Sportsbooks Pay a Steep Price for Prediction Markets

DraftKings will spend $650 million-$700 million of adjusted EBITDA building its Predictions business this year, and the third quarter is where the bill showed up. The company that runs DraftKings Sportsbook cut its full-year profit guidance on November 5, a day after FanDuel’s parent Flutter Entertainment reported that its US business made just $3 million of adjusted EBITDA in the quarter.

Two of the biggest US sportsbooks are paying to get into a product their own industry calls a legal backdoor to sports betting without the state taxes. BetMGM and Caesars Digital, whose third-quarter reports say nothing about prediction markets, grew profits in the same quarter. Here is what an event contract is, why the sportsbooks want one, what it has cost them so far and who is coming out ahead.

What a Prediction Market Sells

A prediction market sells event contracts, which are yes-or-no positions on whether something will happen, from an NFL result to an election. They’re overseen by the Commodity Futures Trading Commission, the federal agency known as the CFTC, and not by state gaming boards. DraftKings runs its version through a subsidiary, GUS III LLC, which offers “federally regulated event contracts under CFTC oversight,” according to its earnings release.

The business model is different on paper. A sportsbook is the house and sets the line. Kalshi, the biggest of the exchanges, makes its money from a fee on each contract instead. “Whereas a sportsbook, it is the house, they set the line,” Sara Slane, Kalshi’s head of corporate development, told CPR News in May. “And so it’s just a much different business model.”

To a customer, though, it’s hard to tell the two apart. “From a bettor’s standpoint, Kalshi and sportsbooks are basically identical,” Victor Matheson, an economics professor at the College of the Holy Cross, told NPR.

Why Sportsbooks Want In

The short answer is reach. A sportsbook needs a license from each state, and DraftKings is live with mobile sports betting in 27 states, Washington, DC and Puerto Rico, which it puts at about 53% of the US population. An event contract under federal oversight can be sold in states with no legal sports betting at all, including California and Texas.

It also reaches younger customers. “Kalshi and Polymarket are signing up customers in states where sports betting isn’t allowed, like California and Texas, and signing up 18-year-olds in states where the legal age to sports bet is 21,” said Brad Allen, an analyst at Eilers & Krejcik Gaming, in the NPR report. “It’s broadly new money, money sportsbooks couldn’t touch.”

The Tax Gap

Then there’s tax. Licensed sportsbooks pay state taxes on gross gaming revenue, which is what they keep after paying winners, at rates that run from 6.7% in Nevada to 51% in New York. NPR put the total at about $4 billion a year. Kalshi says it doesn’t have gross gaming revenue to tax, so in Colorado, for example, it doesn’t pay the state’s 10% sports betting tax.

Flutter spelled out the strategy in its third-quarter letter. FanDuel wants to “use the strength of the FanDuel brand to acquire customers and build a meaningful database in non-sportsbook states,” and it says those customers look like typical sportsbook players but cost less to acquire. It’s the same logic that drove the DraftKings and FanDuel prediction pivot in the first place.

DraftKings Writes the Biggest Check

DraftKings grew revenue 16% to $1,329 million in the third quarter. It still lost money. The adjusted EBITDA loss was $316.9 million, against a $126.5 million loss a year earlier, and the net loss was $468.7 million. Sales and marketing spending jumped to $605.7 million from $360.4 million.

The guidance tells the bigger story. In May the company set full-year adjusted EBITDA guidance of $700 million-$900 million, and it held that range in August. On November 5 it cut it to $175 million-$275 million, “primarily reflecting the Company’s increased Adjusted EBITDA investment in Predictions of $650 million to $700 million.” Revenue guidance fell from $6.5 billion-$6.9 billion to $6.35 billion-$6.55 billion.

Spending Now, Revenue Later

Predictions won’t pay its way this year. DraftKings says its revenue “will be immaterial in fiscal year 2026, as new customer promotions will offset the lion’s share of gross revenue generated.” Another $175 million-$225 million of Predictions investment is planned for the fourth quarter alone.

Strip that spending out and the core business is still earning. DraftKings expects $875 million-$925 million of adjusted EBITDA from the core in 2026, which is the money now paying for the new product.

The 2027 Pitch

Management thinks the spending pays off. For 2027 DraftKings guided to revenue of $7.35 billion-$7.75 billion and adjusted EBITDA of $800 million-$1.0 billion, with Predictions bringing in $650 million-$750 million of revenue while still absorbing $250 million-$350 million of investment.

