Uncapped with Jack AltmanHow To Make Money in Venture | Josh Kopelman, Co-Founder of First Round Capital | Ep. 8
CHAPTERS
- 0:00 – 0:25
Venture cycles and why bubbles still make the best money
Josh opens with a counterintuitive lesson from the dot-com era: even when everyone knows it’s a bubble, exiting early can destroy the majority of your eventual profits. This sets up the episode’s recurring theme that venture returns are lumpy, time-dependent, and often driven by extreme market conditions.
- •Dot-com bubble was widely acknowledged, yet staying in captured most of the upside
- •Exiting ‘prudently’ can mean forfeiting the bulk of profits
- •Venture outcomes hinge on timing as much as selection
- •Sets framing for later discussion on harvest cycles and disequilibrium
- 0:25 – 2:27
Today’s venture landscape: exploding fund count and the LP-side shift
Josh contrasts 2004’s venture ecosystem with today’s: an order-of-magnitude increase in funds and active check writers. He argues the more under-discussed change is on the LP side—new pools of capital with different return targets reshaping what “success” looks like.
- •~850 funds in 2004 vs 10,000+ funds today; far more active check writers
- •More competition for VCs, but generally more options/capital for founders
- •Original LP bargain: accept illiquidity for high IRR/outperformance
- •New LP entrants (e.g., sovereign wealth) may accept lower IRR for scale exposure
- 2:27 – 4:47
Blackstone-ification: scale vs alpha, and the bear case risk
Josh describes the “bull case” where venture becomes more like large-scale alternative asset managers: lower but consistent returns at massive AUM can produce more total profit dollars. The “bear case” is that returns compress too far—toward zero or negative—especially given venture’s tricky math and long duration.
- •Scale model: 12% on $100B can beat 30% on $1B in absolute dollars
- •Lower returns may be ‘fine’ if capital’s cost/expectations are different
- •Bear case: instead of 12–14% IRR, outcomes could be 0–3% or worse
- •Average/median venture outcomes can be poor; power laws dominate
- 4:47 – 8:26
The Venture Arrogance Score: sizing a fund by required exit capture
Josh introduces his “Venture Arrogance Score,” a simple model using fund size and expected ownership at exit to compute how much total market exit value a firm must capture. The exercise highlights how difficult it is for mega-funds to achieve traditional venture multiples without assuming implausible market share of total exits or a huge increase in overall exit value.
- •Two inputs define the model: fund size and % ownership at exit
- •Example: $7B fund at 10% ownership implies $70B exit value per 1x fund return
- •To hit 4x gross, required exit value scales to ~$280B for the fund’s basket
- •Compared to historical annual exit value, implies capturing an unrealistic market share
- 8:26 – 10:49
Does the whole asset class lose money? Power laws, labels, and mixed strategies
Josh and Jack discuss venture as an asset class where the aggregate basket can underperform, making manager selection crucial. Josh points out that the industry lacks clear labeling between “alpha-seeking bespoke venture” and “AUM-driven cash-on-cash” approaches, even though they behave like different products with different success metrics.
- •Median fund historically often fails to return capital; outliers drive returns
- •Buying the entire venture ‘basket’ can be negative in aggregate
- •Venture vs private equity risk/controls differ; even venture has sub-strata (seed/late)
- •Industry doesn’t clearly distinguish alpha-first vs scale/AUM-first venture models
- 10:49 – 15:38
Big private companies and longer holding periods: the duration/IRR trap
Jack offers a counterargument: huge private companies staying private longer can absorb large checks and still produce big multiples. Josh agrees in principle but demonstrates how longer duration crushes IRR—even if headline multiples (e.g., 4x) look identical—making the long-private trend a meaningful risk for LP outcomes.
- •More mega-private outcomes could support large fund deployment
- •Duration matters: same multiple over longer time dramatically reduces IRR
- •Illustration: 4x over 10 years (~27.5% IRR) vs 4x over 18 years (~11.5% IRR)
- •A 2x outcome over very long duration can become unattractive vs safer alternatives
- 15:38 – 18:19
Relevance in venture: activity begets activity (and the Steph Curry analogy)
The conversation turns to how “relevance” creates access in private markets. Josh argues venture has a ‘hot hand’ dynamic where frequent investing can build reputation and deal flow regardless of underlying hit rate in the short term, because unlike sports, you can get credit just for taking the shot.
- •Private markets are access-driven; founders choose the most active, connected investors
- •Relevance often compounds via referrals and visibility (Matthew effect)
- •Venture differs from sports: activity can be rewarded even before outcomes are known
- •Reputation can be created/destroyed quickly via the industry ‘gossip mill’
- 18:19 – 21:03
Relevance vs returns over time: the two-by-two grid and who endures
Josh outlines that relevance without returns doesn’t last, citing time periods when certain fast-moving funds were extremely relevant. He notes some high-performing firms stayed under the radar by focusing on investing quality over hype, and argues enduring franchises require sufficient returns, not just momentary relevance.
