The Twenty Minute VC (20VC): Mercor – From $1M to $500M in 17 Months with Brendan Foody

Host: Harry Stebbings
Guest: Brendan Foody, Co-founder & CEO of Mercor
Release date: September 15, 2025

Episode Overview

In this high-energy episode, Harry Stebbings sits down with Brendan Foody, the 22-year-old co-founder and CEO of Mercor, the AI labor marketplace that has vaulted from $1M to $500M in revenue run rate in just 17 months—making it the fastest-growing company in history. They dive deep into Mercor's growth story, the shifting dynamics in AI data and model training, business strategy, investor perspectives, margins, the future of the AI labor market, and much more. Foody shares candid insights on leadership, industry competition, and what it takes to build a generational company in the AI era.

Key Discussion Points and Insights

Brendan’s Entrepreneurial Origins & Early Hustles

• [05:01] Early Side Hustles:
  Foody describes selling donuts at school as his introduction to entrepreneurship, scaling the side business, encountering competition, and creatively skirting school regulations.

  • “I would buy Safeway donuts, bike to my middle school and sell them for $2 each... My principal called me into their office to try to shut down my donut stand... I moved my donut stand 20 feet over off campus so that they couldn’t police me.” (Brendan Foody, 05:01)
  • His mother’s worry: scaling hustles might lead to riskier pursuits prompted her to send him to Catholic high school for “the straight and narrow.”

• [07:24] Ambition and Founder Psychology:

  • Foody admits to always having grand ambitions, but never anticipated the meteoric success Mercor would achieve, highlighting founder duality—“superiority and inferiority complexes.”

College, Self-directed Learning & Advice to Young People

• [08:03] College Skepticism:

  • Foody recounts substantial income from high school side businesses (sneaker resale and AWS credits consulting). Debated going to college, ultimately appeased his parents, but applied last minute.

• [09:18] Educational Value Shift:

  • Emphasizes most learning now happens online; college is primarily valuable for social growth rather than education:
    • “I listened to almost every Stanford GSB lecture when I was in high school… AI only exaggerates that.” (Brendan Foody, 09:18)
  • Harry shares his own university dropout story, illustrating the shifting opportunity calculus for young entrepreneurs.

Mercor’s Market and Business Model Evolution

• [10:43] AI Labor Market is Not a ‘Body Shop’:

  • Mercor’s unique value: matches top-tier, specialized talent (Goldman Sachs, McKinsey, FAANG engineers) with AI training needs, going far beyond generic crowdsourced labor.

• [12:44] Human Expertise Remains Critical:

  • As models improve, the supply side narrows to the most elite domain experts, but model complexity introduces new frontiers:
    • “So long as there’s things the human is able to do the model’s not able to do… we need humans that help to create those verifiers.” (Brendan Foody, 13:44)

• [14:22] Scaling Laws & Model Improvement:

  • Disputes idea of a plateau in model progress; sees shift from quantity to quality and high-complexity RL (reinforcement learning) environments as the key driver.

Competitive Landscape and Differentiation

• [15:15] Supply Side Differentiation:

  • Mercor wins by focusing on outcomes driven by the top 10-20% of contributors—true power-law dynamics.
  • “When we’re able to find those people that are the 10x contributors, it’s very difficult to recreate.” (Brendan Foody, 16:15)

• [16:48] Measuring Data Quality:

  • Pushes back on claims competitors have no quality algorithms. Mercor deeply invests in assessing and improving data quality via research partnerships.

• [17:45] Customer Concentration and Vendor Dynamics:

  • Labs may multi-vendor initially, but gravitate to partners who drive best results, leading to revenue concentration akin to Nvidia’s dynamic.

Mercor’s Hyper-Growth Story

• [20:57] Explosion after Scale AI Acquisition:

  • Scale AI’s acquisition led to a step-change in demand.
  • “We scaled the business from $1 to $500 million in revenue run rate in the last 17 months, which is the fastest revenue growth of all time. One month faster than Cursor’s time.” (Brendan Foody, 21:13)

• [22:32] Why Scale AI Stumbled:

  • Scale lost focus on product and quality.
  • Mercor’s market advantage: average pay rate of $95/hr vs. Scale’s $30/hr, leading to access to elite talent and better outcomes.

Human vs. Synthetic Data & The Limits of AI

• [23:59] Synthetic Data vs. Human Data:

  • Synthetic data’s role is growing but limited; ultimate frontier still needs elite human “stasis points”:
    • “The total addressable market is bound by the amount of things that humans are better at than models.” (Brendan Foody, 23:59)

• [25:56] AI Evaluation Benchmarks are Broken:

  • Current academic evals often fail to reflect real-world utility.
  • “One of the largest inefficiencies in all of AI research is that the evals … are wholly disconnected from the outcomes that consumers and enterprises actually care about.” (Brendan Foody, 25:56)

• [26:56] The Future of Evals:

  • Calls for evals that truly mirror workplace tasks, e.g., “Create a rubric, similar to how a professor would grade an essay, for grading model performance on job-related tasks.” (Brendan Foody, 26:56)

Founder Lessons: Capital, Growth, and Going Public

• [28:12–30:33] On Valuation & Investor Hype:

  • Mercor has raised at 100x+ revenue multiples, justified by outlier growth and “phenomenal experiences” for customers and talent.
  • Intentionally capital efficient, but open to strategic financings for signaling market leadership.

• [32:45] Staying Private vs. Going Public:

  • Taking advice from Jack Dorsey to remain private as long as possible to keep long-term focus:
    • “Public companies… get more caught up in the quarterly numbers and aren’t as focused as they maybe should be on all the long-term drivers of value.” (Brendan Foody, 33:09)

• [34:01] Private Markets, Bubbles, and Long-Term Value:

  • Too much VC capital funds subpar competitors; true value will be clear over 10 years, not in today’s exuberant cycles.

