Brendan Foody

AI · AI training data
United StatesAge 23

Co-founder and CEO of Mercor

Never held a formal job: first found overseas engineers for American companies at a São Paulo hackathon, dropped out at 20 to build full-time, and two years later turned a recruiting marketplace into AI-training infrastructure

Dropped out to start upFirst-time founderScaled company$100M+ · 2026-04VC-backedSilicon ValleyFrom United StatesBorn Apr 2003Thiel Fellow(2024)

Key numbers

Each figure includes its definition, date, and evidence level.

2003Born · age 23
~20First venture · 2023
Self-reported$10B↗Series C valuation · 2025-10
Self-reported$350M↗Series C funding · 2025-10
Media report$1B↗annualized revenue run rate · 2026-04
Estimateapproximately 22% per person↗estimated founder ownership · 2025-10
Media report$6M↗first-half profit · 2025-06

Key stages

What he was doing at each age

~20

Won the first deal at a hackathon

~20

Dropped out at 20 to build a company

22

Reached a $10 billion valuation at 22

Geographic moves

The city behind each change in direction

  1. California
    2003 Born in the United States
  2. United States
  3. São Paulo
    ~20 Won the first deal at a hackathon
  4. Washington, D.C.
  5. San Francisco
    ~20 Dropped out at 20 to build a company~21 Co-founder and CEO22 Reached a $10 billion valuation at 22

Selected work

Shifted from an AI recruiting and interviewing platform to organizing professionals, evaluations, and training data for frontier model labs

How the business works

sector
AI · AI training data / expert network
users and payers
Model labs and enterprises pay; programmers, lawyers, doctors, finance professionals, and other experts are the supply side, paid by the hour or by deliverable
pricing model
Two eval engagement models: embedded experts and task-based delivery
main costs
expert compensation, operations, and delivery costs
verification level
profit unauditedForbes reported that the company claimed $6 million in first-half 2025 profit, with no public audit; the higher 2026 annualized revenue cannot be used to extrapolate profit
moat
The more reusable the matching and quality tools are across projects, the better the unit economics (analytical judgment)
main risk
It remains an operations-heavy business—the profit margin is the difference between revenue and expert compensation and delivery costs

Mercor organizes programmers, lawyers, doctors, finance professionals, and other experts on one side, and delivers training data, evaluations, and Agent projects to model labs and enterprises on the other. It currently offers two eval engagement models: embedded experts and task-based delivery. Experts are paid by the hour or by deliverable.Official eval offering

Customers pay because recruiting, screening, and quality-managing scarce experts is difficult to build on demand in-house. The difference between revenue and expert compensation, operations, and delivery costs forms the early profit margin; the more reusable the matching and quality tools are, the better the unit economics. But this remains an operations-heavy business. Forbes reported that Mercor claimed $6 million in first-half 2025 profit; the figure was not publicly audited, and the higher 2026 annualized revenue cannot be used to directly extrapolate profit.Forbes

Timeline

  1. Born in the United States

    Starting outUnited States · California

    Verified director records from the UK's Companies House are public only through April 2003, so no specific date is added.

    Source ↗
  2. Validated cross-border engineer matching at a São Paulo hackathon

    About 20 years oldTurning pointUnited States → Brazil · São Paulo · Georgetown student

    With high-school debate teammates Adarsh Hiremath and Surya Midha, tried to match overseas engineers with American companies; the first deal showed the three that talent matching and delivery could be productized.

    Source ↗
  3. Dropped out of Georgetown to found Mercor full-time

    About 20 years oldTurning pointWashington, D.C. → San Francisco · Co-founder and CEO

    Foody serves as CEO and Hiremath as CTO; Midha later moved from COO to chairman. The three founders' first long-term formal job was their own company.

    Source ↗
  4. Shifted from the general recruiting market to an expert network for AI training and evaluation

    About 21 years oldTurning pointUnited States · San Francisco · Co-founder and CEO

    Customer demand expanded from screening engineers to organizing lawyers, doctors, finance professionals, consultants, and others to create evaluation standards and training feedback for models.

    Source ↗
  5. Completed a $350 million Series C at age 22, reaching a $10 billion valuation

    22 years oldMilestoneUnited States · San Francisco · Co-founder and CEO

    Based on the financing price and estimated ownership, Forbes said the three founders became the world's youngest self-made billionaires at the time; this is a valuation of private-company equity, not wealth already cashed out.

    Source ↗

Three founders, not one prodigy

Mercor’s story must name all three people. Brendan Foody is CEO, Adarsh Hiremath is CTO, and Surya Midha moved from COO to chairman. They met on a high-school debate team, later entered different universities, and left together to start the company in 2023.

The value of that shared history is more than “knowing each other early.” Debate training requires quickly researching unfamiliar industries, breaking down arguments, and building evaluation standards—abilities Mercor later used when designing expert evaluations and rubrics for models.

Starting with one cross-border recruiting transaction

At a hackathon in São Paulo, the three tried to help an American company find engineers overseas. The first customer paid $500 per week; Mercor paid about 70% of that to the engineer and kept the service fee. This small transaction validated three things: companies would pay for screening, cross-border talent had price differences, and much of the process could be software-ized.

The real turning point came when AI labs began buying large amounts of human expert feedback. Mercor moved from “recruit people for companies” to “organize professionals to help train and evaluate models,” changing the customer budget, task frequency, and network effects.

The boundary of paper results

The October 2025 Series C pushed Mercor’s valuation to $10 billion. Based on roughly 22% personal ownership, Forbes estimated the three founders’ paper wealth; that is not cash and changes with later financing, liquidation preferences, and company-valuation changes.

Media reports in 2026 used a $1 billion annualized revenue run-rate figure. That likewise does not mean Mercor had recognized $1 billion in revenue in a complete fiscal year.

My read

Brendan Foody is a market-minded founder who first found a global talent-price gap, then turned a recruiting network into an AI-training supply chain. The easiest mistake is to reduce the story to “a 22-year-old dropout became a billionaire”; the real key is that the three founders repeatedly changed their position in the transaction: from matching engineers, to automating screening, to organizing high-quality expert judgment.

The first lesson is to validate a transaction with a human service before productizing the repeated parts. Their first deal was not a grand platform, but one customer willing to pay weekly. Second, watch why customers buy instead of clinging to the initial category: AI labs did not really need more résumés, but manageable, evaluable expert capacity. Third, the roles of CEO, CTO, and chairman among three high-school friends show that young teams still need clear positions of power and responsibility.

The caution is that this result depends heavily on frontier-lab spending cycles, Silicon Valley capital networks, and extreme work intensity. Mercor’s valuation, annualized revenue run rate, and founders’ paper wealth are three different measures; none automatically proves long-term profit or a durable moat. Ordinary founders should learn to find change in real transactions, not treat dropping out or a high valuation as the cause.

Next founder