How game studios can sell data to AI companies

A game studio's archive is unusually complete. Source code sits next to the bug that broke it and the fix that closed it, design documents explain why a system works the way it does, and the game itself is a runnable environment.

That makes studio data a fit for buyers building coding agents and RL environments. The rights are scattered, though: engines, asset stores, platforms, publishers, contractors, performers and players each hold a piece.

What game studio data do AI companies buy?

AI companies buy game studio data that shows software and content being built and fixed. Code with commit history and linked QA tickets is the closest match to what coding agent buyers describe.

Our guides to selling source code and Jira and issue tracker history cover the two most valuable groups in detail.

Who buys data from game studios?

Several buyers' published scopes fit studio data, though none publishes a game studio price. DataVendor lists codebases, tasksets with their RL environments and raw data, lets vendors set their own prices and keeps 20% of vendor earnings when they are withdrawn.

Protege licenses video and spatial and physical data alongside other domains. SimpleClosure AssetHub sells or licenses startup source code and workspace data to AI labs and RL environment builders, for operating companies and those winding down. Our guide to selling data from a closed company covers studios that are shutting.

Can you license code and art built on asset store content?

You can usually license your own code and art, but bought assets have to come out first. Unity's support guidance says all content sold or distributed on the Unity Asset Store cannot be used to train AI or ML models, and points to section 2 of the Asset Store EULA.

Epic's Fab marketplace lets sellers add a NoAI tag, which its documentation says indicates that an asset must not be used for generative AI data collection. Engine source code is a separate question under your engine license, so treat it as third-party material. Our vetting standard requires third-party content to be identified and either cleared or removed.

Who owns art and code made by contractors?

Contractors own the art and code they make unless a contract transfers it. Under 17 U.S.C. section 101, a commissioned work is a work made for hire only if it falls in a listed category, such as part of a motion picture or other audiovisual work, and the parties expressly agree in a written instrument signed by them. Employees' work within the scope of employment belongs to the studio under section 201(b).

Publisher deals add another layer. A development agreement can give the publisher ownership of the game IP or approval rights over its use, so read yours before including any title a publisher funded.

You should not license identifiable player data. The CCPA, at California Civil Code section 1798.140(ad), treats making personal information available for valuable consideration as a sale, and section 1798.140(ae) lists the contents of a consumer's mail, email and text messages, unless the business is the intended recipient, as sensitive personal information.

Children's data carries the heaviest risk. In December 2022, Epic Games agreed to a $275 million civil penalty for COPPA violations, which the Justice Department called the largest ever imposed for a COPPA violation. In the amended COPPA Rule, with a compliance date of April 22, 2026, the FTC says disclosures of children's personal information to train or otherwise develop artificial intelligence technologies are not integral to a service and would require separate verifiable parental consent.

Mobile titles face platform rules too. Apple's App Review Guideline 5.1.2(i) requires developers to clearly disclose where personal data will be shared with third parties, including with third-party AI, and to obtain explicit permission first. Guideline 5.1.2(ii) says data collected for one purpose may not be repurposed without further consent unless otherwise explicitly permitted by law.

Do voice actors and motion capture performers need to agree?

Voice actors and motion capture performers hold rights that a data license must respect. The SAG-AFTRA 2025 Interactive Media Agreement, which members ratified on July 9, 2025 and which runs to October 31, 2028, requires a performer's consent to the creation and use of their digital replica to be in writing and subject to a reasonably specific description of intended use.

The union's published summary does not address licensing recordings as training data, so read each performer contract and leave voice and capture sessions out unless the rights are clear in writing. Illinois' Biometric Information Privacy Act also lists a voiceprint as a biometric identifier.

If you hold UK or EU data

UK studios whose games are likely to be accessed by children must follow the ICO's Age appropriate design code, a set of 15 standards whose foreword says children's data should not usually be shared. UK GDPR Article 8A also tests whether reusing player data for a new purpose is compatible with the purpose it was collected for.

How do you sell game studio data through Sell My Data To AI?

You start with a consultation about your titles, engines, publisher deals and contractor agreements. We do not take custody of your raw data. Buyers see redacted samples under a mutual NDA, and full data moves only after you sign a license directly with the buyer.

Every buyer is verified for company identity, intended use, security and retention commitments and onward transfer limits before seeing a sample. You can withdraw at any point before signing. Book a consultation to start.