Brendan
Foody

Skip Finals. Build a Decacorn.

Spring 2023. Georgetown University's sophomore class was grinding through finals. Brendan Foody was not. He had already decided - months earlier - that he wasn't going to take them. He walked out of his dormitory, flew to San Francisco, and started building what would become one of the fastest-scaling companies in the history of venture capital.

The move didn't shock his co-founders. Adarsh Hiremath and Surya Midha - his best friends from Bellarmine College Preparatory in San Jose, his partners on the #1-ranked national debate team - had already made the same call. Hiremath from Harvard. Midha from Georgetown. Three dropout letters. One company. One São Paulo hackathon where it had all clicked into place a few months prior.

"I knew I wanted to drop out before finals my sophomore year," Foody told Fortune. "I just didn't go to finals."

What they'd figured out in São Paulo was deceptively simple: connect companies that needed skilled technical work with engineers abroad, handle the logistics, pocket a percentage of each deal. Their first client paid $500 a week for a developer. Mercor paid the engineer roughly $350 - 70% - and kept the rest. Nine months later, they had a $1 million annual run rate. Seventeen months after that, $500 million.

From Debate Rounds to AI Rubrics

Foody's childhood reads like a venture capitalist's origin story checklist. Sold donuts in 8th grade. Built websites as a high school side hustle with his co-founders. Founded Stealth, a cloud consulting business, and Seros, a startup aiming to bring cheap computing to the developing world. All before college. All before 18.

But the debate team might have mattered most. At Bellarmine College Prep - the Jesuit school in San Jose where all three founders met - Foody co-captained the squad that ranked #1 in the nation. He placed 6th nationally in individual extemporaneous speaking, a discipline that rewards rapid synthesis of evidence under pressure. If you want to understand how Mercor builds rubrics for AI training, it helps to know that its CEO once competed in timed rounds constructing arguments from scratch about current events. The cognitive infrastructure maps directly.

The Pivot That Built a Decacorn

Mercor launched as an AI-powered hiring platform - matching companies with software engineers, automating resume screening and candidate vetting. It was good. It grew fast. But the bigger opportunity was adjacent, and Foody saw it early.

The AI labs weren't just hiring engineers. They were hungry for something different - professionals who could evaluate AI outputs in specialized domains. Financial memos drafted by a Goldman analyst. Legal briefs reviewed by a Cravath attorney. Medical summaries checked by a practicing physician. These were tasks the models aspired to master, and the only way to get there was through human feedback at scale.

Mercor became the infrastructure layer for that feedback. A kind of talent exchange for the AI economy, running 24 hours a day, paying out more than $1.5 million every single day to a network of over 30,000 contractors worldwide. The average hourly rate on the platform exceeds $85 - meaningfully higher than most gig economy platforms, and intentionally so.

"Models might be superhuman at Olympiad math," Foody has noted, "but they still can't draft an email the way a senior banker would."

What $10 Billion Looks Like at 22

In October 2025, Felicis Ventures led a $350 million Series C into Mercor at a $10 billion valuation - five times the $2 billion tag from just eight months earlier. Benchmark and General Catalyst returned. Robinhood Ventures came in new. The raise made Foody, Hiremath, and Midha the youngest self-made tech billionaires in history, edging out Mark Zuckerberg who hit that threshold at 23.

When Felicis closed the deal, they sent Foody custom Mercor Air Jordan 1s. He uses a standing desk in Mercor's open-plan office. He works 9am to 10pm most days. He has not taken a single day off in three years.

He doesn't frame this as sacrifice. "People burn out when they work hard on things that don't feel compounding," he told Fortune. "I see the ROI of my time every day." Before Mercor, he says, work was something he had to discipline himself into. After it, something he couldn't switch off.

The Talent Web Around Mercor

The people Foody has assembled around Mercor tell you something about how seriously the company is taken. Sundeep Jain, former Chief Product Officer at Uber, joined as President in May 2025. Shaun VanWeelden, who ran human data operations at OpenAI, came on as Managing Director. Aaron Langerman - their former debate coach from Bellarmine - joined to help scale operations. Former debate coach. Running operations at a $10B company. Mercor's origin story runs all the way through.

The client roster is no less striking. OpenAI. Meta. Google DeepMind. Anthropic. When Mercor hired 25 Math Olympiad winners in a single day for one AI lab, the client's reaction was the kind you don't forget: "Holy shit, this is crazy." Foody didn't disagree.

Phoebe Gates - yes, Bill Gates' daughter - used Mercor to hire the team for Phia, her AI shopping startup. The platform has become, quietly, a who's-who of the AI economy's talent infrastructure.

Building Benchmarks to Prove a Point

In January 2026, Mercor launched APEX-Agents - a benchmark of 480 tasks drawn from real scenarios at Goldman Sachs, McKinsey, and Cravath. The results were clarifying: even the best AI agents barely cleared 25% on a first-try basis for tasks that a seasoned professional could complete in 2.5+ hours. The benchmark wasn't designed to embarrass the labs. It was designed to make the case for what Mercor sells: that human expertise in the loop remains non-negotiable for high-stakes knowledge work.

Mercor's APEX (AI Productivity Index) had already mapped AI performance across 200 professional tasks - financial memos, legal briefs, medical analysis - before agents were added to the equation. The data gave Foody something sharper than a pitch deck: evidence.

The Longer View

Foody is optimistic about automation to a degree that surprises some people. He thinks AI will displace roughly two-thirds of knowledge work. He thinks this is, on balance, good. His read on economic history: every major technical revolution has eventually made life better, even when the short-term disruptions were real and painful. He's aware of the Luddite critique. He just doesn't find it persuasive at the civilizational scale.

"We'll automate maybe two-thirds of knowledge work," he told Fortune. "And that'll be incredible, because it lets us do things like cure cancer and go to Mars."

He calls Mercor's work a bridge - between the economy we have and the one coming. The platform creates a new category of work for the transition period: domain experts who train the systems that will eventually outperform them in some tasks, while humans figure out what higher-order things come next. His mother's charity, meanwhile, works to revitalize downtown Menlo Park. The person building global AI infrastructure still has local roots.

Foody's former debate coach joined the company. His high school friends run it with him. The investors gave him sneakers. He still hasn't taken a day off.