DraftKings Writes a $650 Million Check to a Market With a Court Date

Revenue grew 16%, but the guide swung to an EBITDA loss as prediction-market spend tripled — against a base case that assumes a Supreme Court ruling in mid-2028.

DKNG · Earnings Call · 2026-11-06

The guide that went backwards

DraftKings' third quarter looked, on the top line, like a clean beat-and-raise. Revenue rose 16% year over year to $1.33 billion, monthly unique payers climbed 13% to 4.1 million, and core consumer volume grew 22%. Sportsbook net revenue grew 36% on 12% handle growth with roughly 120 basis points of margin expansion to 6.4%, and the CEO claimed the company is “basically tied with FanDuel for number 1 in the industry on sports betting this quarter” — Jason D. Robins, CEO · 2026-11-06. Then the CFO, Alan Ellingson, cut full-year revenue guidance to $6.35–6.65 billion and swung adjusted EBITDA to a loss of $175–275 million, from the $700–900 million guided 90 days earlier. Third-quarter adjusted EBITDA was negative $317 million. The culprit is one line item: adjusted EBITDA investment in predictions, now $650–700 million for 2026. What makes this a story rather than a miss is the size of the step-change. On the May call, Jason Robins framed the budget as “we are thinking we're going to probably invest about $200 million to $300 million all in on Predictions this year” — Jason Robins, Chief Executive Officer · 2026-05-08. Ninety days ago, when Shaun Kelley asked whether they would go above target, Robins answered: “if we see something like that line up this fall, then yes, I think it would be the wise move” — Jason Robins, CEO · 2026-08-07. That is exactly what happened. The spend roughly tripled in a single quarter — and the market is treating it as a cost, not a call option.

What $650 million bought

The bull case is unusually concrete for a nascent business. 1.6 million customers have engaged with predictions year to date. The 2026 cohort is already run-rating to roughly $350 million of annualized gross profit and is expected to pay back in about 20 months — inside the two-year target, and faster than the core book. October annualized consumer volume reached $10.1 billion, nearly 3x July, while annualized total volume traded hit $56.8 billion, up more than 5x. Take rates are running near 11.1%, of which DraftKings captures about two thirds through its own brokerage, exchange and market maker. Combos are already 27% of consumer volume. In a few months it became the number three exchange by volume. The genuinely new wrinkle is the word that now sits atop the company's own keyword list: non sports. Nearly 20% of prediction customers already trade non-sports markets, and per-customer non-sports volume is on par with sports. Robins, almost casually: “we put basically no effort into it, and it is already approaching 10% of our sports volume” — Jason D. Robins, CEO · 2026-11-06. He floated non-sports eventually being 25–30% the size of sports today. That is not a diversification footnote; it is the entire answer to the legal question below. The market-level non sport theme also shows up in the global keyword set this quarter, so this is not purely a DraftKings construction.

The binary management volunteered

Two keywords are new to the top of DraftKings' vocabulary and neither came from the sell side: Supreme Court and sports contracts. For three quarters the regulatory conversation was about the CFTC affirming jurisdiction — a tailwind. Now the framing is litigation, with a base case of a mid-2028 ruling.

Base case forecast is a midyear 2028 Supreme Court ruling, so we should have those customers fully paid back by then... But even if there is a ruling that shuts down sports predictions in July 2028, still think that cohort should pay back in 4 years or less.
Jason D. Robins, CEO · 2026-11-06

Robin Farley of UBS put the sharpest version of the bear case to him: if sports predictions survive, states plausibly legalize sportsbook anyway and DraftKings captures that demand regardless; if they don't, non-sports alone looks like a six-year payback nobody would underwrite.

Why the need to be in prediction market business if... in those 2 scenarios, if it stays or if it goes, it seems like the non sports is not a payback period you would wanna be in.
Robin Farley, Analyst · 2026-11-06

Robins' answer — that you cannot count on state legalization, and that the payback period math clears without any credit for it — is the single most falsifiable claim on the call.

What management stopped saying

The silence is informative. The Super App — the organizing idea of the 2026Q2 call, top-five keyword then — now registers among the largest decliners in the company's own momentum history. So does "Promo". In their place: Prediction customer, cohort economics, and a phrase that has never appeared before, harvest period, used to describe 2028 in the scenario where sports contracts survive. The iGaming reversal deserves its own paragraph. In August the CEO conceded “we have not performed the way that we had wanted to over the last several quarters” — Jason Robins, CEO · 2026-08-07. Now acquisition is up 76% year over year, gross revenue +7% accelerating to 12% in October, and share gains for three consecutive quarters — but net revenue grew just 2% because the new customers arrived with promotions attached, and roughly 60% of the ~$100–150 million implied 2026 EBITDA reduction traces to sports outcomes rather than operations. Cannibalization, the obsession of the August call (“there is very minimal, if any, cannibalization happening” — Jason Robins, CEO · 2026-08-07), is now declared settled: handle up 12% with share gains for a fourth straight quarter.

The tape versus the story

The stock is down 28.5% over 90 trading days and 31.9% from its July 8 peak of $27.17; it sits 74% below the 2021 high. At a $9.2 billion market cap, the multiple has compressed hard: price to revenue is 3.6x, down roughly half year over year from a peak of 11.6x at the end of 2021. Note that the fundamentals feed lags the call: the latest filing covers the August quarter, when revenue was $1.4 billion, down 5% year over year, with operating income of negative $68 million. Meanwhile liabilities at 86.7% of assets, up 27 points over four years — a balance sheet carrying more weight into a period of peak discretionary spend. That is the tension in one frame: a management team asking the market to underwrite a 20-month payback and a 2027 step-down to $300 million of prediction investment, while the market marks the equity as if the $650 million were pure expense. Robins' own words betray the frustration: “I know everyone's a little surprised at how much we spent, and I wish you all could see every data point that we see” — Jason D. Robins, CEO · 2026-11-06. He also pre-committed that “I will be surprised if we spend more in terms of total investment next year on predictions” — Jason D. Robins, CEO · 2026-11-06 — barring a positive regulatory shock. So the whole thesis collapses to a date. If the 2026 cohort pays back on schedule, the 2027 guidance of $800 million to $1 billion of EBITDA with core margins near 18% and a $300 million investment step-down becomes the base case, and the de-rating was an entry point. If payback slips, the company has spent two years of core cash flow on a product whose existence is being litigated. Rarely does a single earnings call hand investors a stopwatch this explicit.