Flutter after Q3: who keeps the $100?

The argument in brief

Of every $100 bet with FanDuel over the past year, about 65 cents stayed with FanDuel, on the firm’s estimate.

FY2026 group EBITDA guide, against $2,970m in February

$2,495m

Net debt / EBITDA, last twelve months

4.6x

Buyback

Paused

FLUT, close on 4 Nov 2026

$81.79

Firm's position

Long

Spider Eyes Research holds a long position in Flutter Entertainment plc.

Flutter owns FanDuel, the largest US sportsbook, and a large betting business outside the US. Its profit guide has fallen at every results date this year, its debt is high, its buyback has stopped, and its chief executive has been in the job for five weeks. So the question for anyone holding the shares is whether there is a way out.

By a way out the firm means three things: EBITDA, operating profit before depreciation and amortisation, stops falling in 2027; net debt falls below 4x EBITDA by the end of 2027; and it falls below 2.5x, the top of management’s target, by FY2029, when buybacks can restart.

The largest question is one Flutter doesn’t control. Kalshi is the largest of the prediction markets, exchanges that sell contracts on sports results under federal rules, outside state gaming law and state gaming tax. Federal appeal courts have split on whether states can stop them, and the Supreme Court has been asked to decide. So the firm values the shares on the three ways it can end: Kalshi wins, the courts stay split, or the states win. Its answer: there is a way out by about FY2029 if the courts stay split or the states win, narrowly in the first case. If Kalshi wins, it takes longer.

The firm stays long because the ruling is lopsided on its values: a states’ win adds more (+47%) than a Kalshi win takes away (−25%). It gives no odds on either, and it stays long until the picture is clearer. The next date is 7 December, when Kalshi’s response at the Supreme Court is due. That is a filing, not a decision.

How it got here

The quarterly tests kept being met while the year got worse.

Flutter’s test for the third quarter was US EBITDA of about break-even, and at $3m FanDuel passed it. The third quarter is the seasonal low, and the guide now needs about $474m of US EBITDA in the fourth, at its midpoint. Yet the US investment that was meant to be a phase now runs into 2027, and International’s organic revenue grew under 2% over nine months.

Part of this year’s fall is bad luck. Sports results have cost about $247m of EBITDA, on the firm’s sum of Flutter’s disclosures, with the fourth quarter counted only to 29 October.

Before about $200m of losses on its own prediction-market app, FanDuel makes about $915m this year, on the firm’s arithmetic. That is level with $922m in FY2025, and about $410m short of about $1,325m in February’s plan.

Who keeps the bet

Customers may have responded to a better price. One quarter can’t show it yet.

Before results and promotions, a FanDuel sportsbook customer loses about $14 of every $100 bet. That is the hold, and for three years FanDuel raised it, mainly with same-game parlays, bets that combine several outcomes from one game. Then the hold stopped rising and the customer started to resist. In the fourth quarter of 2025 Flutter itself linked its very high revenue margins to lower engagement into 2026.

In the third quarter FanDuel gave back 47% of the hold as promotions, up from 42% a year earlier, and handle rose 13%. The World Cup was about 6% of that handle, so growth without it was about 6%, on the firm’s arithmetic, with one fewer NFL week and some new states. That 13% came with one of the smallest rises in promotions of the past four quarters. Over twelve months handle rose only 1.6%, even with new states. FanDuel’s share of sportsbook gross gaming revenue has gone from 41% to 38% to 36% over the last three third quarters.

There are two readings. In one, FanDuel has hit a ceiling on what it can take, and needs ever more generosity to hold its place. In the other, which is management’s account, it has simply moved to the market’s level of generosity, a one-time reset to a lower margin, and growth then comes from volume. The firm treats the second as a hypothesis, and fourth-quarter handle, with no World Cup, is the next reading. All three outcomes below assume it. On the first reading, with margin slipping and underlying handle growth of 2.5%, not 4%, the split value is about $77 and leverage in FY2029 is 2.70x, on the firm’s model.

Prediction markets: the deciding question

The largest question for Flutter’s value is a ruling it doesn’t control.

To a customer a sports contract looks much like a bet. To its operator it is a federally regulated product, so it can reach states with no legal sportsbook and, it argues, states with one. FanDuel’s sportsbook needs a state licence and pays state gaming tax.

What has gone where

The firm found no measure of the handle lost state by state since January, so the 6% is the weakest number in this note, and it rests on one forecast. Prediction-market volume isn’t handle either: it counts contracts at $1 each, not dollars staked. Most of it doesn’t appear to come from legal books, but from states without one, offshore betting and new customers. Flutter says the impact on its customers is limited, and in February put it at low single-digit points of handle growth, on research it hasn’t published.

