The conversations are preliminary, and the company may ultimately decide not to proceed with a round, or the terms could shift.
Sources indicate Mercor has informed potential investors that it has secured a term sheet at that valuation, the sources said. These active fundraising discussions have taken place over the past several weeks, less than a year after Mercor's previous funding round at a $10 billion valuation.
A person familiar with the situation attributed part of the growth to rising demand from enterprise and application-layer firms that are building their own models and rely on Mercor for data to accelerate development. However, the revenue figure represents the total amount billed to clients, not Mercor's net share after paying contractors, who receive 60‑70% of that amount, according to another source.
Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter
At the time, Meta Platforms Inc. was a client and responded by indefinitely pausing its work with the startup. "The impact on customer and contractor data was very limited," a Mercor spokesperson said, citing the findings of an internal investigation.
OpenAI and Anthropic continue to use Mercor's services, a person familiar said. Meta declined to comment, and neither OpenAI nor Anthropic replied to requests for comment.
Founded in 2023, Mercor has quickly become a key player in helping AI developers build more capable models. It competes with other data-labeling startups like Surge AI and Scale AI; Scale AI was valued at over $29 billion last year after receiving funding from Meta. Mercor has reported profitability from day one, according to earlier statements to Bloomberg, and currently pays its contractors over $4 million per day. The San Francisco-based firm employs roughly 400 full-time staff.
Mercor's rapid ascent reflects the broader boom in the AI training data market, where startups like Scale AI have also achieved multibillion-dollar valuations. The company's profitability from inception, rare in the startup world, is driven by its lean operations and a global workforce of contractors who handle tasks such as data labeling and validation. With 400 full-time employees and daily contractor payments exceeding $4 million, Mercor has built a scalable model that appeals to top AI labs.
The demand for high-quality training data has surged as AI models become more sophisticated. Mercor's business model relies on a global network of contractors who perform tasks like data labeling and validation, enabling clients to improve model accuracy. Despite the security incident, the company's rapid revenue growth underscores its critical role in the AI supply chain.
As major tech firms and startups alike race to train more advanced models, data providers like Mercor are seen as essential partners. The security breach with Meta, however, highlights the risks that come with handling sensitive data and could influence future client relationships. Despite that setback, the continued backing from OpenAI and Anthropic signals confidence in Mercor's ability to recover and scale further.
Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets