AI Data Access: How Labs Use Mercor for Exclusive Insights

The Shift in AI Data Acquisition Strategies

Rather than entering into costly data agreements with established corporations, artificial intelligence laboratories are increasingly focusing on recruiting former senior-level personnel from these very companies to leverage their specialized industry expertise. This trend was highlighted by Mercor CEO Brendan Foody during his appearance at TechCrunch Disrupt 2025 on Tuesday.

Mercor's Role as an Intermediary

Foody positioned Mercor’s platform as a primary facilitator, connecting ex-employees of investment banks, legal firms, and consulting agencies with AI labs aiming to automate processes within those sectors. Notable clients of Mercor include prominent entities such as OpenAI, Anthropic, and Meta.

The competitive landscape is being reshaped, as evidenced by the potential for models to automate the operations of companies like Goldman Sachs. Consequently, AI labs are reliant on services like Mercor, as their clientele are often hesitant to directly provide the data necessary for automating core business functions. Instead, they seek contractors possessing prior experience and a willingness to develop models for automation.

Mercor's Business Model and Growth

Mercor, founded by the 22-year-old Foody, compensates industry professionals at rates reaching $200 per hour for completing forms and generating reports used in AI training. The company currently employs tens of thousands of contractors and distributes over $1.5 million to them daily.

Despite these substantial payouts, Mercor maintains profitability due to the high value placed on the acquired data by AI laboratories. In less than three years, the startup has achieved an annualized recurring revenue of approximately $500 million and recently secured funding at a $10 billion valuation.

Disruption and the Future of Work

Incumbent organizations have legitimate concerns regarding Mercor’s expansion, as their proprietary knowledge may be disseminated through former employees utilizing the platform, potentially leading to automation of their workflows. Foody acknowledges this market inefficiency but refrains from labeling it a “loophole.”

He suggests that some companies are proactively adapting to this evolving work paradigm. Foody envisions Mercor’s marketplace potentially establishing a new gig economy model, akin to the impact Uber had over a decade ago. Notably, Uber’s former chief product officer, Sundeep Jain, recently joined Mercor as president.

“Certain companies are embracing the inevitable changes, while others express apprehension about potential disintermediation and direct engagement between customers and AI labs,” Foody stated. He believes the former group will ultimately be proven correct.

Navigating Ethical and Legal Considerations

Mercor actively strives to prevent contractors from engaging in corporate espionage – the unlawful acquisition and transfer of confidential business information. However, this proves challenging given the nature of its workforce.

Many Mercor contractors are former employees of organizations with strict data confidentiality protocols, and some continue to hold their previous positions while contributing data on a part-time basis. While contractors are instructed against uploading documents from their former employers, Foody concedes that unforeseen occurrences are possible given the scale of the operation.

Foody posits that an employee’s knowledge resides with the individual, not the company – a perspective that diverges from many corporate viewpoints. Furthermore, some of Mercor’s job descriptions blur the line between soliciting an employee’s expertise and requesting access to their company’s data.

Data Access and Competitive Landscape

Currently, Mercor is seeking a CTO or co-founder from a startup who can grant access to a “substantial, production codebase” for AI evaluation or model training. Mercor confirmed that several startup CTOs have expressed interest, though contract details remain undisclosed.

Mercor was among the first companies to actively recruit and compensate highly skilled professionals for AI model training. Initially, data vendors like Scale AI relied on contractors in developing nations for basic labeling tasks. However, competitors like Surge and Scale AI have recognized the need for expert knowledge to enhance AI model performance and have begun developing specialized training “environments.”

Market Dynamics and Future Expansion

Mercor has benefited from challenges faced by Scale AI, particularly after Meta’s investment and subsequent hiring of Scale AI’s CEO. Mercor’s valuation has increased fivefold in the past year, though it remains smaller than Surge and Scale AI, both valued at over $20 billion.

Currently, a significant portion of Mercor’s revenue originates from a limited number of AI labs. However, the startup intends to expand its partnerships to encompass other industries. Foody anticipates demand from legal, financial, and medical sectors for assistance in leveraging their data to train AI agents – a core competency of Mercor.

“Ultimately, ChatGPT will surpass the capabilities of leading consulting firms, investment banks, and law firms,” Foody predicts. “This transformation will fundamentally alter the economy, fostering abundance for all.”