A newly reported trend reveals that artificial-intelligence labs are increasingly turning to the startup Mercor as a workaround for accessing data that companies are unwilling to directly share. Mercor offers a platform that connects AI-developers with former insiders from industries like banking, consulting and law, who are paid to describe internal workflows, complete tasks and essentially provide domain expertise that replaces hard-to-obtain corporate datasets. Source companies that are hesitant to hand over raw data nonetheless find themselves indirectly contributing: their former employees monetize their experience on Mercor, and AI labs get the structured knowledge they need to train advanced models. Mercor is already valued at roughly $10 billion and is servicing major labs, with plans to expand into other sectors. Another outlet notes the company’s explosive growth, describing its revenue run-rate and deep-domain-expert sourcing model. While proponents see this as an efficient way to fuel AI progress, critics question the ethics and confidentiality risks of tapping insiders rather than formal data-sharing agreements.
Sources: Yahoo News, TechCrunch
Key Takeaways
– AI labs are increasingly bypassing traditional data-sharing deals by working with platforms like Mercor that leverage former insiders’ knowledge and workflows.
– The model raises serious questions about corporate confidentiality, intellectual property and whether companies’ internal processes can become de-facto “shared data” through third-party intermediaries.
– Mercor’s rapid ascent — massive valuation, heavy investor interest — highlights how lucrative and competitive the behind-the-scenes domain-expert and knowledge-work market has become in the AI supply chain.
In-Depth
In the evolving landscape of artificial intelligence development, one of the toughest bottlenecks has been access to high-quality domain-specific data: structured, granular, proprietary workflows or internal processes that major companies often guard zealously. Traditional approaches have relied on licensing agreements, partnerships, or in-house data generation. But as the pace of AI innovation quickens, labs are increasingly tapping into workaround strategies. Enter Mercor — a startup designed to bridge the gap between the vast knowledge locked inside corporations and the voracious data hunger of AI labs.
According to the TechCrunch piece published on October 29, 2025, Mercor has built a marketplace that recruits former senior employees from investment banks, law firms and consulting organizations and pays them to carry out tasks like filling out forms, describing internal processes or writing reports. Those tasks produce structured knowledge which AI labs then use to train foundational models. The article explains that many companies are reluctant to allow their internal data – whether trade logs, operational workflows or confidential documents – to be used by outside AI developers, fearing automation might erode their competitive edge. As a result, these companies unsuspiciously become indirect suppliers: their ex-employees now monetize the experience they built up inside and outsource it via Mercor’s platform.
While Mercor’s transformational potential is clear — the firm reportedly has a multi-hundred-million-dollar run-rate and valuation near ten-billion dollars — the model draws significant scrutiny. On the one hand, this structure offers a faster, more flexible way for labs to acquire domain depth without negotiating lengthy contracts or navigating restrictive NDAs. On the other hand, it raises ethical, legal and risk-management issues. For example, to what extent does a former employee’s knowledge truly belong to them versus being proprietary to the company they left? And if an insider shares structured insights that reconstruct sensitive internal processes, does that de facto become an unauthorized transfer of corporate data?
Critics point out that while Mercor claims to forbid uploading internal documents, the risk of tacit information leakage is real, especially when knowledge workers operate outside formal employment structures. The value of that domain-specific knowledge is high – for instance, a former investment bank quant or law-firm partner may know key decision-making flows that can be used to model and automate tasks within an AI system. Mercor’s business model essentially monetizes this by turning human expertise into training data.
From a conservative perspective, one might argue that this model represents both entrepreneurial ingenuity and a potentially destabilizing shift in how corporate intellectual property is managed. Companies traditionally built competitive barriers via confidential data, proprietary workflows and hidden processes. If that advantage can be circumvented by a marketplace that pays insiders to translate it into generic knowledge, the competitive edge may erode over time. Furthermore, the governance frameworks around such transfers are still nascent: many companies may not have anticipated that ex-employees would funnel their institutional know-how to third-party AI labs.
Legally, there may be risk. If an insider shares information covered under non-compete clauses or confidentiality agreements, firms could assert trade-secret misappropriation. Indeed, though not covered in the original article, the broader industry has seen lawsuits over such transfers (for example, between major data-labeling firms). That underscores the importance of clear legal protections and strong IP governance.
From the AI labs’ standpoint the appeal is obvious: You gain access to high-value, domain-specific expertise quickly, and your model training pipeline becomes richer. For firms pressured to bring advanced models online rapidly, this knowledge-work circuit may be an efficient shortcut. But the question remains: at what cost to corporate security, competitive markets and the rule of intellectual-property law?
In sum, Mercor’s emergence signals a larger trend in the AI ecosystem: the commoditization of human domain expertise as training input, and the growing market for knowledge-work services that stand between corporations and AI labs. It prompts a reflection on whether our governance regime around corporate knowledge transfer, insider expertise and data rights is keeping pace with the speed of technological change. Firms attentive to safeguarding their intellectual assets — especially in sectors like finance, consulting and law — would do well to revisit how they classify, protect and manage the institutional knowledge of their workforce. At the same time, investors and technologists might view the growth of marketplaces like Mercor as a new frontier in AI infrastructure: not hardware or cloud services, but the human-knowledge ecosystem that trains the machine.