Where the money is in making games in 2026

The other half of the question. Prize money is next door; this is about getting paid to build.

Start with the number that reframes everything.

The median

The median game on Steam grossed 249 dollars in 2025.

After Valve's 30 percent, the developer keeps about 174 dollars. For 2026 releases the picture is, if anything, worse: half of all games released this year have earned under 290 dollars.

Now put a cost next to it. Median development cost for an indie title runs 30,000 to 60,000 dollars.

The median game recovers between 0.3 and 0.6 percent of what it cost to build.

That is a ratio of roughly one to a hundred and twenty at best, one to two hundred and forty at worst. Not a thin margin. Not a difficult market. A different order of magnitude.

Yes, but the average is fine

This is where most coverage goes wrong, so let us be careful.

Roughly 12 percent of Steam indies clear a million dollars in lifetime revenue, and the top 0.5 percent average 7.2 million. Those are real numbers and they describe real careers.

They also sit alongside this: of the roughly 20,000 games shipped on Steam in 2025, about 300 passed a million dollars. That is 1.5 percent of one year's cohort.

Both figures can be true, because they count different things: everything ever released and still selling, versus one year's arrivals measured early. The honest reading is that the lifetime figure includes the survivors of a decade in which far fewer games competed, and a game released in 2026 is not entering that market.

If you want one sentence: the distribution is not skewed, it is vertical, and it has been getting more vertical every year that supply doubled.

So where does the money actually come from

Not from selling copies, for most people. Here is the honest list.

Somebody else's budget. In 2025, 56 percent of developers self-funded. Publishing deals or project-based funding covered about 28 percent, crowdfunding 11 percent. Read that the other way round: for nearly a third of developers, the money arrives before the game does, and the game is the deliverable rather than the product.

Work for hire. Contract development, ports, tools, contributing to somebody else's title. Unglamorous, invisible in the sales charts, and the actual income floor for a large part of the industry.

Platform creator economies. Roblox paid out about 1.5 billion dollars to creators in 2025. It is also the platform where the median creator earns 1,575 dollars a year, so it is not an escape from the distribution problem, it is the same problem at a different address. The full numbers are here.

Web publishers. Poki, CrazyGames and the HTML5 sponsorship model, which pays for traffic rather than for copies. First deals run in the hundreds to low thousands per month. Covered in detail here.

The 0.5 percent. Building a game that sells is a real path, it just should not be the plan you present to a bank.

What actually changed

The temptation is to say the market got worse. That is not quite it.

Production cost fell and revenue per title fell faster. A game that cost 200,000 dollars in 2018 might cost 40,000 to build now. That looks like an improvement until you notice that median revenue fell from thousands to hundreds over the same stretch, because supply grew eleven-fold while the audience grew six-fold.

Cheaper to make, and much harder to be seen. The bottleneck moved from the workshop to the shelf, and nothing about that is fixed by making the workshop cheaper still.

Which means the interesting work is on the shelf. Discovery is the constraint. It is expensive enough that Roblox told the stock market its bookings would fall because it changed how players find things. When a discovery change moves a public company's guidance, it is not a small problem.

The honest advice

If you want a salary, the money is in work for hire and platform contracts. Boring, real, and it pays monthly.

If you want to own something, build for a platform economy or build cheap enough that the median outcome is survivable. A 5,000 dollar game earning 4,000 is a different life from a 60,000 dollar game earning 174.

And do not plan around the mean. The mean revenue per game on Steam is a real division that almost nobody experiences. Every business plan built on it has been wrong in the same direction for a decade.

Sources

• Median gross revenue for a game on Steam in 2025 of 249 dollars, roughly 174 to the developer after Valve's 30 percent; 50 percent of 2026 releases under 290 dollars
• Median indie development cost of 30,000 to 60,000 dollars; most indie developers earning under 20,000 dollars a year
• Funding split 2025 per GDC State of the Game Industry: 56 percent self-funded, about 28 percent publishing or project-based, 11 percent crowdfunding
• 12 percent of Steam indies above one million dollars lifetime, top 0.5 percent averaging 7.2 million; approximately 300 of the roughly 20,000 titles shipped in 2025 passing one million
• Compiled from Steam Page Analyzer, games-stats.com Steam Q1 2026 analysis, Fungies.io indie market analysis and Shane The Gamer's indie statistics review
• Roblox creator payouts and distribution as set out in our own Roblox piece

Retrieved 4 August 2026. The recovery ratio and the coverage percentage are ours, calculated from the median revenue and median cost figures above. Where two sources describe different populations, both are shown rather than reconciled.