US software companies with offshore engineering teams: who owns the code and records?

Under US copyright rules, code employees write in their jobs belongs to the employer, but code from independent contractors, including offshore agencies and freelancers, generally becomes the company's only through a signed assignment. For a data license, the IP terms in each development contract decide which repositories, pull requests and tickets can be included.

The contract decides, not the location

Where code was written matters far less than the paper behind it. Under US copyright law, work an employee creates within the scope of the job is a work made for hire, so the employer is treated as the author and owner, as the Copyright Office's Circular 30 on works made for hire explains. Work from an independent contractor is different: it counts as made for hire only if it falls into one of the nine categories in the definition at 17 U.S.C. 101 and both parties signed a written agreement saying so. Software is not named among those categories, so contractor code usually becomes the company's through a signed written assignment rather than by default.

For a US software company with engineers in Kraków, Bangalore, Manila or Medellín, each engagement model therefore needs its own answer. Local law in the developer's country can add requirements of its own, which is why rights are checked contract by contract rather than assumed.

This is general information, not legal, tax or financial advice. Confirm with your own counsel before acting.

Engagement models and where rights usually sit

Engagement modelDefault position under US rulesWhat the paperwork should show
US employeesThe company owns work made in the jobEmployment agreement with confidentiality and invention terms
Employees of the company's own offshore subsidiaryThat subsidiary, as employer, absent other termsAn intercompany development or IP assignment agreement in favor of the US company
Employer-of-record staff abroadDepends on the chain: the provider employs, the company directsAn assignment running from developer to provider to the company
Offshore development agencyThe agency, whose own employees wrote the code, unless it assignsThe IP clause in the master services agreement and in each statement of work
Freelancers hired directly or through a marketplaceThe freelancer, unless assignedA signed contract with a present assignment, plus the marketplace's terms
Open-source and third-party codeThird parties, under their licensesExcluded from scope, with license notices kept

When the offshore team belongs to a sister company inside a larger group, the question widens into intercompany ownership of shared group systems.

What records these companies hold

SystemRecordsWhy AI buyers value them
Git hosting (GitHub, GitLab, Bitbucket)Commits, branches, pull requests, review commentsShows how changes are proposed, critiqued and corrected, step by step
Issue tracker (Jira, Linear, Azure DevOps)Tickets, sprints, bug reports, acceptance criteriaLinks work to requirements and records whether it shipped
CI/CD pipelinesBuild and test runs, deployment logsPass and fail outcomes tied to specific changes
Incident and on-call toolsIncident timelines, postmortemsDiagnosis under pressure, with root cause and fix
Chat (Slack or Teams)Handoff threads, design debatesReasoning and decisions that never reach the tracker
Wiki and design docs (Confluence, Notion)Architecture decision records, runbooksDecisions written down with context and alternatives

AI developers are building agents that carry out multi-step engineering work, and training or evaluating them needs records of how real teams plan, review, test and fix. Those records sit mostly inside companies and are thin on the public web. Epoch AI researchers project that, if current trends continue, language models will fully use the stock of public human-generated text sometime between 2026 and 2032, a forecast with wide uncertainty that raises the value of permissioned, non-public records.

Distributed teams add something specific. Engineers split across time zones cannot lean over a desk, so handoffs, review comments and tickets tend to be written out in full. The data type brief on software engineering workflows covers which of those records matter most.

Which software companies fit

  • Product companies that own the code they ship: B2B SaaS, vertical software, platforms with long release histories.
  • IT services firms and MSPs whose internal tooling, runbooks and ticket histories are their own.
  • Headcount at the US company: 50+ full-time employees at peak, with contractors excluded. Offshore agency developers and freelancers do not count toward that number, however large the team. Whether staff employed by a foreign affiliate count is settled during qualification rather than assumed up front.
  • History: a multi-year operating record, with repository and tracker history that survived migrations, for example from older version control to Git.

Custom development agencies that build software for clients are a weaker fit. Their client work usually belongs to the clients, and data that belongs to someone else, without consent, is a red flag. Their own internal records may still qualify. If a group is deciding whether to sell or close a US software subsidiary, the comparison of selling, winding down or licensing records shows where licensing fits.

