Unity Won the Game Count and Lost the Money. That's Not a Duopoly.

On September 12, 2023, Unity told the people who built games on it that they would soon be charged a fee every time one of those games was installed. Not sold — installed. A pirated copy, a re-download, a charity bundle: a meter on the runtime itself. The reaction was not a complaint; it was an exodus in slow motion. Inside eleven days the company apologized and started backing out.⁵ Inside a month, the CEO was gone.³ And inside twelve months, the fee was dead and the high-dollar end of the market had quietly walked across the street.⁴⁶ The whole episode is usually filed as a pricing blunder. It was something larger: the moment a company with the most games on earth proved it no longer had the most valuable ones.

The story everyone tells is that game engines are a tidy duopoly — Unity and Unreal, two giants of roughly equal size splitting the world. Almost every word of that is misleading. They are not equal, and they do not split the same world. One owns the volume; the other is taking the money.

Unity has the games. Unreal has the dollars.

Look at where developers actually ship. In 2024, 51% of all games released on Steam ran on Unity, against 28% on Unreal Engine.⁶ By that count, it isn't close — Unity is the default for indies, mobile studios, and the long tail of small releases, the way a spreadsheet is the default for a back office. But count the money instead of the games and the picture inverts. By revenue, Unreal pulled 31% of Steam game sales in 2024, up from 19% the year before, and passed Unity's 26% for the first time since 2018.⁶ The same Steam, two opposite leaderboards. Unity wins the headcount; Unreal wins the cash register. That gap is the whole strategic story, because the games that generate revenue are the games studios pay real money to make — the high-fidelity, high-budget titles where Unreal's rendering pedigree is the obvious choice.

This is why "duopoly" is the wrong word. A duopoly implies two firms competing for the same prize on the same terms. These two are climbing different mountains. Unity's moat is ubiquity — millions of small studios who already know its tools and don't want to relearn another. Unreal's moat is fidelity — the engine that big-budget studios reach for when the visuals are the product. The danger for Unity isn't that Unreal will steal its indie horde. It's that the horde was never where the profit lived.

How a pricing memo became a structural wound

Unreal's revenue surge did not happen in a vacuum. The crossover landed in the same window as the runtime-fee disaster, and that is not a coincidence. The fee was meant to extract more from successful games — exactly the high-revenue titles Unity most needed to keep. Instead it taught those studios that the cost of staying could change retroactively, on a meter they couldn't predict. Unity walked it back in stages: a partial retreat on September 22, 2023, dropping the charge for free-tier users and grandfathering games on older engine versions⁵, then a full cancellation in September 2024, swapping the install meter for a cleaner seat-based subscription — Unity Pro up 8% to $2,200 a seat, Enterprise up 25%, effective January 1, 2025.⁴ The price went up; the trust did not come back. By the time the fee was buried, Unreal had already booked the revenue crossover.

The leadership change tells you how seriously the board took it. The official line was retirement, but the SEC filing reads differently: the board letter — signed by Roelof Botha as Chairman — accepts the resignation with the phrase "I am writing to confirm my acceptance of your resignation so that you may pursue retirement."⁹ The official framing is retirement; the abruptness of an effective-immediately departure, weeks after a policy firestorm, invites its own reading. You do not lose a chairman and CEO in a single 8-K over a fee you believe in. The fix wasn't to soften the policy; it was to remove the person who shipped it.

Why a 5% royalty can be the cheaper deal

Compare the two pricing philosophies, because they reveal the moats. Unreal charges no per-seat fee at all. A studio keeps the first $1 million of gross revenue per title each year and pays 5% above that — and if the game ships on the Epic Games Store, the Unreal royalty is waived entirely.⁸ That structure aligns Epic's payday with the developer's: Unreal only gets rich when the game does. Unity's seat license does the opposite — it charges you the same whether your game earns nothing or millions, which feels efficient to a profitable studio and punitive to a struggling one. The runtime fee briefly tried to bolt success-based pricing onto a seat-based engine and got the worst of both: the unpredictability of a royalty with none of its "we win when you win" goodwill.

The honest counter: Unity isn't collapsing

The fair objection is that this reads too much doom into one bad year. And it's true that Unity is not falling apart. Its 2024 revenue came in at $1,813M, down from $2,187M the year before — but most of that drop was deliberate housecleaning, not erosion.¹² Strip out divested businesses and one-time items and the "strategic portfolio" Unity actually runs was down only about 1% year over year.¹ The headline number was always softer than it looked; the underlying engine business held. So the bear case isn't that Unity is shrinking — it's that the company is still far larger by dollars than Unreal Engine's standalone licensing line, which third-party estimates put at roughly $275M for 2023, a rounding error inside Epic's Fortnite-driven business.⁷ By that measure Unity, with $614M in Create Solutions revenue in 2024, still dwarfs Unreal as a licensing business.¹

That's real, and it's the steelman. But it cuts the other way too. Epic doesn't need the engine to be a profit center — it has Fortnite-driven revenue dwarfing the engine's licensing line⁷ — which means Epic can waive Unreal royalties for store titles and offer terms Unity structurally cannot match.⁸ Unity has no Fortnite. The engine is the business, which means Unity has to monetize the exact developers Epic can afford to subsidize. A free $200K revenue ceiling and a $2,200 seat are not weapons against a competitor that can give the high end away.⁴ The asymmetry isn't just in market share. It's in who can afford to lose money to win.

So forget the duopoly. Picture two doors into the same building. Through Unity's door pour the millions — the hobbyists, the mobile shops, the prototypes that never ship. Through Unreal's door walk the studios with budgets, and they carry the money. Unity built the wider door and called it dominance. Then it tried to put a turnstile on the people who were actually paying, and watched them quietly find the other entrance. The engine count still favors Unity, and probably will for years. But the revenue crossover in 2024 made the real shape of the market visible: this was never two giants splitting one prize. It's one company that owns the crowd, and another that's taking the till.