If your company has built its own software for years, you may be holding code that AI labs will pay to train on. Buyers don't price all code alike, and a few signals raise the price more than the rest. Read on to find out which code buyers ask for first, what adds to its value, and how to check your own codebase against the list.
What kind of code AI labs pay for
Buyers tell us they pay most for code with a rich history from a complex product a company built in-house. Four signals lead their lists.
| Signal | What buyers ask for | Why it matters |
|---|---|---|
| The product | A complex product your own team built and ran in production | Code that ran real workloads carries real edge cases and real tradeoffs |
| The language | Back-end languages, with C++, C#, Go and Rust preferred | These are the languages large back-end systems are built in |
| Who wrote it, and when | High-quality code written before large language models by on-shore teams, or code from in-house teams working with frontier models | Buyers ask how the code was written as well as what it does |
| The record around it | Git history, comments, and the Jira tickets that go with the code | It shows how the code changed, why, and which request each change answered |
Our guide to how to sell source code to AI labs covers why labs started paying for private code and how they price it.
The languages buyers prefer
C++, C#, Go and Rust are the four languages buyers ask us for first. All four are back-end languages, the kind large systems are built in.
Buyers call these four preferred, and code in other languages still licenses. The other signals on this page apply to any language.
Complex products beat simple apps
Buyers pay more for a product that does something hard. A fintech ledger or a logistics routing engine is worth more than another CRUD app.
In-house matters too. A product your own team designed, shipped and maintained carries years of decisions (what broke, what got rewritten, what was traded off), and labs want their coding models to learn from that record.
When the code was written, and who wrote it
Buyers give extra credit to two kinds of code:
- High-quality code that on-shore teams wrote before large language models (pre-LLM code).
- Code that in-house teams write today working with frontier models.
If your product was built before AI coding assistants arrived, say so and show it. The commit dates in your Git history are the proof.
Git history, comments and Jira tickets
Buyers pay a bonus for the record around the code, and a snapshot with no history is worth little to them. Three parts of that record count most:
- Your Git history shows how the code evolved, what broke, and how it was fixed. Export every branch and every commit.
- Your comments, in the code and in code review, capture the engineering judgment behind a change (what was accepted, what was rejected, and why).
- Your Jira tickets connect a request to the change that resolved it and to the outcome. Tickets in Linear or another tracker do the same job.
A repository with its ticket history and the engineering chat that references both is a workflow trajectory in raw form, and it's worth considerably more as a connected package than as separate exports.
Check your own codebase
Six questions tell you where your code stands:
- Did your own team build the product, and did it run in production?
- Is the back end written in C++, C#, Go or Rust?
- Do you have years of Git history, with every branch still there?
- Was the code written before AI coding assistants, or by an in-house team working with frontier models?
- Do your commits link to tickets in Jira or another tracker?
- Does your company own all of it, with contractor work assigned and open-source dependencies listed?
The more of these you can answer yes to, the stronger your position. You don't need all six.
What it's worth
Typical full-company licensing deals start at six figures, and code is one part of what a company licenses. For code on its own, the public figures come from companies that shut down, and they're the small end of the market (one codebase-buying platform lists a 5,000 dollar baseline per repository, and closure deals have run from about 10,000 to 100,000 dollars).
An operating company with the signals above is in a different conversation. Buyers price a whole engineering organization's history, connected across systems and spanning years, as an operational archive. Our guide to how much your company's data is worth to AI labs covers the public price points and the six factors behind them.
Rights to clear first
You can only license code you own. Before a buyer looks at anything, confirm three things:
- Every contributor, contractors and agencies included, assigned their work to your company.
- Your open-source dependencies are listed, with copyleft and redistribution-restricted code identified.
- Secrets, credentials and customer data are out of the history, and personal information in tickets and commit messages is replaced with pseudonyms.
If your code was written for clients, our guide to what agencies and consultancies can license to AI labs covers what you own and what your clients own. Our guide to how long it takes to license your company's data covers what buyers check and how the timeline runs.
You license it and keep it
Licensing your code doesn't end your use of it. You keep ownership, keep building, and keep selling your product. You can license selectively (a retired product line, a legacy system that's been replaced, an internal tool with years of tickets behind it), and a non-exclusive license leaves you free to license again. Our guide to exclusive vs non-exclusive data licenses covers the difference.
If your company is closing, our guide to selling your code and data to AI labs when shutting down covers that route.
Where to start
The quickest way to find out what your engineering history could license for is to take our free data value assessment (ten questions about your systems, history and industry, about five minutes). You get an estimate, and a scoping call after that gets you a real number.
Where Troveo fits
Troveo is a data licensing marketplace. We work with more than 40 active buyers and we've paid rights holders more than 20 million dollars. Our business data program covers code alongside documents, messages and operating systems. You decide what's in and what's out, we handle the packaging and every buyer conversation, your company is never named publicly, and you're paid on every sale. Start with the data value assessment, or talk to us about what your engineering history might be worth.