An AI Recruiter Just Hit a $10 Billion Valuation. Here Is Exactly What It Automates - And What It Does Not.

In October 2025, AI hiring platform Mercor closed a $350 million Series C funding round led by Felicis, with participation from Benchmark, General Catalyst, and Robinhood Ventures, valuing the company at $10 billion. Eight months earlier, in February 2025, Mercor had been valued at $2 billion. That is a five-fold increase in under a year, and by the middle of 2026 the company continues to scale, having grown from $1 million to over $1 billion in booked revenue in roughly 20 months.

Founded in January 2023 by three college dropouts - Brendan Foody, Adarsh Hiremath, and Surya Midha, all former high school debate partners in San Jose - Mercor's rise makes its founders, at age 22, among the youngest self-made billionaires in the world. For recruitment agency owners, the more important question is not how fast the company grew, but exactly what it automates, and what that means for the fee conversation with clients.

What Mercor actually does

At its core, Mercor's platform runs full candidate evaluations before any human recruiter or hiring manager looks at a profile. Candidates complete a roughly 20-minute video interview, which an AI system scores across multiple dimensions - not simply keyword matching against a CV, but an assessment closer to what a skilled interviewer would produce. The company reports this cuts hiring cycles by around 80% and reduces cost by roughly 70% compared to a traditional recruitment agency, drawing on a pool of more than 300,000 professionals. Its fee - roughly 30% of first-year compensation - sits in the same order of magnitude as a traditional contingency or retained search fee.

That last point is worth pausing on. Mercor is not primarily competing on price. Its fee is comparable to what a human agency charges. What it is proving is that the screening and shortlisting step - historically the most labour-intensive part of a recruiter's week - does not require a human in the loop at all, delivered at a scale and speed no boutique agency can match.

An important caveat: who Mercor actually serves

It would be a mistake to read Mercor's growth as a signal that mainstream corporate recruitment is being wholesale replaced. The company's dominant current customer base has increasingly skewed toward AI labs and technology companies sourcing highly specialised domain experts - scientists, doctors, lawyers, and other professionals - to perform AI model training work, rather than general corporate hiring across every industry. Mercor reportedly manages more than 30,000 contractors through this model, collectively paid over $1.5 million a day.

Some of this shift has been accelerated by ripple effects elsewhere in the market: after Meta's $14.3 billion investment for a 49% stake in rival data-labelling firm Scale AI, concerns about Scale's neutrality reportedly led some major AI labs, including Google and OpenAI, to redirect work toward alternatives like Mercor. This context matters because it means Mercor's core screening technology - not its current customer mix - is the part of the story most directly relevant to recruitment agencies. The mechanism (AI evaluates and scores candidates before a human sees them) is real and proven at scale. The claim that it has already replaced general corporate recruiting end to end is not yet supported by the same evidence.

Where the fee argument actually goes

For a recruitment agency owner, the useful exercise is not deciding whether Mercor specifically will take a given piece of business. It is recognising which part of the traditional recruitment fee is being priced by the market as automatable, at scale, today.

Historically, a recruitment fee has bundled several distinct activities into one number: sourcing candidates, screening and shortlisting them, managing the interview process, and providing judgement about fit, risk, and negotiation. Mercor's technology - and the wider category of AI hiring tools it represents - makes a credible case that the sourcing and screening components can be done by AI, at comparable cost to a human agency's own fee, before a client conversation even begins.

That leaves the remaining components of the fee to justify themselves on their own terms:

• The read on cultural and team fit - judging whether someone will actually thrive with a specific manager, in a specific team, under specific pressure, which an AI-scored interview does not directly assess
• Honest counsel - being willing to tell a client that the role, the salary band, or the hiring brief itself is wrong, something a scoring algorithm has no incentive or capacity to do
• The relationship - being the first call a client makes before a role is even formally opened, which requires an ongoing human relationship no platform can substitute for
• Risk ownership - standing behind a placement's success over the following months, rather than handing over a shortlist and moving on

The practical test for any recruitment desk

A simple diagnostic: think of the last placement where a client explicitly chose your shortlist over a cheaper alternative. Can you state, in one sentence, the specific reason they chose you? If the honest answer involves speed, volume, or "we found good candidates" - all things an AI-scored platform can now credibly claim to do as well or better - that is the exposed part of the fee. If the answer involves a judgement call, a piece of counsel, or a relationship that predates the search itself, that is the part of the business worth building the pitch around going forward.

Mercor's $10 billion valuation is not a story about one company undercutting recruiters on price. It is a market signal, backed by serious institutional capital, that the screening function specifically has become commoditised technology. The agencies that will hold their fees over the next several years are the ones who can already answer, specifically, what they are charging for once screening stops being the answer.