Two billion dollars.
That's how much prediction markets had committed to World Cup contracts before Mexico kicked off against South Africa on Wednesday. The number ran through analyst notes, CNBC segments, and Kalshi press releases. Bernstein called the tournament a "watershed moment" and projected $5–10 billion in total prediction market volume before it ends.
US sportsbooks, by comparison, project $2.8–4.3 billion in total handle for the entire tournament.
So prediction markets are bigger than sportsbooks now? No. The numbers aren't measuring the same thing.
The Conversion
When a bettor wagers $100 on the USA at a sportsbook, that's $100 of handle. When a trader buys a "USA wins" contract on Kalshi at 72 cents and later sells at 85 cents, the platform counts both sides. Entries, exits, price adjustments — each generates volume. A single underlying bet produces multiple volume events.
H2 Gambling Capital ran the most rigorous comparison available. After the Super Bowl, they stripped out non-comparable segments — season-long futures, halftime entertainment props (including over $100 million wagered on Bad Bunny's first song) — and found prediction markets generated $633 million in actual Super Bowl bets. Apply standard single-game conversion and the handle equivalent was approximately $317 million.
Licensed sportsbooks generated $1.412 billion. In states where both prediction markets and sportsbooks compete head-to-head, prediction markets captured 7% of handle. Not the 114% that raw volume numbers suggested. Seven percent.
The World Cup numbers follow the same pattern. Gaming America calculated that Kalshi's $123 million in World Cup winner volume equates to just $5.4 million in sportsbook handle — a 23-to-1 ratio, inflated further by lopsided odds on favorites.
What the Volume Hides
So prediction markets are a rounding error? Also no. The volume inflation is real, but three structural facts live outside the handle comparison.
Geographic monopoly. California, Texas, and Florida — roughly 85 million people — have no legal sportsbooks. Kalshi operates in all three under its CFTC license. ADI Predictstreet, FIFA's first-ever official prediction market partner, offers World Cup contracts via Fanatics Markets in 23 states including all three. H2GC found prediction markets "over-index" in states without sportsbook competition. Most of that Super Bowl action didn't come from New York or New Jersey. It came from markets where prediction platforms are the only game.
Industry convergence. DraftKings invested $300 million building its own prediction exchange. Coinbase hit $100 million in annualized prediction revenue within two months of launch. Robinhood acquired a CFTC-registered exchange. Every major platform in gambling, crypto, and fintech is building prediction market capability — except BetMGM.
Regulatory patchwork. Three federal bills now target prediction markets — one banning sports contracts outright, one regulating them with state authority and insider-trading rules, one targeting government officials. The AGA, 100,000 Unite Here casino workers, and 60+ tribal nations are filing legal challenges. The most likely outcome isn't extinction. It's a state-by-state patchwork, mirroring how sports betting itself legalized from 2018 onward.
The Pecking Order
Deutsche Bank published World Cup handle projections by operator. FanDuel: ~$1.3 billion. DraftKings: ~$1.1 billion. BetMGM and Caesars combined: $300–500 million.
BetMGM's World Cup problem isn't prediction markets stealing customers. It's that BetMGM is already a distant third among sportsbooks — and the only major operator not building exchange capability.
Meanwhile, the platforms everyone compares prediction markets to — Kalshi ($22 billion implied valuation), Polymarket ($20 billion) — are valued higher than the sportsbook operators they supposedly compete with. Flutter's US operation: $18 billion. DraftKings: $11 billion. Companies generating 7% of handle are valued at 200% of the operators generating the other 93%. The denominator mismatch has infected the valuation market too.
Which Story
Choose volume, and prediction markets look enormous — a $2 billion World Cup phenomenon that makes sportsbooks look like legacy infrastructure. Choose handle, and they shrink to 7% of the competitive market, over-indexing in states where they face no competition.
Both are true. Neither is the point.
For BetMGM, every denominator tells the same story. In volume: Kalshi's reported Q1 sports revenue of $304.9 million exceeds BetMGM's $203 million — though if Kalshi's figure measures volume rather than net revenue, the comparison is exactly the denominator problem this piece is about. In handle: BetMGM is third and losing share. In capability: every competitor is building exchange products while BetMGM isn't. In geography: prediction markets reach 85 million Americans that BetMGM cannot.
The question isn't whether prediction market volume is inflated. It is. The question is what BetMGM's management thinks they're optimizing for — market share in a category that may not exist as a standalone business in five years, or a bridge to the converging market that everyone else is already building.
There is no denominator where that question goes away.