The sportsbook industry spent half a decade chasing market share like it was oxygen. Then it promised Wall Street it had grown up. Prediction markets just ripped that script in half, and the NFL season is about to prove it.
Jay Snowden saw it coming on PENN’s Q2 call. “We’re anticipating that there could be a bit of an arms race as we head into football season,” the PENN CEO said, before sharpening the blade: competitors are locked into “very aggressive, irrational” spending as prediction products hit their first full football campaign. PENN is sitting this one out. DraftKings and FanDuel are not. That gap is the entire story.
I keep coming back to the mechanism, not the press releases. Traditional handle growth has cooled off the double-digit pace that defined the early PASPA years. Operators need a new customer funnel. Prediction markets, regulated under the CFTC rather than state gaming commissions, give them one that reaches states where online sportsbooks remain illegal. FanDuel frames FanDuel Predicts exactly that way: acquire the relationship now, convert later when the jurisdiction flips. DraftKings built a universal app that auto-switches users between sportsbook and predictions based on legality. That is not a side product. That is a customer acquisition engine with different regulatory friction.
The Cannibalization Story Does Not Survive Contact With the Tape
DraftKings’ Q2 numbers make the spend thesis look rational on paper. Customer acquisition jumped nearly 75 percent year over year. Sports-consumer volume, the blended sportsbook-plus-predictions figure, rose 15 percent. More than 600,000 customers touched the predictions product in the first half. Annualized trading volume roughly quintupled from April to July. Management claims customer overlap with the leading pure prediction operator sits around 1 percent in legal sportsbook states, and estimates that 80 to 90 percent of prediction volume inside those states comes from professional syndicates and institutional traders, not the recreational bettor who used to define their book.
If those figures hold, prediction markets are not stealing sportsbook handle. They are attaching a different user base and, crucially, planting flags in dark states. That is why DraftKings is putting an incremental $200 million to $300 million into Predictions this year. Flutter is staring at $250 million to $300 million in category expenses against roughly $50 million in market-making revenue. Leading operators are projecting more than $500 million in combined adjusted EBITDA losses just from building and marketing these platforms. That is not a hobby budget. That is a second land grab.
I have watched this movie before. The early U.S. sportsbook expansion burned billions on bonuses, celebrity spots, and stadium naming rights while unit economics stayed ugly. The industry spent 2024 and 2025 talking about discipline, LTV over CAC, and profitable growth. Then the World Cup dropped a record $20 billion in prediction-market volume, Super Bowl LX cleared more than $1 billion on the prediction side against the AGA’s $1.76 billion traditional sportsbook estimate, and the arms race instinct returned. Digital ad impressions for online sportsbooks fell nearly 14 percent in 2025 while prediction and event-contract ads surged. Roughly 15 percent of the sports betting ads consumers saw last year did not have to clear state responsible gaming rules. Different regulatory lane, different cost structure, same fight for attention.
Football Season Turns the Test Into a Stress Test
NFL is the industry’s highest-leverage acquisition window. Preseason already started. Kickoff is September 9. Every promo dollar, every app-install bid, every lobby takeover is about to get more expensive because three large operators plus Kalshi and the rest are bidding against each other for the same September attention. PENN narrowed its Interactive adjusted EBITDA loss to $9.5 million from $62 million a year earlier and is guiding a full-year Interactive loss of about $20 million. Snowden is betting that staying disciplined while rivals torch cash will look smart by February. The rivals are betting that owning the prediction customer, especially the one who lives in a state without a legal book, is worth the near-term bleed.
For bettors, the surface effect is obvious: juicier deposit matches, more risk-free entries, louder creative. The deeper effect is structural. If 80 to 90 percent of the prediction flow is professional, the recreational user is still the marketing target even when the volume is not. Operators will keep subsidizing the casual side to build the brand while the sophisticated side provides the liquidity. That is the same dynamic that turned early sportsbook books into loss leaders wrapped around a hope of lifetime value.
I am not convinced every dollar of this spend converts. Maturing markets punish irrational CAC. But I am convinced the spend is real, the NFL window will make it visible, and the operators who treated prediction markets as a temporary sideshow misread the product. DraftKings and FanDuel are not defending a sportsbook. They are buying optionality on the next regulatory map. PENN is defending a balance sheet. One of those approaches will look like strategy in twelve months. The other will look like the last cycle all over again, just with different contracts and fewer state compliance disclaimers.
The industry said it was done with arms races. The earnings calls just said otherwise. Football is about to keep score.
