How AI is reshaping work and who gets to do it, according to Mercor's CEO - Equity Recap

Podcast: Equity

Published: 2026-01-02

Duration: 25 min

Summary

This episode explores how Mercor is transforming the AI labor landscape by connecting highly skilled professionals with AI labs, addressing the challenges of corporate data sharing and workforce automation.

What Happened

In this episode of Equity, host Anthony Ha engages with Brendan Foody, the CEO of Mercor, a startup that has rapidly established itself as a $10 billion intermediary in the AI data marketplace. Mercor employs highly skilled contractors, often former employees from top firms like Goldman Sachs and McKinsey, to create valuable training data for AI labs such as OpenAI and Anthropic. This innovative model allows companies to benefit from expert knowledge without directly sharing sensitive internal data.

Foody shares insights into Mercor's origins, revealing that it began as a college project among friends who sought to automate hiring processes. As the AI industry evolved, Mercor pivoted to focus on sourcing high-level expertise required to train sophisticated AI models. Foody highlights the structural efficiency of having experts train AI agents rather than having them perform repetitive tasks, which could lead to significant shifts in knowledge work across various industries. He also acknowledges the concerns of companies like Goldman Sachs, which may view AI as a potential threat to their traditional business models but emphasizes the potential for AI to create new opportunities for growth.

Key Insights

Key Questions Answered

How does Mercor source its contractors?

Mercor primarily hires highly skilled contractors who often have backgrounds at prestigious firms such as Goldman Sachs, McKinsey, and White Shoe law firms. These contractors are paid competitive rates, averaging around $95 an hour, to provide their expertise and create valuable training data for major AI labs. The company has seen rapid growth, paying out $1.5 million a day just 18 months after identifying this market opportunity.

What challenges do AI labs face regarding data sharing?

AI labs encounter significant challenges in obtaining data directly from corporations, primarily due to concerns about confidentiality and competitive dynamics. Companies are wary of sharing sensitive data that could ultimately empower competitors. As a result, Mercor plays a crucial role by acting as an intermediary that allows these companies to share their insights without directly compromising their proprietary information.

How is AI expected to change traditional business models?

Brendan Foody suggests that AI models like ChatGPT are poised to outperform traditional consulting firms, investment banks, and law firms, leading to a radical transformation in the economy. This shift may create an environment of abundance, but it also raises concerns among established firms about being disintermediated. Companies that embrace AI technology may find new avenues for growth, while those that resist may struggle to keep up.

What types of expertise does Mercor seek for AI training?

Mercor looks for individuals with specialized skills in fields such as law, finance, and technology, specifically targeting professionals who can contribute to training sophisticated AI models. This includes former investment bankers, lawyers, and software engineers. The focus is on sourcing high-level expertise that can help automate complex tasks and workflows, moving away from the need for low-skilled crowdsourcing.

How does Mercor ensure confidentiality with its contractors?

To maintain confidentiality, Mercor establishes contracts with all experts that explicitly prohibit the sharing of proprietary company data. This legal framework helps alleviate concerns from corporations about potential data leaks while allowing Mercor to utilize the valuable insights and expertise of its contractors to train AI models effectively.