Chief executive Jason Robins said more than 1.6 million customers have used Predictions so far this year. “The third quarter marked a pivotal moment for DraftKings, as the Predictions offering we have been building is now powering meaningful growth for our business,” he said. Chief financial officer Alan Ellingson said the core business is funding the push “at attractive payback periods.”

FanDuel Pays in a Different Way

Flutter’s US segment, home of FanDuel, grew revenue 6% to $1,455 million in the third quarter. Adjusted EBITDA fell 94% to $3 million from $51 million, which Flutter put down to “investment in our prediction market offering and Alberta launch.” Brazil did most of the damage to Flutter’s third-quarter loss at group level, but the US number is the one that matters here.

FanDuel Predicts is still small. It averaged 66,000 monthly players in the quarter and handled $119 million of volume, up 102% from the second quarter, and its revenue “was not material.” October volume was 271% higher than July’s after FanDuel moved its sports contracts to Crypto.com and put Predicts inside its main app.

A Guidance Trail That Keeps Falling

Flutter’s US profit target has dropped twice since the summer. In August it cut US adjusted EBITDA guidance by $210 million to a $760 million midpoint, mostly to put more money behind its sportsbook offer. In November it cut again, to $715 million, blaming sports results that went against it. That August update also warned that Predicts revenue in the second half would be “offset by customer acquisition investment costs.”

The less obvious cost is advertising. Chief executive Dan Taylor wrote that FanDuel is “beginning to see an impact on our overall media costs as prediction market operators increase marketing spend.” Flutter expects its prediction market investment to continue in 2027 “at similar or higher levels.”

The Two That Stayed Out

BetMGM, which is owned by MGM Resorts and Entain, had a quiet, profitable quarter. Net revenue rose 7% to $710 million and adjusted EBITDA more than doubled to $86 million from $41 million. It reiterated full-year guidance and expects to land at the lower end of $2.9 billion-$3.1 billion in revenue and $300 million-$350 million in adjusted EBITDA. Entain also stuck to its own 2026 guidance in October.

Caesars Sportsbook and the rest of Caesars Digital grew net revenue 4.8% to $326 million, and adjusted EBITDA rose to $44 million from $28 million. Neither BetMGM’s update nor Caesars’ release mentions prediction markets.

Licenses Come First

Their bosses explain why. Nevada and other state regulators have told licensees that event contracts are wagers under state jurisdiction, and that offering them can put a license at risk even if the betting happens in another state, Bloomberg reported. “We will not put any of our licenses at risk,” Caesars chief executive Tom Reeg said on a call in October 2025.

The tone hasn’t softened. At the Global Gaming Expo in September, MGM Resorts chief executive Bill Hornbuckle had a message for the exchanges. “If they want to play by the rules, they want to come to Nevada and get a license and do what we all do, then God bless them.”

The scorecard comparison isn’t perfect. Online casino is BetMGM’s biggest product, bringing in $483 million of its $710 million, and Caesars runs a much smaller digital business. But the split in the third quarter is hard to miss.

Who Is Winning So Far

The exchanges are, at least on valuation. On October 29 Kalshi announced a $1 billion Series G at a $40 billion post-money valuation, co-led by Wellington and Sequoia. It says more than 30 million Kalshi accounts have been opened, that it handles over 75% of prediction market trading volume and that its annualized non-sports revenue is now more than $1 billion.

For the sportsbooks, the case rests on the long game. Robins has said prediction markets do best in states without legal sports betting, such as California and Texas, and DraftKings’ bet is that those customers will be worth more later. FanDuel says the impact on what its existing sportsbook customers bet has been limited. Both are spending heavily to find out.

The Regulatory Wildcard

All of this sits on an unsettled legal question. Whether event contracts are gambling is at the center of more than 20 pending federal lawsuits, and the answer decides whether the state gambling laws that bind the sportsbooks apply to their Predictions apps too. Flutter says the duration of the marketing pressure will depend in part on how regulation develops, and DraftKings’ 2027 plan assumes it can keep selling Predictions at scale.

Neither company expects the spending to stop. DraftKings has budgeted $250 million-$350 million for Predictions in 2027, and Flutter expects to spend at similar or higher levels than this year.