America's largest online gaming operators are projecting more than $500 million in combined lost adjusted EBITDA tied to building and marketing prediction market platforms ahead of the 2026 NFL season. That is not a marketing budget line — it is a deliberate, publicly disclosed hit to profitability, taken simultaneously by companies that spent the previous two years telling investors that the era of unprofitable customer acquisition was over.
Something changed. Understanding what it was matters for anyone who bets, and for anyone watching how the wider gambling industry allocates capital.
The Numbers
Legal US sportsbooks are projected to take a record $32.3 billion in NFL handle this season. Prediction markets are projected to trade $36.8 billion on NFL outcomes — more than double last season's volume.
Read those two figures together and the strategy explains itself. Prediction markets are on track to move more NFL money than the entire regulated sportsbook industry, in their second serious season. Meanwhile sportsbook handle, while hitting a new high, is growing at its slowest year-over-year rate since the Supreme Court cleared the way for nationwide sports betting in 2018.
That is the shape of a maturing market being flanked by a faster-growing adjacent one.
Why Operators Are Paying $500M to Compete
The instinctive reading is that prediction markets are stealing sportsbook customers. The more careful reading — and the one operators themselves seem to be working from — is that prediction markets are reaching people sportsbooks legally cannot.
Prediction markets operate under CFTC oversight as federally regulated exchanges, which means they are available nationwide, including in states with no legal sports betting. Texas and California are the obvious prizes: two of the three largest state economies in the country, tens of millions of adults between them, and no regulated sportsbook access.
For an operator that has spent seven years and enormous sums lobbying unsuccessfully for legalization in those states, a federally regulated product that simply works there is not a competitive threat. It is a market-access solution.
That reframes the $500 million. It is not defensive spending against a rival; it is the entry cost into markets that were previously closed.
What the Operators Are Building
Flutter is integrating FanDuel Predicts into a unified "one app" before the regular season, combining sports betting, casino, fantasy sports and prediction markets in a single product. DraftKings has launched its own super-app incorporating DraftKings Predictions.
The one-app strategy is the interesting part. Rather than running prediction markets as a separate product for a separate audience, both companies are folding them into the existing customer experience. A user in New Jersey sees prediction contracts alongside traditional bets; a user in Texas sees only the contracts. Same app, different regulatory surface.
Penn Entertainment CEO Jay Snowden set expectations bluntly on an analyst call, saying the company anticipated "a very aggressive, irrational marketing spend, advertising, and new customer acquisition approach this football season." When an operator uses the word "irrational" about their own sector's behaviour in advance, it is generally a signal they intend to participate anyway.
The Regulatory Fight Underneath
None of this is settled. Prediction markets' nationwide availability rests on a federal regulatory classification that multiple states are actively contesting, arguing that sports event contracts are functionally sports betting and therefore subject to state gaming law. Litigation has escalated across several states through 2026, and the CFTC's own posture has shifted more than once.
The $500 million bet is therefore also a bet on a legal outcome. If prediction markets' federal status holds, the spending buys entry to Texas and California. If it does not, operators will have spent heavily building a product whose primary competitive advantage disappears.
That is an unusual amount of regulatory risk to accept voluntarily. It suggests operators either have confidence in the federal position or have concluded that the cost of being absent if it holds exceeds the cost of being present if it does not.
What This Means for Bettors
Several practical implications, in rough order of immediacy.
Promotional value is about to spike. An acknowledged "irrational" acquisition war means unusually generous sign-up offers, boosted odds and retention promotions through the NFL season. For a disciplined bettor, promotional periods are the most reliable source of positive expected value in the entire sports betting product — the underlying markets are efficient, but the promotions frequently are not. Our sportsbooks guide covers what to look for, and reviews of BetOnline Sportsbook, SportsBetting.ag and TigerGaming Sportsbook cover the operators serving players who prefer offshore books.
Prediction market pricing is different from sportsbook pricing. A prediction contract is a two-sided market where you trade against other participants, with the exchange taking a fee. A sportsbook line includes a built-in margin — the vig — that you pay on every bet regardless of outcome. In theory, exchange pricing should be tighter. In practice it depends entirely on liquidity, and thin markets on an exchange can price worse than a competitive sportsbook line.
Understand what you are actually buying. A prediction contract settles at $0 or $1 and is priced between. Buying at $0.60 means the market implies a 60% probability. That maps directly onto implied odds, and if you are comfortable converting between formats, the comparison to a sportsbook line is straightforward. If you are not, you are trading a product you cannot price — and probability-to-odds conversion is the one skill worth learning before you place a single contract.
The Poker Parallel
Poker players watching this should recognise the dynamic, because poker lived through it.
The 2003-2006 online poker boom was funded by exactly this kind of acquisition spending: operators burning enormous sums to capture players in a market growing fast enough to justify it. The spending stopped when the growth did, and the operators who had built durable products survived while the ones who had bought traffic did not.
The relevant lesson for players is about timing. The best value available to a customer in any gambling vertical arrives during the acquisition phase, when operators are paying above rational value for signups and volume. That window is open now in US sports betting and prediction markets. It will not stay open, because $500 million a year in foregone EBITDA is not a permanent posture — it is a land grab with an end date.
The same logic applies to poker promotions. Sites competing for players offer better bonuses and rakeback than sites with settled market positions, which is why offshore rooms fighting for US traffic consistently offer terms that regulated markets do not match.
Where This Lands
Three scenarios, roughly:
Prediction markets hold their federal position. Operators get Texas and California access, the $500 million looks cheap in retrospect, and state-by-state sports betting legalization loses much of its urgency.
States win the litigation. Prediction markets get pushed into state-by-state licensing, their structural advantage disappears, and $500 million was spent building a redundant product.
Something in between. Partial carve-outs, negotiated frameworks, or a federal-state settlement that preserves prediction markets in some form with meaningful constraints. This is the likeliest outcome and the hardest to price.
For bettors, none of those scenarios argues for waiting. The promotional environment is at its most generous right now, during the uncertainty — and our payments guide covers how to move money safely while the spending war runs.