How casinos actually make money

Start with why anyone wants this business. Casino EBITDA margins run around 41% at the top against 25% for luxury goods, 21% for cruise lines, 19% for hotels and 11% for restaurants. Gambling out-earns almost every consumer industry there is, and it does it with a mathematical guarantee rather than a brand.

That guarantee is the house edge, and it is enormous at the casual end. Blackjack gives the house 2%, craps 2%, baccarat 5%, roulette 5%, slots 15%, and keno 40%. A player can win a hand; a player cannot win a thousand hands. Everything else in the building exists to increase the number of hands.

Which is why the free food is not generosity. The Wynn hands out $200–400 million a year in complimentary rooms, meals and drinks to its gamblers — a number that looks enormous until you set it against gambling revenue, where it works out to roughly 9%. As long as the lobster keeps someone at the table, the maths holds.

The whole resort is a funnel. Casino revenue has been the majority of the Wynn's gross earnings for 17 consecutive years, and on a dollar basis it grosses two to three times as much from gambling as from every non-gambling business combined. The hotel is only in demand because of the casino, and its capacity is fixed.

The reason this is not a crowded market is that states ration it. Commercial casinos are still banned in over half the U.S., sports betting is illegal in 21 states, and real-money online casinos are illegal in more than 35. Supply is throttled deliberately, which is what lets states tax it so hard.

And they tax it hard. Rhode Island takes up to 74% of gross slot revenue and Delaware 57%, against 9% in South Dakota and under 7% in Nevada — on gross earnings, not profit. For every $100 lost on a Delaware slot machine, the state keeps $57.

There is one structural loophole, and it is the largest single fact about American gambling. Indian reservations are self-governing and outside state jurisdiction, so a state that has banned casinos cannot stop one opening on tribal land inside its borders. There are now more tribal casinos than commercial ones — 523 against 468.

Against that backdrop, FanDuel and DraftKings looked like a clean arbitrage. No dealers, no janitors, no cashiers, no free prime rib; infinite virtual tables against a hard cap on physical ones; and one storefront serving every state where the game is legal. On paper the margins should be better than the Wynn's.

They raised accordingly — $11 million, $70 million, $275 million and $80 million at FanDuel against $34 million, $41 million, $300 million, $224 million and $119 million at DraftKings — and reached billion-dollar valuations in five years on an activity that was legal by omission rather than by permission.

The fantasy business underneath was never viable. In 2017 DraftKings grossed $191 million and lost $73 million; FanDuel did $124 million and lost $44 million. The pots were too small, most of the entry fee goes back to the winners, and the average user generated $28 a year. The 2015 loss alone was $509 million.

The 2018 Supreme Court repeal was the lifeline, and the pivot did work on the top line. Revenue went $192 million, $226 million, $323 million, $615 million, $1,296 million and $2,240 million — but sales and marketing went $156 million, $145 million, $185 million, $495 million, $982 million and $1,168 million alongside it. Between 50% and 80% of every dollar earned goes straight back out to buy the next one.

But volume was never the real constraint — licensing was. Sports betting remains illegal in 21 states, 42% of the country, six years after the repeal, and some states permit it in person but not online. A platform business that cannot reach half its market cannot reach critical mass, so the only lever left is squeezing the states it already has.

And those states charge a premium precisely because an app is worth less to them. New York taxes online sportsbooks 50% of gross revenue against 10% for retail; Arizona, Massachusetts and New Jersey all tax online above retail. A casino brings jobs, tourism, construction and outside money; DraftKings brings a download.

In several states it cannot even hold its own license. To operate in New Jersey, Iowa, Mississippi, Pennsylvania and Colorado, DraftKings has had to partner with a local brick-and-mortar casino, which hands over the license in exchange for a slice of every dollar taken statewide. The incumbent gets paid for existing.

Stack it up and the business is illegal in half the country, taxed at up to 50% where it is legal, revenue-shared with a casino in several states, and priced against rivals it cannot undercut — while paying league and data licensing fees on top. There is very little left at the bottom.

Meanwhile the house it was meant to disrupt keeps clearing double digits. Core casino operating margins run 43% at the Wynn, 50% at MGM Resorts and 52% at Caesars, and they have stayed in that band through a pandemic. The edge does not care about labor costs or commodity prices.