Updated October 2026
Selling a software agency or IT company
Software agencies and IT consultancies sell on their client roster and their engineers. Their archives hold what AI coding buyers prize most: requirements turned into tickets, tickets turned into code, code reviewed and revised. Turing's Project Lazarus licenses code, tickets, communications and decision records from operating and winding-down companies, by its own account.1
The field
How many agencies clear the floors
The Census counts 10,560 US custom software development firms with 10+ employees and 4,725 with 30+, plus 6,733 IT consulting and systems integration firms with 10+ and 2,871 with 30+. Most buyers want 20 to 30+ employees (Cognor 20+, micro1 30+), and Troveo states no minimum; micro1 names technology first among its industries, Pewter lists software and IT.2,3,4,5,6,7,8
| Fact | Value and source |
|---|---|
| US custom software development firms and agencies (NAICS 541511) | 10,560 with 10+ employees; 4,725 with 30+ |
| US IT consulting and systems integration firms (NAICS 541512) | 6,733 with 10+ employees; 2,871 with 30+ |
| Turing Project Lazarus | Licenses code, tickets, communications and decision records from operating and winding-down companies; up to $1M in one payment |
Whose code
Most of an agency's code is written for clients
An agency writes code under statements of work, and the code usually ships to the client. That is the central question for any data license, and it is listed below for the licensing company. Polyshares says it excludes data held on behalf of clients. What is less tied to any client: internal tools and libraries, estimation spreadsheets, architecture decision records, code-review guidelines, onboarding docs for new engineers and the internal discussion in Slack or Teams.9
SimpleClosure reports Slack in 95% of the workspaces it evaluated and GitHub in 82%, and says roughly 37% of raw code disappeared after removing third-party, generated and duplicate code. Agencies that copy starter templates across projects should expect a large haircut on code volume.10,11
Terms
Process and payment
Turing says it signs an NDA from first contact, samples within a week and pays once, up to $1M, its own ceiling. Mercor buys from GitHub, Confluence, Slack and 100+ other tools. Like most buyers it charges sellers nothing, and payment typically follows 2-8 weeks after signing (Mercor 2-4 weeks, Cognor 3-6, Polyshares Net 30-60). SimpleClosure says companies need not be shutting down to use its marketplace.12,13,14,15,16,17,18,19
If the agency is being acquired, the acquirer will want to know what has been licensed. Foley reports training-data rights becoming a separate diligence category in AI deals.20
Payment and fee terms differ more than prices do. Most buyers, Polyshares among them, charge sellers nothing, while Jottings takes 20-30% of proceeds and FileYield 8%.19,21,22
FAQ
Questions, answered with sources
That depends on your contracts and is a question for the licensing company. Polyshares says it excludes data held on behalf of clients.9
SimpleClosure says roughly 37% of raw code disappeared after removing third-party, generated and duplicate code.11
Up to $1M in one payment, its own ceiling, not a typical amount.1
We do not answer legal questions. Ask the licensing company: their team reviews this with you before anything is signed.
Before you sign
Legal questions go to the licensing company
We do not answer legal questions. Ask the licensing company before you sign: their team reviews these with you.
Questions for this kind of business
- Which repositories and tickets are ours, which are clients', and how will you verify that?
- Do customer, vendor or platform contracts, NDAs or API terms restrict this?
Questions for any buyer
- Is the license exclusive, and for how long?
- Can you sublicense the data, and to whom?
- Which AI labs will receive it, and will you tell us?
- How long do you keep it, and how is it deleted?
- What is your liability cap, and who indemnifies whom?
- Will the contract ban re-identification of people in the data?
- Who removes personal data, by what method, and do we approve it first?