When Keelan Miskell wound down his crypto startup this year, he initially assumed the company’s data had little value: users were anonymous and the firm only a few years old. He nonetheless sold the digital archives and converted them into cash — roughly $25,000 for the codebase and a still-pending sum likely around $15,000 for workspace data stored in Google, Notion, Figma and Slack. “At least all that stuff we were doing for two and a half years is worth something to someone somewhere,” he said.
Services and buyers emerge amid rising shutdowns
Startup shutdowns have increased in recent years: estimates from Carta and AngelList show annual rises between 26% and 56%. An end-of-life ecosystem has grown around closures, including advisers, “death doulas,” and firms handling legal, tax and operational wrap-up. A new strand in that ecosystem is the resale of company data to AI development teams.
SimpleClosure, a company focused on servicing shutting startups, added data-sales to its offering in April. Its Asset Hub platform helps founders sell code and workplace data — emails, chats, and documents. Buyers include foundation AI labs, vendors that build simulated workspaces for AI training, and established tech firms developing their own models. SimpleClosure has handled more than 150 deals and expects to return over $10 million to startups by the end of 2026. Dori Yona, SimpleClosure’s CEO, says transactions that were once rare and “somewhat taboo” have become routine: 47% of SimpleClosure’s deals now include code or data sales.
Competitor Sunset has also rapidly grown a data-resale business. Brendan Mahony, Sunset’s founder and CEO, says the firm values a dying client’s assets by head count, years of operation, data volume and how the team used its tools. He likens operational-exhaust data to wood shavings: once thought worthless, it can hold considerable value.
Sudden demand as public datasets run dry
Demand for defunct companies’ archives surged last year as AI training labs reached limits with public and synthetic datasets. Surge AI, for example, buys data to build “reinforcement-learning gyms” where client AI agents practice tasks ranging from debugging complex IT architectures to drafting appellate briefs. Nick Heiner, who leads that work at Surge AI, says he looks for archives that reflect a breadth of situations, including disagreements and difficult judgment calls. Heiner notes that one company’s Slack history, if it contains varied and deep conversations, can sometimes be worth more than another company’s Slack, email and calendars combined.
Real archives avoid common failures of synthetic worlds, such as impossible date references. Yet the degree to which real-world data improves model performance remains unclear: no company has released tests comparing training on real archives versus alternatives. Pricing reflects that uncertainty: buyers quote widely different amounts because valuations are based on research theses estimating how many useful tasks, for example, can be extracted from 100,000 lines of code.
Large public deals and legal concerns
Bigger transactions have occurred despite the ambiguity. In August, amid Spirit Airlines’ bankruptcy, Google outbid the AI-data company Mercor to buy the carrier’s decades of data — including 100 million emails, 500 million Microsoft Teams messages and 176,000 employee records — for $10 million. Bloomberg reported in September that SpaceX’s AI unit was considering purchases of data from struggling startups.
These deals raise legal and ethical questions. Spirit’s flight attendants’ union objected to the sale on privacy grounds, and Greg Ewing, an attorney advising AI startups, has called such data “legally radioactive.” Suppliers employ data scientists to scrub personally identifiable information and to exclude HR records, and buyers say they also perform scrubbing. Ewing warns, however, that private details frequently appear in emails and calendar entries — doctor appointments, school pickup arrangements or date-night plans — and describes the market as a “Wild West of AI training” only beginning to move toward proper licensing agreements and vetted uses.
Who benefits from the wind-down proceeds?
Proceeds from these sales typically flow to senior creditors and investors rather than to employees. Ifeoma Ajunwa, a law professor at Emory University and author of The Quantified Worker, characterises this as a form of “captured capital.” Legal precedent allows monitoring to assist workers in doing their jobs, but Ajunwa argues that justification disappears once a firm ceases operating. “Employers should have express consent before selling workers’ data,” she says, noting there is no federal law banning such sales.
Outlook
Many founders and investors see the market as a way for failing businesses to recover some value, but wind-down data sales have limits: founders often lack visibility into who buys their data and how it will be used, and there is little assurance of fair pricing. Data prices may stabilize over the coming year unless an AI-market correction occurs first. In any case, the new class of tech undertakers appears poised to continue appraising and monetizing what remains.
