The remains of failed startups are becoming training material for AI systems. Wind-down services are brokering sales of code and workplace archives, including email, Slack, Notion and Figma data, to AI labs and companies that build simulated work environments for agents.
SimpleClosure says its Asset Hub has handled more than 150 deals and expects to return over $10 million to startups by the end of 2026. One founder received about $25,000 for a codebase and expected roughly $15,000 more for workspace data. The buyers want records of real decisions, disagreements and difficult tasks that synthetic datasets may not reproduce.
The training benefit is not yet well quantified. Companies have not published direct comparisons showing how much these archives improve models, and prices vary because buyers estimate how many useful tasks a dataset contains.
Privacy is the larger unresolved issue. Suppliers say they remove personally identifiable and human-resources information, but ordinary calendars and messages can still expose medical appointments, family routines and other personal details. Sale proceeds also tend to go to creditors and investors rather than the employees who created the records. Without a broad US law requiring consent, valuable operational history can change hands after the organization that collected it disappears.