As Keelan Miskell closed the doors on his crypto venture earlier this year, he underestimated the potential value of the data accumulated during its short lifespan. However, he has since transformed that digital footprint into tangible assets, netting approximately $25,000 from the sale of his codebase and an additional pending amount of around $15,000 for data stored across platforms like Google, Notion, Figma, and Slack. Reflecting on the experience, Miskell noted, “At least all that stuff we were doing for two and a half years is worth something to someone somewhere.”
This transaction was facilitated by SimpleClosure, a company pioneering a niche market focused on assisting defunct startups. With estimates from Carta and AngelList indicating a significant rise in startup closures—between 26% and 56% annually—a new ecosystem is evolving. This includes coaches, so-called “death doulas,” and companies specializing in the legal and operational intricacies of winding down. SimpleClosure has recently expanded its services to include data sales, enabling founders to monetize their operational data and code. The company’s Asset Hub platform has successfully brokered over 150 deals and anticipates returning over $10 million to struggling startups by 2026. SimpleClosure’s CEO, Dori Yona, remarked that these previously rare and sensitive transactions are becoming increasingly commonplace, with nearly half of their business deals now encompassing data or code sales.
The demand for data from shuttered startups gained momentum last year as AI training labs reached the limits of available public datasets. Organizations like Surge AI are actively purchasing these archives to create “reinforcement-learning gyms” where AI agents can practice complex tasks. Nick Heiner, who leads the development of these environments, emphasizes the unique value of diverse conversation histories found in platforms like Slack, which can be more beneficial than a larger but less varied dataset. While the advantages of using real-world data over synthetic alternatives remain to be fully quantified, significant transactions involving data from failed firms are on the rise. For instance, Google made headlines in August by outbidding competitors for Spirit Airlines’ extensive data during its bankruptcy, underscoring the high stakes in this emerging market.
However, the legal landscape surrounding these transactions is murky. Concerns about privacy and the ethical implications of selling employee data have surfaced, particularly following objections from unions like Spirit’s flight attendants, who challenge the legality of such sales. Greg Ewing, an attorney specializing in AI startups, has flagged the potential risks of handling sensitive information, noting that while data scrubbing is standard practice, the presence of identifiable details can still pose significant challenges. Furthermore, as these transactions unfold, questions arise about who truly benefits from the sale of this data, with proceeds often directed toward senior creditors rather than employees. Legal expert Ifeoma Ajunwa highlights the necessity for explicit consent from employees regarding the sale of their data, underscoring a significant gap in current federal regulations. As this new market continues to evolve, founders must navigate an uncertain landscape, balancing the recovery of value from their failed startups with the ethical and legal complexities surrounding data ownership.
Source: AI labs are buying data from dead startups to train their models via MIT Technology Review