For FanDuel, 6% is about $3.5bn of handle, $280m of revenue and $100m of EBITDA, if it loses in proportion. Add the cost of its own app and prediction markets take about $300m a year from US EBITDA, on the firm’s estimate.

The competition can afford to stay. Kalshi raised money on 29 October at a valuation of $40bn, nearly double May’s, a private price, against about $14.2bn of market value for Flutter. PitchBook, as reported by InGame, estimates its 2026 adjusted EBITDA, before stock compensation, at $1.2bn. The firm could find no audited figures. But if that is even roughly right, the leader may be funding itself.

The law

The Third Circuit sided with Kalshi and the Ninth and Sixth against, and New Jersey has asked the Supreme Court to decide. Kalshi’s response is due on 7 December, and the Court could say early in 2027 whether it will hear the case. A ruling, if it does, is unlikely before mid-2027, on the firm’s reading. The petition has the NFL, the American Gaming Association and 39 states plus the District of Columbia behind it. Other courts may rule first: the Fourth and First Circuits and Massachusetts’s highest court have heard, or were due to hear, argument.

The map shows the ground in dispute. FanDuel’s sportsbook reaches states holding about half of the US population. Its app offers sports contracts in 18 more jurisdictions, with about 42%, on the latest list the firm found. On one tracker’s count, including one non-binding opinion, 14 of FanDuel’s 25 sportsbook jurisdictions have moved against these contracts, and 8 of the other 26. Most of the rest signed the brief for New Jersey. A states’ win would remove them from FanDuel’s own app too, leaving state law to decide.

Politics may matter later and stays out of the model. Early reports on 4 November show Democrats winning both houses of Congress, with some races uncalled. They could press the federal regulator on these contracts, but bills to bar them are a long way from law.

The three outcomes

The firm starts from 4% underlying handle growth, H2 Gambling Capital’s estimate for this NFL season, which it applies to all handle to FY2029. It takes off the World Cup in 2027, then varies four things: the handle diverted, the US margin, the loss on FanDuel’s own app and the exit multiple.

The firm assumes a states’ win reverses the decline rather than restoring the past: handle comes back over two years, and Kalshi loses its most profitable contracts wherever states enforce. If Kalshi wins, it assumes handle falls about 2% and stays flat for three years, and FanDuel gives up some margin to keep its customers.

FanDuel’s own answer

FanDuel is far behind, and growing quickly from a small base. FanDuel Predicts had 66,000 monthly players and $119m of volume in the third quarter against $12bn of US handle, with revenue Flutter calls immaterial, and it will cost about $200m of EBITDA this year. Taylor, the new chief executive, says it wasn’t competitive when he first looked at it in the spring. The firm doesn’t state a market share: Flutter’s volume and a broker’s share estimate can’t be reconciled.

The money may sit elsewhere. A market-maker quotes prices on other companies’ exchanges and earns the spread, and Flutter does this with the pricing it built for its sportsbook. This growth “will help fund our prediction market customer acquisition spend”, the release says. Flutter expects about $50m of revenue this year, about $40m of EBITDA, and has declined to raise that. If third-quarter revenue was about the $11m rise in other US revenue, the fourth quarter has to bring in over $30m, on the firm’s arithmetic.

How big could it get? Flutter won’t say, so the firm sized it, on about $180bn a year of sports volume in dollars paid, annualised from the broker Macquarie’s latest month, with 75% sports on the firm’s assumption. Macquarie’s forecast for the whole year implies nearer $140bn, about a quarter less.

Paying for this year’s roughly $200m of losses would take about $250m of revenue, a 14–18% share at a 1% spread. That is possible, but it isn’t what the firm assumes: about $170m by FY2029, which the model carries at 35 cents on the dollar, though Flutter’s own figures imply about 80. Trading firms such as Susquehanna have long practice at this, and a states’ win would probably shrink it. That is partly the point: market-making pays if the exchanges thrive, and the sportsbook gains if they don’t.

The new chief executive

He spoke for the first time as chief executive this quarter, and he was blunt about the group he now runs.

Dan Taylor took over on 1 October, succeeding Peter Jackson. He isn’t new to Flutter. He ran the International division, which earned more than $2.2bn of adjusted EBITDA in 2025, became Flutter’s President in May, and played what Flutter calls a central role in FanDuel’s sportsbook plan. So he is an International operator who now has to fix a problem that sits mainly in the US, and his own division’s organic revenue grew under 2% over nine months.

The Q3 release carried his first shareholder letter, and the same day he led his first call. The letter’s central sentence is that Flutter has fantastic brands and significant scale, “but we are not getting enough from these advantages”. He set four priorities: build on US leadership, capture the prediction-market opportunity, drive more value from International, and simplify the operating model, all of which “will require clear leadership and tough decisions”.

