March 2026).
  • Three discrete drivers: 2024 US election, sports event contracts, and mobile UX parity with sportsbooks.
  • December 2025 ≈ $13bn was a cyclical peak; H1 2026 will test the durable base.

What the curve actually looks like

Aggregate monthly notional volume across Polymarket and Kalshi traced an S-curve with three distinct phases.

Phase one — pre-election base (Jan 2024 to Aug 2024). Monthly volume oscillated between $60m and $400m. The audience was small, crypto-native, and concentrated in politics and culture markets.

Phase two — election surge (Sep 2024 to Nov 2024). The 2024 US presidential cycle pulled in mainstream attention and pushed monthly volume above $3bn in October and over $5bn in November. Polymarket dominated this phase.

Phase three — sports inflection (Dec 2024 to Dec 2025). Kalshi listed sports event contracts and rapidly closed the volume gap. By Q4 2025, aggregate monthly volume crossed $13bn, with sports accounting for the majority share. (Source: NEXT.io, March 2026.)

Why the rate slows from here

A 130× expansion in 24 months is not a sustainable run rate; it's a category being discovered. Three forces will compress the slope.

The election bump unwinds. US political volume in non-election years runs at roughly 15–25% of election-year levels in equivalent markets. Without a 2026 federal election of comparable salience, 2024 totals are an unreliable baseline.

Sports contracts hit a regulatory ceiling. State cease-and-desist letters have already removed access in several jurisdictions and are working through the courts. Even a partial loss for the venues caps the addressable market until federal preemption is resolved.

Marginal-user economics get harder. Phase three users were acquired largely through earned media and category novelty. The next cohort costs more to acquire and trades smaller tickets.

The 2026 base case

Where 2026 volume lands depends on two outcomes:

  • A favourable federal ruling that protects sports event contracts → upper band.
  • An adverse Supreme Court ruling or expanded state action → lower band.

Either outcome resolves much of the uncertainty around the category's medium-term trajectory.

Why volume and sportsbook handle are hard to compare

Comparing prediction market volume with sportsbook handle flatters prediction markets in three ways (notional volume is not handle, win rates differ, retention curves differ) and discounts them in one (sportsbook handle excludes the illegal market, which is several multiples larger).

What we watch

  • Monthly Kalshi sports volume share — a leading indicator of cease-and-desist effectiveness.
  • Polymarket monthly active wallets — a measure of net new user growth versus churn.
  • Prediction-market take-rate per dollar of notional — a measure of how much of this volume is actually monetisable revenue.
The European lens

European betting-exchange volume followed a slower but structurally similar curve in the early Betfair era (2002–2008). The lesson from that arc: a one-time category surge driven by novelty and a marquee event does not predict the steady-state monthly volume. Steady-state arrives 18–36 months later and typically settles at 20–40% of the peak. Applied to US prediction markets, that points to a 2027–2028 steady state somewhere between $3bn and $5bn per month in dollar terms — still a serious category, but smaller than the cyclical headline.

#category

#explainer

Related