# Mercor seeks $20B valuation as AI training spend pours into human data > Source: > Published: 2026-07-09 19:25:53+00:00 [Mercor](https://www.mercor.com/) is in early-stage talks to raise new capital at a $20 billion valuation, [TechCrunch reported](https://techcrunch.com/2026/07/09/mercor-is-in-talks-for-a-20b-valuation/) Thursday, citing [Bloomberg](https://www.bloomberg.com/news/articles/2026-07-09/ai-training-startup-mercor-discusses-20-billion-valuation?srnd=phx-technology), a second valuation reset in less than nine months for [Brendan Foody (@BrendanFoody)](https://x.com/BrendanFoody/status/2074193875695214907) and the AI training marketplace he co-founded. The talks remain early, according to TechCrunch, but Bloomberg reported that Mercor told investors it had already received a term sheet at the new valuation. A $20 billion price would double Mercor's last confirmed valuation from October 27th, 2025, when Mercor raised a [$350 million Series C at a $10 billion valuation](https://techcrunch.com/2025/10/27/mercor-quintuples-valuation-to-10b-with-350M-series-c/) led by Felicis Ventures, with Benchmark, General Catalyst and Robinhood Ventures also participating. Foody, Adarsh Hiremath and Surya Midha have become the young-founder faces of the human-data boom. The three were 21-year-old Thiel Fellows when Mercor raised its $100 million Series B at a $2 billion valuation in February 2025, and Forbes later described them as Bay Area high school friends and debate teammates who became the world's youngest self-made billionaires after the October 2025 Series C. Mercor began as an AI-driven hiring marketplace, then shifted toward supplying AI labs with domain experts - doctors, lawyers, scientists, bankers and engineers - who create and judge the data used to train frontier models. The valuation report landed the same day Mercor announced two pieces of narrative support for that price: Foody wrote on X that Mercor's annualized revenue run rate had crossed $2 billion, and Mercor said it would acquire [Deeptune](https://deeptune.ai/), a reinforcement-learning environments startup. The run-rate figure is the number investors are being asked to underwrite. At $20 billion, Mercor would be valued at roughly 10 times Foody's self-reported annualized revenue run rate. That multiple looks restrained compared with some AI application rounds only if the revenue is treated like software revenue. Mercor is a marketplace that bills clients and pays contractors, and [Sacra's recent Mercor report](https://sacra-pdfs.s3.us-east-2.amazonaws.com/mercor.pdf) describes Mercor's reported revenue as total customer spend before contractor payouts, rather than net revenue retained after paying experts. That distinction matters because gross marketplace volume carries different margins and durability than subscription software. Mercor has leaned into the scale of the underlying payment flow. In a [May 8th engineering post](https://www.mercor.com/blog/when-you-go-from-2-million-a-month-to-2-million-a-day/), Mercor said it had gone from $0 to a $500 million run rate in 17 months, crossed $1 billion in annualized revenue run rate earlier in 2026, and was paying more than $2 million per day to more than 30,000 weekly active contractors. Mercor also said in that post that its payments infrastructure had been stretched by growth from $200,000 to more than $14 million a week on the platform. ### Deeptune is the strategic tell Mercor's [Deeptune acquisition announcement](https://www.mercor.com/blog/mercor-to-acquire-deeptune/) shows where Foody wants Mercor to sit in the AI supply chain. Mercor said Deeptune builds software environments for reinforcement learning, including recreations of enterprise applications from spreadsheets to Salesforce, and that Mercor had already been a Deeptune customer. Mercor said the entire Deeptune team will join Mercor as Mercor expands in New York. The acquisition price was not included in Mercor's announcement. Deeptune had raised a [$43 million Series A](https://fortune.com/2026/03/19/andreessen-horowitz-ai-startups-deeptune-series-a/) in March 2026 led by Andreessen Horowitz, with 776, Abstract Ventures and Inspired Capital participating, according to Fortune. Andreessen Horowitz separately wrote that Deeptune was focused on training environments for AI models. That move pulls Mercor beyond labor matching. Mercor already supplies the expert layer: people who write tasks, create verifiers and judge model performance. Deeptune adds the software layer where agents practice the work. If the bottleneck in AI training is moving from static datasets to interactive environments, Mercor is trying to own more of the production stack rather than stay a broker of expert hours. Foody framed the Deeptune deal in those terms. In Mercor's announcement, he wrote that frontier models can pass exams but often fail inside real workflows because they lack experience. Mercor's experts, he wrote, are already writing evals, while Deeptune provides the software platform where that expertise can become training environments at scale. ### The risks did not disappear The new valuation talks also arrive after a difficult spring for Mercor. In April, [TechCrunch reported](https://techcrunch.com/2026/04/09/after-data-breach-10b-valued-startup-mercor-is-having-a-month/) that Mercor had suffered a data breach and that multiple contractors had filed lawsuits after alleged exposure of personal data. [WIRED reported](https://www.wired.com/story/meta-pauses-work-with-mercor-after-data-breach-puts-ai-industry-secrets-at-risk/) that Meta paused work with Mercor while investigating the incident, and that OpenAI was investigating whether its proprietary training data may have been exposed. HR Dive later reported that plaintiffs filed class-action lawsuits in California federal court alleging damages tied to the breach. Those issues cut directly against Mercor's pitch. Mercor handles sensitive training workflows for AI labs and personal information from contractors. A marketplace that depends on elite customers and expert labor has to convince both sides that Mercor can manage security, payments and trust at the same pace as Mercor's revenue growth. The fundraising talks show investors may be willing to price through that risk if Mercor's revenue run rate is real and durable. The Deeptune acquisition gives Mercor a cleaner story than raw contractor scale: Mercor wants to be the system that builds, staffs and measures the environments where AI agents learn economically useful work. The unresolved question is what Mercor is actually selling investors in this round. TechCrunch's report does not include a round size, lead investor or whether the $20 billion figure is pre-money or post-money. Those details will determine whether this is a capital raise to fund acquisitions and infrastructure, a mark-up round to establish a new private-market price, or both. For now, Foody has the market's attention because Mercor sits in a spending category the AI labs cannot avoid. Models still need humans to define good work, score edge cases and turn messy expert judgment into training signals. Mercor's bet is that the same labor marketplace that supplied those humans can become the software layer around them before AI labs or rival data vendors do it themselves.