Rights and confidentiality pitfalls

  • Assignment on payment. Some agency contracts transfer IP only once invoices are paid in full, so disputed or unpaid invoices can leave gaps.
  • The first contract. Early offshore engagements are sometimes signed in a hurry without an IP clause; later renewals may be cleaner than the years they replaced.
  • Client-owned work. Customizations built under customer contracts may belong to those customers.
  • Vendor accelerators. Frameworks and libraries the agency reuses across clients usually stay with the agency.
  • Secrets in history. API keys, passwords and customer data can sit in old commits, logs and tickets. How those are redacted or de-identified is settled with the company before any work starts.
  • Personal data. Developer names and email addresses in commit metadata, and customer details in tickets, need handling rules too.

The paper-trail screen

Partners do not read contracts. Ask the CTO or CFO whether they know the answers:

  • Does every development contract, for employees, agencies and freelancers, assign IP to the company in writing?
  • Are any years or teams covered only by statements of work with no IP clause?
  • Is the code the company's own product, rather than work owned by clients?
  • Can third-party and open-source code be identified and left out?
  • Can someone export repositories and tickets with full history intact?

Uncertain answers do not end the conversation. They tell the rights review where to look first, and the scope can shrink to the years and repositories with clean paper.

Who can introduce these companies

RoleWhere they see the signal
Fractional CTOs and engineering advisersThey read vendor contracts and run the repositories
Software M&A advisersIP diligence exposes assignment gaps and the depth of history
Advisers who set up offshore entitiesThey draft the intercompany agreements and see the captive center's scale
Offshore delivery leadersThey know the client's executives and the engagement history, and must be open about their own firm's role and any IP it keeps
PE operating partnersPortfolio companies with offshore engineering hubs

Fractional CTOs can adapt the introduction email templates for fractional CTOs. For a company with an overseas parent, check the points on referring a foreign-owned US company.

Partners earn 25% of the eligible platform fees SourceX actually collects from the referred company's licensing deals, capped at $100,000 per referred company, payable only after the buyer pays and SourceX receives its fee. The reward is funded from SourceX's share, so the company's payment is untouched, and no reward is guaranteed.

Conversation starter

Next step

Run a candidate through the paper-trail screen, then the company fit checker. If it holds up, register as a partner and send the CTO your referral link.

  1. Step 1Share your linkSend your personal link to a company you know.
  2. Step 2Company appliesThe company applies itself at /apply.
  3. Step 3Buyer selects and paysThe buyer selects and pays for the data and SourceX receives its fee.
  4. Step 4You get your rewardYour share of SourceX fees becomes payable.

Common questions

Do offshore contractors count toward the 50+ full-time employee baseline?

No. The baseline is 50+ full-time employees at peak with contractors excluded, so agency developers and freelancers abroad do not count, however many there are. A company with a small US staff and a large outsourced team may not qualify on headcount. Whether employees of the company's own foreign affiliate count is settled during qualification.

If an agency wrote most of the code, can the company still license it?

Possibly, if the agency assigned its rights to the company in writing, typically through the master services agreement or each statement of work. Pull request discussions and tickets written by agency staff raise the same question. If assignment is missing for some years or teams, the scope can be narrowed, or the agency's consent sought by the company.

What about code written by the company's own subsidiary abroad?

Under US rules, the subsidiary that employs the engineers is the starting point for ownership, so the question becomes whether an intercompany agreement assigns the work to the US company. Some groups have development or cost-sharing agreements that do this; others never put one in place. That has to be settled in rights review before anything is offered to buyers.

Is the source code itself licensed, or only the tickets and reviews?

The company decides. Scope is agreed with each company and can include repositories, pull request discussions, tickets, design documents or only some of these. A company may hold back sensitive modules, and third-party code is left out. Redaction and de-identification rules are settled up front, and delivery happens only under an executed agreement.

Can a custom software agency that builds for clients qualify?

Its client deliverables usually belong to its clients, so they are out of scope without client consent. The agency's own records can still qualify, such as internal tooling, project management history, support tickets, proposals and delivery playbooks, provided it meets the size, history and rights baseline and has an authorized sponsor.

Free resources

By SourceX Partnerships Team · Published 2026-10-09 · Updated 2026-10-09

Know a US company with valuable proprietary data?

Become a referral partner from anywhere we support, get your link and introduce an owner or authorized decision-maker.

Refer a company →

I own a business

Explore licensing your company's data to AI developers worldwide. Start a short assessment; no uploads needed.

Start an assessment