Yes—reporting describes data brokers acquiring codebases from failed startups for resale or packaging to AI labs, alongside interest in broader company records such as Slack conversations, tickets, and work histories. But the evidence does not establish that every reported approach became a sale to a lab, nor does it reveal a reliable market-wide total. The key distinction is between a reported acquisition, an offer, ongoing talks, and a completed AI training-data deal.
What is being offered to AI companies?
Source code is only one possible asset. The Information has reported that data firms, including Turing and AfterQuery, acquired codebases from failed startups for resale or packaging for AI labs. Its reporting also describes interest in records that show how work was done: Slack conversations, tickets, code, and other operational records. These are reported examples, not proof that every archive was sold to a model developer.
As an Amazon Associate I earn from qualifying purchases.
A commercial seller-facing service markets a broader set of potential assets: private repositories, datasets, tasksets, reinforcement-learning environments, and closed-company workspaces that may include Slack, email, tickets, wikis, and finance records. That page describes what it seeks to market; it does not independently establish that every category has been sold.
These records may offer more context than a repository alone. Code can show what a product contains, while tickets, discussions, and documentation may capture decisions, debugging, and collaboration. Whether a particular archive is complete, useful, or suitable for training depends on its contents and context; the available reporting does not establish a standard package or quality measure.
#1 Best Overall
What the reporting does—and does not—establish
Coverage combines different stages and kinds of activity. An intermediary’s acquisition of a codebase is not the same as a lab buying it, and a lab’s discussion with a company is not evidence of a completed license. The Information has separately reported discussions involving OpenAI and Anthropic over specialized data deals, including genomics, as well as conversations involving Google DeepMind and biotechnology startups. Those reports broaden the story to active companies, but describe discussions rather than confirmed purchases.
- Reported acquisition: A data firm is reported to have obtained a startup codebase. That alone does not identify a final buyer or prove model training use.
- Offer or marketing: A seller-facing service describes assets it is willing to market. That is not transaction evidence for every listed asset type.
- Negotiation or discussion: A lab and a data holder are reported to have discussed a deal. A discussion does not establish that they agreed or completed a transaction.
- Completed transaction: This requires transaction-specific evidence; the reviewed reports do not support treating every approach or discussion as one.
The reviewed sources also do not establish a dependable count or total value for sales of defunct-startup data. Individual offers or claims should not be treated as typical prices or industry totals. That gap does not prove there are no sales; it means the available evidence does not quantify the market reliably.
Rank #2
How a defunct startup’s data differs from an active company’s data
| Question | Closed-company archive | Active-company data deal |
|---|---|---|
| What is described? | Reported codebase acquisitions and market interest in operational records such as communications, tickets, and work history. | Reported discussions about specialized datasets, including genomics, between AI labs and companies. |
| What stage is established? | The Information reported data firms acquiring codebases for resale or packaging; that does not by itself establish a completed lab purchase. | The Information described discussions, not confirmed completed purchases. |
| Potential arrangement | Acquisition by an intermediary and subsequent resale or packaging are reported; the sources do not establish the terms of any particular deal. | Specialized data licensing was discussed; the sources do not establish final terms or completion. |
A shutdown does not make an archive equivalent to a clean, owner-controlled dataset. A closed company’s records can contain information created by employees, contributors, customers, and other parties, while the reported active-company talks concern specialized data held by businesses that may still be operating. The evidence here does not settle who controls any particular archive after insolvency or acquisition, or what a specific contract allows.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Who can offer the code and records?
A seller-facing service says an owner should own an asset outright or have written permission from contributors, and says employees cannot sell an employer’s code or data. That is a market participant’s stated screening practice, not a legal opinion or a universal rule. The reviewed sources do not establish who has authority over any specific company’s records, what its contracts say, or what permissions are required in each jurisdiction.
Employee and customer records raise separate privacy and rights concerns from code ownership. A workspace could include personal or confidential material even when a company has shut down. Closure alone does not establish that customer, employee, contributor, or contractual rights have ended; authority and permissions need to be assessed for the particular records and circumstances.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Why personal-data marketplaces are related—but different
Axios reported that Verb launched a consumer-data marketplace on August 6, 2026. Users can select activity categories and offer personal data to companies. Axios also reported privacy concerns and said the marketplace excludes categories including passwords, text messages, and health data from analysis. Verb is an adjacent example of people being offered a way to monetize data; it is not evidence that the company buys archives from dead startups.
Rank #4
Verb CEO and co-founder Cyrus Beschloss told Axios: “Verb is the place that people ‘will go to monetize something that for decades we’ve given away for free,’” referring to that consumer-data marketplace. The remark should not be read as a claim about startup archives.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteQuick Recap
What readers can responsibly conclude
- There is reported buyer interest in startup code and in records of real work, not just polished datasets.
- Evidence of an intermediary acquisition, a seller’s offer, a lab discussion, and a completed lab purchase are different claims and should not be collapsed into one.
- Permissions, ownership, privacy, and contractual obligations matter for each archive; the reviewed reporting does not settle them for a particular company.
- No reliable market-wide sales count or value is established by the available sources.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