What he has done so far fits that. Flutter is cutting up to 1,400 UK and Ireland roles and has announced a further $500m of gross operating cost and capital expenditure savings by 2029. He argued against selling assets to speed up the deleveraging, for now.

What he didn’t give matters as much. There were no group or US profit numbers for 2027, and no targets. Those wait for his strategy in February. New chief executives have a reason to reset expectations before rebuilding them, so the firm wouldn’t rule out 2027 guidance below market expectations.

The 2027 guidance has a gap. Promotions at similar levels, marketing intensity elevated, prediction-market spending similar or higher, cost of sales unchanged: analysts pressed on the call on how that adds up, and did not get a direct answer. The firm’s split case assumes US EBITDA falls to about $659m in 2027, from $715m this year. That is the firm’s estimate, not Flutter’s, and February’s guidance will test it.

The governments

International is shielded from prediction markets, and exposed to tax.

Governments are re-pricing their share of the profit. By 2027 the steps Flutter has disclosed add up to about $1.07bn of annual EBITDA before its own responses and about $0.73bn after them, on the firm’s rough count and as a floor. The lower figure is about a quarter of FY2025’s group EBITDA.

The largest operator pays the most in dollars, but it is also best placed to respond. FanDuel passed most of Illinois’s per-wager tax on through a 50-cent fee. In the UK, Flutter says it cut roles rather than marketing, and took share. Brazil’s provisional ban in September brought an $873m impairment but cost only about $20m of this year’s EBITDA, before cost actions that offset it. So the question isn’t whether rates rise. It’s whether scale gains outrun them.

Over nine months they did not. Reported International revenue rose 11%. Flutter’s own organic measure, which also removes India, rose 1.6%, on the firm’s sum of its reconciliations, and about 4.4% without this year’s sports results, against Flutter’s target of 5–10%. For 2027 management expects International EBITDA to grow at the high end of 5–10%, with revenue growth slightly lower. The UK bill alone rises from $235m net this year to $339m, as the betting duty arrives in April.

The firm assumes International EBITDA grows 7%, and that is less demanding than it looks. FY2027 gets about $169m back from normal sports results and about $82m from growth, against about $104m more UK tax, so about $5m of kept savings reaches it, on the firm’s arithmetic.

Debt first, buybacks later

The CFO’s order of priorities is clear. The timing isn’t.

“Our immediate capital allocation priorities are to deleverage the balance sheet, while still investing in those organic growth opportunities,” the release says, and the buyback is paused “for the foreseeable future”. On the call Rob Coldrake, the CFO, put the order plainly: invest, delever, and only then consider returning more capital. Pressed on whether buybacks could restart before the 2.0–2.5x target, he accepted there’s a compelling case at this price and gave no level.

Three months earlier, at an investor conference, he intended to end 2026 with leverage starting with a three, and said buybacks weren’t too far out. Leverage is now 4.6x, and the guide implies about 4.2–4.3x at year end, on the firm’s arithmetic. That is a miss of at least 0.2x inside three months.

The firm thinks he has the order right, even at this price. Since 2024 Flutter has spent $1,372m on buybacks at an average of about $202, about 0.55x of this year’s EBITDA. A $250m tranche today would buy about 1.8% of the company, and repaying debt with it would save about $12m a year after tax, a 4.7% return. If the firm’s split case is right, the buyback earns far more than that. The case for debt is risk: a company at 4.6x, with the largest question still in court, shouldn’t bet its balance sheet on the answer.

Refinancing moved the revolving facility and term loan A to 2032, the next large maturities are about $1.1bn in 2029 and $3.7bn in 2030, and there’s about $1.7bn of available cash. So the firm reads this as a constraint on choices and not a solvency question. The financial covenant changed on 3 November, from net total leverage of 5.20x to net first-lien leverage of 5.70x. All of Flutter’s debt is secured loans or notes, so on the firm’s reading that’s about half a turn more room. Even with everything going wrong together, as below, the firm’s model reaches 5.12x of total leverage in FY2029. The firm found no explanation of why Flutter moved it now.

Is there a way out?

Yes, by about FY2029 if the courts stay split or the states win. If Kalshi wins, it takes longer.

Flutter would need about $4.3bn of EBITDA to reach 2.5x on today’s net debt, and nobody should expect that. In the split case it gets there by doing two things at once: EBITDA of $3,155m and about $2.9bn less net debt. Here is where that EBITDA comes from, on the firm’s assumptions.

The 35 cents is the firm’s assumption, before the rise in marketing and overheads, and net of it the US earns about 14 cents. FanDuel earned 35.5 cents on added revenue in FY2025 and 27 over the two years to then, and the ceiling is about 42, because cost of sales takes about 58% of revenue. It is what makes lost handle expensive.