- •2018–2021: founders frequently wanted Tiger/SoftBank intros due to relevance
- •Long-term durability needs returns; hype without substance fades
- •Examples of quieter, strong performers (e.g., Founder Collective, IA Ventures, Altos)
- •Endurance comes from ‘checks deposited’ (real outcomes), not just activity
- 21:03 – 26:36
First Round’s positioning: small-fund alignment, ownership targets, and portfolio construction
Josh explains how First Round competes amid mega-funds by emphasizing alignment (carry over fees), meaningful early ownership, and focused value-add in the first 24–36 months. He also breaks down their portfolio sizing logic, emphasizing time diversification, expected entry ownership, and the mortality curve from round to round.
- •Alignment: smaller fees; partners’ economics depend heavily on carry
- •Strategy: be the best partner in the first 2–3 years (PMF, team, pricing, culture)
- •Portfolio approach: ~70–80 seed investments, deployed over ~2.5–3 years
- •Target ownership: ~12.5–15% at entry; aim for ~8–9% at exit via pro rata
- 26:36 – 32:12
Hyper-harvest and disequilibrium: where venture profits are actually made
Josh argues venture doesn’t make money in normal markets—it makes money in extreme greed periods when multiples expand dramatically. He shares data showing profit concentration into short windows (both historically and at First Round), and describes the discipline required to hold through ‘obvious’ bubbles to capture outsized returns.
- •Returns concentrate in short windows; example: 83% of profits in last 3 years of 1980–2000
- •First Round: 90%+ of returns in a ~36-month period across 20 years
- •Profits come from irrational disequilibrium, not equilibrium or mild cycles
- •The hardest discipline: not selling too early even when froth is obvious
- 32:12 – 33:56
When the market is frothy: secondaries, partial de-risking, and founder alignment
Jack asks what to do when the next hyper-harvest arrives. Josh recommends avoiding top-ticking, using tenders/secondaries to take partial liquidity (e.g., 1–2x fund return) while staying long, and explicitly sharing the math with founders so decisions reflect both timing and personal risk profiles.
- •Don’t try to perfectly top-tick; consider partial liquidity instead
- •Tenders/secondaries can lock in outcomes and improve IRR by shortening payback
- •Founder alignment: don’t compete with the company; be transparent about incentives
- •Secondary can reduce founder pressure to sell early for personal liquidity
- 33:56 – 40:52
‘Software is eating the world’—but the missing assumption was margin superiority
Josh reframes the Andreessen essay as transformative for venture’s aperture, making many non-traditional categories venture-fundable. But he argues the industry implicitly assumed software would bring margin superiority; in many sectors that hasn’t materialized, leading to multiple compression back toward traditional industry valuations.
- •Essay changed venture framing: expanded what VCs considered fundable
- •Core implicit premise: software winners would enjoy superior margins (and thus multiples)
- •Many ‘software-enabled’ sectors didn’t achieve margin superiority; valuations reverted
- •Result: 2020–2021 vintages include companies priced for margins they didn’t realize
- 40:52 – 49:01
AI’s promise vs the underwriting question: people-and-problems over themes
Josh is bullish on AI’s societal impact (science, healthcare, energy, education), but notes many startups are essentially productivity plays against labor categories. He explains First Round doesn’t try to theme-invest early; instead they focus on founders and problems, assuming early solutions are often wrong and themes become obvious too late for seed.
- •AI likely enables major scientific and societal advances
- •Many AI startups target automating or augmenting specific job categories
- •First Round’s seed strategy: back founder + problem, not a fully formed solution
- •Themes often become clear too late for early-stage advantage
- 49:01 – 56:46
Operating like a company: Brett’s role, decision-making as product, and the ‘game tape’
Josh explains why First Round invests in an operating model uncommon in venture: running the firm like a company with product development, experiments, and strong management leadership (Brett’s role). He details their structured decision process—including a 36-question rubric—to capture reasoning, reduce bias, and build a searchable learning system over time.
- •Most VC firms are ‘poorly run’ operationally; First Round tries to operate like a company
- •Brett functions like a CEO/operator partner driving products, experiments, and process
- •Decision-making is treated as IP: capture, analyze, and learn from meetings
- •36-question pre-discussion rubric creates better debates and long-term learning archives
- 56:46 – 58:55
Future of First Round: no ‘Jiro’ endpoint—build an adaptive machine
In closing, Josh rejects the idea of a static, perfected model; he believes venture demands continuous reinvention as technology and markets change. The aspiration is to build an iterative organization—culture, processes, and experimentation—that can keep discovering what works across future cycles, including AI-driven changes to how the firm operates.
- •Complacency is dangerous; what ‘good’ looks like will change over time
- •First Round aims to evolve continuously rather than repeat past playbooks
- •AI is already transforming internal workflows and will keep changing the job
- •Goal: an adaptive machine that creates future ‘hits,’ not an ‘80s band’ replaying old songs
