The startup that pays doctors and lawyers to train AI models is now worth twice what it was eight months ago, if new funding talks go through.
Eight months ago, Mercor was worth $10 billion. Now it’s in talks that could double that figure to $20 billion, according to Bloomberg. For a company founded by three college dropouts in 2022, that trajectory is the kind that makes venture capitalists cancel their weekend plans.
Mercor connects major AI labs with domain experts, think doctors, lawyers, financial analysts, and media professionals, to generate high-quality training data for machine learning models. The business is simple in concept and lucrative in practice: AI companies need humans who actually know things, and Mercor supplies them at scale.
The numbers behind the valuation #
Mercor raised $350 million in a Series C round in October 2025, vaulting its valuation from $2 billion to $10 billion in a single step. That earlier round, the Series B, had closed in February 2025. So the company went from a $2 billion valuation to discussions around $20 billion in roughly eight months.
Total funding raised since the company’s founding now exceeds $490 million.
Revenue tells a similar story. Mercor’s annualized gross revenue is projected to approach $2 billion by mid-2026, up from $1 million not long ago.
The daily economics are striking. Mercor reportedly pays out between $1.5 million and $2 million per day to domain experts for their contributions to AI training pipelines.
Co-founders Brendan Foody, Surya Midha, and Adarsh H became some of the youngest self-made tech billionaires following the October 2025 round.
Why the AI training market keeps inflating #
The technical term for what Mercor facilitates is reinforcement learning from human feedback, or RLHF. The simpler version: AI models learn better when actual experts tell them what good looks like. A language model trained partly on a cardiologist’s corrections to its medical reasoning is more useful than one trained purely on text scraped from the internet. Mercor industrialized that correction process.
What investors need to watch #
A $20 billion valuation for a company approaching $2 billion in annualized revenue implies a multiple that would raise eyebrows in most sectors. The question investors are wrestling with is whether that revenue is durable or whether it compresses as AI models become more capable of self-evaluating.
There’s also an operational risk that materialized in plain sight. In late March 2026, Mercor disclosed a data breach tied to a supply-chain attack. For a company whose entire value proposition rests on connecting sensitive domain expertise, including medical and legal knowledge, to AI systems, a security incident is not just a PR problem. It raises questions about the company’s ability to protect proprietary workflows and expert-contributed data at scale.
The breach did not appear to derail the valuation conversation, at least not visibly. But it is the kind of event that tends to matter more in due diligence than in headlines, and sophisticated investors evaluating a $20 billion price tag will scrutinize the incident and Mercor’s response closely.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our