Games With AI Disclosures Have Grossed an Estimated $660M on Steam
A few months ago, we ran the numbers on how many Steam games disclosed using Generative AI. At the time, we were at 7,818 titles—about 7% of Steam's total library. That number has since climbed to 10,258—now 8% of all Steam games.
That raised a question: are these games actually making money?
Using the Boxleiter Method (Review Count * Price * 35 = Ballpark Gross Sales) for gross revenue, we can estimate that the games voluntarily disclosing GenAI usage have a combined gross of about $660 million on Steam.
Breakdown by Revenue Bracket
Again, keeping in mind that the Boxleiter Method is the roughest back-of-the-envelope calculation, it suggests that out of the 10,258 titles with AI disclosures:
12 games have grossed 8 figures.
33 games hit 7 figures.
170 are in the 6-figure range.
487 hit 5 figures.
The rest (9,556) came in below that.
What Are These Games Using GenAI For?
Among the 45 $1M+ estimated grossing titles, developers cited the following uses (with overlap across categories):
In-game Art: 49%
Voiceover / Dialogue: 27%
Preproduction / Concept: 13%
Marketing & Promotion: 11%
Translation / Localization: 9%
Upscaling & Technical: <5%
Runtime Generative Systems: <5%
Other Uses: <5%
A listing of each game’s GenAI disclosure is available in this spreadsheet.
Stated Motivations Behind AI Use
The most frequently cited reasons among the $1M+ earners are:
Budget Constraints: Devs cited cost/time reduction; keeping production within tight budgets.
Faster Prototyping: Generating placeholder art, voices, or ideas to test and refine concepts before full production; used to accelerate early design and iteration.
Audience Reach: Described as a way to make localization, translation, and multilingual voiceovers affordable enough to enter markets that might otherwise be out of reach.
More Content: Teams say they’re using AI tools to scale up the volume of auxiliary, decorative, or repetitive assets they need to create bigger games than they would otherwise.
Improved Quality: Some are using AI-based tools for upscaling and similar refinements to raise visual fidelity beyond what they can do in-house.
Some of these categories overlap a bit (for instance, resource constraints underlie most of them), but in short, they’re currently citing “more production with fewer resources” as the big why (as opposed to, e.g., using it at runtime).
Most Disclosures Describe Light Use
Most developers cite GenAI usage for small, peripheral tasks. The typical note is brief (AI helped with a Store Page image, a background painting, a few icons, or temporary voice lines during prototyping). Even among higher-profile titles, disclosures emphasize limited scope (“minor components,” “no gameplay impact,” “under 1% of content,” etc.).
The Bigger Picture
Taken together, this indicates GenAI use is present across Steam, including higher-earning titles. For now, disclosures suggest it’s mostly confined to smaller work: art generation, dialogue polish, localization, and similar. Still, its presence in titles grossing tens of millions suggests it’s becoming part of mainstream production, and is not limited to smaller experimental projects. Future data will clarify how its use evolves!
Addendum
Since we initially pulled the data for this post, Arc Raiders has launched with more than 30,000 player reviews, over 90% positive. It cites limited GenAI tooling use, consistent with current disclosure patterns for other titles with similar review counts.
