How artificial intelligence labs use Mercor to obtain data that companies do not share

Instead of signing expensive contracts with companies for their data, artificial intelligence labs are now trying a recent tactic: using former senior employees of those companies to gain their industry knowledge, Mercor CEO Brendan Foody said Tuesday at TechCrunch Disrupt 2025.

Speaking on stage, Foody hailed the Mercor marketplace as one of the essential channels connecting former employees of investment banks, consulting houses and law firms with artificial intelligence labs looking to automate these industries. Some of Mercor’s clients include OpenAI, Anthropic and Meta.

“There is an argument that Goldman Sachs doesn’t like the idea of ​​having models that can automate their value chain,” Foody said, citing the Wall Street giant as an example. “It definitely changes the competitive dynamic, and that’s one of the reasons labs need us. Their customers don’t want to give them the data to automate large parts of their value chains, so they need to hire contractors who have previously worked at these companies, understand these workflows, and are willing to train models to automate them.”

Foody, Mercor’s 22-year-old co-founder, says his startup pays industry experts up to $200 an hour to fill out forms and write AI training reports. The company now has tens of hundreds of contractors and says it gives them greater than $1.5 million a day. Still, Foody says the startup stays profitable because AI labs are willing to pay much more for that beneficial data.

In lower than three years since its founding, Mercor has grown its annual recurring revenue to roughly $500 million and recently raised $10 billion in financing.

Established companies across the economy have good reason to oppose Mercor’s advancement because knowledge about their industry may very well be slipping out the back door by former employees in the startup’s market, which could ultimately be used to automate their work. Foody admitted it could expose market inefficiencies, but he would not call it a “gap.”

Foody says some companies are already benefiting from this “new future of work.” He was amused by the idea that Mercor’s marketplace could create a recent form of gig economy, much like Uber did greater than a decade ago. (Earlier this yr, former Uber chief product officer Sundeep Jain joined Mercor as president).

“There are companies that are taking advantage of this and realizing that the world is going to change very quickly,” Foody said. “There is definitely another category of companies that are afraid and concerned that they will be disintermediated and that their customers will go directly to AI labs or application layer platforms. My hunch is that this first category will turn out to be on the right side of history.”

While Mercor tries to mine knowledge across industries, Foody said his startup seeks to prevent contractors from committing corporate espionage — the illegal act of stealing proprietary information, trade secrets or mental property from one company and selling it to one other.

But that’s easier said than done. Most of Mercor’s employees are former employees of law firms, investment banks and other industries that keep their data very confidential. Foody said some Mercor contractors proceed to work in their day jobs and simply submit data on the side, and also stated that contractors have been instructed not to submit documents from their previous workplace. However, he admitted that given the scale of his startup, “some things happen.”

Foody argues that the knowledge in an worker’s head belongs to the worker himself, not his company – a more generous view than many companies would take. Moreover, in some Mercor job advertisements, the startup balances on the border between the worker’s demand for knowledge and his company’s data.

For example, Mercor is currently searching CTO or co-founder of a startup who “can authorize access to a significant production codebase” for AI assessments or potentially AI model training. In an email, Mercor told TechCrunch that several startup CTOs took advantage of the offer but declined to disclose details of their contracts.

Mercor was one of the first data startups to recruit highly expert research employees in the US and pay them large sums to train artificial intelligence models. In the early days of the AI ​​boom, data providers like Scale AI hired contractors in third world countries to perform fairly easy labeling tasks. Now, most of Mercor’s competitors – including Surge and Scale AI – have recognized that AI labs need experts to improve their AI models. Many data providers have also begun training “environments” to improve the ability of AI agents to perform real-world tasks.

Mercor clearly benefited from Scale AI’s woes: many artificial intelligence labs stopped working with Scale AI after Meta made a large investment in the startup and hired its CEO. Mercor has quintupled in value in the last yr, but is still smaller than Surge and Scale AI, which are valued at greater than $20 billion.

Currently, most of Mercor’s revenue comes from just a few artificial intelligence labs, but Foody says the startup plans to partner with other industries in the future. He believes that companies in the legal, financial and medical industries will need assistance in using their data to train AI agents – something Mercor specializes in.

“Over time, ChatGPT will be better than the best consulting firm, better than the best investment bank and better than the best law firm,” Foody said. “This will radically transform the economy, which will be a fundamentally positive force that will help create abundance for all.”