Now the three tests.

The split case passes the third test by under $40m of FY2029 EBITDA, once the lost cash is counted, so with little to spare. In all three outcomes the first test passes only because International’s sports results return to normal: against $2,742m without those results, FY2027 EBITDA falls in all three. If Kalshi wins, the other two miss, the first by 0.02x, and leverage reaches about 2.4x a year after FY2029, with that year’s EBITDA and cash flow held, on the firm’s arithmetic.

Other things move the answer. At 9.2x a Kalshi win costs $12 a share and US costs at their bad setting $14: on the downside, the ruling is larger only with the change in multiple it may bring. Any one of the routes below, at its bad setting, takes the split case to 2.5x or above in FY2029, so one is enough to miss or tie the third test.

The danger is that they fail together. A Kalshi win makes promotions and marketing harder at the same moment. On the Kalshi path, with the other cost, tax and revenue routes at their bad settings and a larger loss on FanDuel’s own app, the firm’s model gives $17 a share, FY2029 EBITDA of $1,905m and leverage of 5.12x, still rising. The firm doesn’t treat that as an outcome. It is what no way out looks like.

The shortcuts are closed for now. Selling $400m of EBITDA at 8x would take leverage from 4.6x to about 3.9x, and at 6x to about 4.3x, on the firm’s arithmetic and before tax. The split case reaches 3.9x by the end of 2027 anyway. Management prefers scale, and that’s its view rather than a closed door.

The dials

Ten readings the firm marks each quarter, set against the split case.

Group EBITDA is the US plus International less corporate cost. Under those sit ten things the firm will track from note to note. Flutter controls some directly: marketing, the prediction-market loss and the cost programme. Whether the business grows depends more on those it doesn’t control: what customers will pay and what governments will take.

What the price leaves room for

The firm values each outcome on a multiple of profit, because the way out runs through profit.

The method

The firm doesn’t use peer multiples. Each outcome is a DCF at one 8.4% WACC, running from net debt at 30 September and ending with a value at the end of FY2029 set as a multiple of EBITDA. That exit is about 85–87% of enterprise value, and one turn, one times EBITDA, is worth about $14 a share. Net debt and other claims of $11.1bn rank ahead of about $14.2bn of equity, so small changes in the value of the business are large changes in the shares.

The fall from 17.0x to 10.1x is the share price: enterprise value fell about 47% while EBITDA fell 12%. The firm’s exit starts from 9.2x, the last row, because FY2029 assumes normal sports results, so that is today’s multiple on the same footing, though its build returns only International’s $169m. It assumes the same 9.2x if the courts stay split, one turn less if Kalshi wins, and one turn more if the states win. These are judgements. It doesn’t assume a return to the old multiples, or any buyback or dividend after FY2029.

Because 9.2x comes from today’s price, the split value moves nearly one for one with the shares: about $67 at a $60 price, about $102 at $100. As a check, the three exits imply perpetual growth after FY2029 of 1.6%, 2.1% and 2.4%, on the firm’s arithmetic. That moves with the price too: for the split case, about 1.1% at $60 and 2.8% at $100.

This replaces the firm’s August range of $86–$134 and its exit on revenue. From August’s central $114, the lower multiple, new settings and lower handle take about $35 off, and dating and the balance sheet add about $7. Two costs are worth stating. A WACC one point higher takes about $3–5 off each value, with the exit held. Deducting share-based pay, about $235m a year, which adjusted EBITDA excludes and these cash flows don’t deduct, would take about $4.

The values

These are the firm’s implied values and the share prices they imply; this is not a price target. They hold US marketing and overheads the same across the outcomes. If those moved with the ruling, the range would be about $48–$132.

Three prices for FanDuel

Fox, the US broadcaster, holds an option to buy 18.6% of FanDuel, granted in connection with Flutter’s purchase of The Stars Group. It’s a formula set years ago, and it’s now hard to square with anything else on this page.

Divide the strike by 18.6% and you get what Fox would be paying for all of FanDuel:

So Fox’s price for FanDuel alone is more than the market’s price for the whole group. Flutter’s own model values the option at only $100m, down from $560m in December. The firm thinks it highly unlikely that Fox ever exercises, and counts nothing for it. The option matters here as a measure of how far FanDuel’s value has fallen.

What the price is consistent with

At $81.79, on the split case’s cash flows, the price is consistent with an exit of about 8.9x. Nothing here says the shares are obviously cheap. The firm’s case is the shape of the outcomes, not the gap to any one value.

Bear case, then bull case

Where the position stands

What remains is a FanDuel that may be holding its customers at a lower take than it once had, a large International business that has to outrun governments, and a balance sheet that takes until about FY2029 to mend. Most of that can be worked on by management. The largest question can’t.

What the firm is watching