Roblox's 'Grow' Is a Promise to Its Creators, Not Its Investors

The Roblox Developers Conference opened this week with the usual promise: new ways to play, build, and grow. Heading into it, the stock had climbed about 20% over the previous month. That looks like the market buying the story again. The numbers underneath it tell a different one.

The useful thing to notice is who "grow" actually belongs to. Roblox the company does not make the games on Roblox the platform. Independent developers do, and they keep roughly 30 cents of every dollar a player spends while Roblox keeps about 70. So when the company talks about growth, the first question is growth for whom—and in 2025, the answer was mostly the creators.

That's not a coincidence. It's the model's one real lever.

Roblox pays its developers a share of what players spend on the virtual currency, Robux, when they cash out. In September 2025 it raised that rate by 8.5%. The effect was on purpose: take less out of each transaction and the economy grows large enough that total money in still rises anyway. That is the Laffer logic applied to a digital marketplace. For a year it looked brilliant. Developers made a record $1.5 billion in 2025, up 63% from $923 million the year before, and Roblox's own revenue rose 36% to $4.9 billion.

Paying your suppliers more and getting bigger yourself is a pleasant trick. It works up until the day it stops working.

In the quarter just reported, bookings—the money players actually commit to buying Robux, the truest gauge of demand—grew only about 8% year over year, at the low end of guidance. And for the current quarter, Roblox guided bookings down 14% to 18%. That is a first. In its public history, the company has never guided to shrinking bookings.

The most interesting part is that the slowdown was partly self-inflicted. Roblox rewired its discovery algorithm away from viral, high-monetizing hits and toward "evergreen" games built for long retention, betting that a user who stays for years will spend more over a lifetime than one who churns. At the same time, a rollout of identity and age verification added friction to the sign-up flow—friction that landed hardest on the under-13 US and Canada audience, historically its best payers. Both are deliberate trades of today's money for tomorrow's. Whether the trade is good is the whole question.

Here is how to test it. Do not watch the daily active user count. Roblox grew that impressively in 2025, averaging 126.5 million and closing the year near 144 million. That much works, and it's the figure the conference is designed to celebrate. The number that matters is bookings per daily active user—how much each person who shows up actually spends. The retention bet only pays off if that figure stabilizes and then climbs as loyal users mature. If DAU keeps rising while spending per user keeps sliding, then "growth" is more people showing up on a smaller economy, and the minus sign in bookings is the model speaking rather than a single bad quarter.

The valuation makes the stakes concrete. Even after a staggeringly bad twelve months—the stock hit a 52-week low just above $34 after peaking near $142—Roblox still carries a market cap around $33 billion, roughly 5.8 times trailing sales, against an operating margin of about negative 24%. Roblox does not make money; it moves a lot of money and keeps a thin, currently negative slice. Investors are still paying growth multiples because the platform's size is real. But a price-to-sales premium assumes sales are going to grow.

So the stage in San Jose is where Roblox tells its developers there are new ways to grow. The version for shareholders is narrower and harder: can the money per user grow again, or does Roblox's growth keep leaking out to exactly the people it's working hardest to attract? That isn't a product question that gets answered at a keynote. It's a bookings-per-user number you can watch over the next two quarters. Until it turns up, the recent bounce is confidence priced in before the evidence arrives.