Revenue, Margins, and Sustainability in the AI Boom

• [36:01] Revenue Durability & Retention:

  • Extreme pilot churn is a red flag; strong retention and customer love signals durable value.

• [36:55] Margins Still Matter:

  • While Mercor maintains strong gross/net margins, many peers chase growth at the expense of margin—dangerous in competitive, low-switching-cost markets.

• [37:49] Capex and the Supercycle:

  • Capex is less a concern if you’re investing on a 10-year horizon and carefully pick the right bets.

AI Models, Market Structure, and Talent Economics

• [40:53] Big Model Providers & Industry Structure:

  • Most big model players are likely already established due to massive capex required, but still room for innovation and startup breakthroughs.

• [41:42] Talent Costs & Motivation:

  • Talent is more expensive than ever; high pay matters, but so does mission and upside for “missionaries, not mercenaries.”
  • Startups can’t match $100M in cash, but equity appreciation and purpose provide compelling value for top contributors.

• [43:18] Undervalued Teams:

  • DeepMind (Gemini Flash models) highlighted as an underrated player, especially in evals and smaller, efficient models.

• [44:03] Many Models or Monoliths?:

  • Shift from belief in model specialization to a future of both large general-purpose and specialized models, as headroom remains for foundational model generalization.

• [45:07] Sovereignty in Model Markets:

  • Partial value in geography-specific models (e.g., Mistral in EU law), but expects global generalist models to win at scale.

Building Company Culture in Hyper-growth

• [46:11] 996 Work Culture & Execution:

  • Early Mercor operated at extreme intensity, but never mandated hours. Now moving to “output” over face-time, especially as execs join.

• [47:57] Mission-Driven Culture:

  • Focus on hiring “people that love what they do and are obsessed… rather than specific hours.”

Leadership, Risk, and Personal Growth

• [48:43] Taking Risks:

  • Foody admits Mercor has run conservatively (capital effectiveness) to ensure durability; sometimes wonders if they should “start burning hundreds of millions a year.”

• [51:29] Demand Outstrips Supply:

  • Mercor turns down projects daily; supply is main constraint, and expanding capacity is the focus.

• [52:09] Advice on Founder Liquidity:

  • Foody advises taking only minimal secondary to avoid distraction, signaling commitment long-term.

Notable Quotes & Memorable Moments

• On Early Hustles:
  “I moved my donut stand 20 feet over off campus so that they couldn’t police me.” (Brendan Foody, 05:01)

• On Current AI Evals:
  “One of the largest inefficiencies in all of AI research is that the evals … are wholly disconnected from the outcomes that consumers and enterprises actually care about.” (25:56)

• On Talent Market:
  “Our average marketplace pay rate is $95 an hour, to put that in frame of reference, whereas scale and search generally pay about $30 an hour.” (22:32)

• On Efficiency and Growth:
  “Too many people think about valuation through the lens of market comps and revenue multiples and not enough through the lens of what's possible with this company.” (Brendan Foody, 28:29)

• On Private vs. Public:
  “It allows you to stay very long term oriented... public companies still get more caught up in the quarterly numbers.” (33:09)

• On Sustaining Company Culture:
  “It's been much more about hiring people that like give a shit and love what they do and are obsessed with it.” (47:57)

Key Timestamps for Important Segments

• [05:01] – Early entrepreneurial stories and lessons from selling donuts
• [09:18] – College, self-directed learning, advice to the next generation
• [10:43] – Mercor’s market, why it’s not a “body shop”
• [12:44] – The role of elite human contributors as AI models climb in complexity
• [14:22] – Shifts in AI model improvement: scaling laws, data quality
• [15:15] – Differentiating on supply side, value of “10x contributors”
• [16:48] – Data quality measurement and being a research partner
• [17:45] – Vendor concentration, revenue mixes, parallels to Nvidia
• [20:57] – Mercor’s hyper-growth, effect of Scale AI acquisition
• [23:59] – Synthetic data vs. human data, limits of model self-improvement
• [25:56] – Broken state of AI evals and what’s needed for real-world benchmarking
• [28:12] – Leadership at high growth, personal change as a young founder
• [32:45] – Pros and cons of staying private vs. IPO
• [36:01] – Revenue sustainability, margins, and the durability of AI businesses
• [40:03] – The elasticity of engineering roles in the age of AI
• [44:03] – The future industry structure: monolithic vs. specialized models
• [46:11] – Evolving the work culture, from 996 to output-driven
• [51:29] – Supply constraints, doubling demand, and focus on quality over growth for its own sake

Closing Quick-Fire Round Highlights

• AI Myth:
  “We’ll have superintelligence in three years that’s better than humans at everything—I think it’s totally wrong.” (Brendan Foody, 52:48)

• CEO for a Day:
  “Model customization is a really exciting opportunity… API will have low switching costs, not much pricing power. It’s not a good business.” (53:00)

• Most Admired Investor:
  “Jeff Bezos… I’ve admired Amazon so much and just the early clarity of thought… I’d love to learn from him.” (54:51)

• Best Retroactive Advice to Self:
  “Focus on Foundation Model Labs… If I’d realized that nine or 12 months sooner, that would have been even more exciting.” (55:23)

Final Takeaway

This episode is a masterclass in navigating explosive growth in the AI era, balancing capital discipline with ambition, and finding competitive edges in the high-stakes labor market for AI development. Foody’s candor on leadership, model improvement, and business fundamentals offers rare — and hard-earned — wisdom for founders, operators, and investors alike.

For more episodes and information: www.20vc.com