The Twenty Minute VCBenchmark's GP, Everett Randle on Why Mega Funds Will Not Produce Good Returns
CHAPTERS
- 0:00 – 2:03
Everett Randle joins Benchmark: what this conversation will really focus on
Harry introduces Everett Randle as Benchmark’s newest partner and frames the episode around how venture changes in an AI-dominated market. The cold open tees up two through-lines: new AI investing frameworks and why mega-funds struggle to generate great net multiples.
- •Benchmark’s newest GP and why his perspective matters now
- •The episode’s core tension: AI-era frameworks vs legacy VC playbooks
- •Teaser: mega-funds, Tiger’s legacy, and Benchmark’s biggest threat
- 2:03 – 3:35
Lessons from Mary Meeker: using numbers to tell an 8–10 year story
Everett argues Mary Meeker is misunderstood as purely quantitative; he sees her as deeply qualitative through numbers. He explains how modeling should help investors visualize the future state of a business, not trap them in spreadsheet logic.
- •Mary’s “read the matrix” approach: numbers as narrative
- •Long-horizon visualization (e.g., household penetration vs growth rates)
- •Quant frameworks are useful when they sharpen story, not replace it
- 3:35 – 6:27
Lessons from Peter Thiel: organizational design as a conviction machine
Everett describes how Thiel builds firms with incentive structures that constantly test investor conviction. The standout example is personal co-investing alongside the firm, turning “pro rata” decisions into a real integrity and conviction check.
- •Firm design matters as much as picking companies
- •Personal co-invest programs as a conviction litmus test
- •Scrappiness expectation: find a way even without liquidity
- •Culture of incentives over explicit rules
- 6:27 – 8:23
Inside Founders Fund: intense truth-seeking and flat hierarchy
He demystifies Founders Fund’s ‘black box’ by describing the internal culture: confrontational debates that are possible because relationships are strong. The goal is brutal truth-seeking rather than political deference.
- •IC intensity interpreted as sibling-style honesty
- •Psychological safety enables direct disagreement
- •Flat culture prioritizes truth over hierarchy
- •How that environment shapes decision quality
- 8:23 – 11:35
Lessons from Mamoon Hamid: seeing excellence early and developing taste
Everett credits Mamoon with teaching that spotting excellence requires seeing it up close early in your career. He also highlights Mamoon’s refined taste for ‘consumer-like’ B2B products with high user love and strong founder/product/market pattern recognition.
- •“See excellence up close” to calibrate your bar
- •Boardroom exposure as training for pattern recognition
- •Taste built around people + product + market fit
- •Common thread in major wins (love/engagement in B2B)
- 11:35 – 14:06
A personal miss: passing on OpenAI at $32B and learning to trust intuition
Everett recounts how deal structure and dilution fears spooked him out of the OpenAI round—despite strong belief in the product. The lesson: sometimes the product trajectory overwhelms ‘sensible’ concerns, and investors must trust intuition when a platform shift is obvious.
- •Private equity instincts can create venture blind spots
- •Structure/dilution risks were real but ultimately secondary
- •OpenAI growth/product utility outweighed all other factors
- •Learning to trust intuition (Josh Kushner as an example)
- 14:06 – 20:21
OpenAI vs Anthropic vs the coding stack: what wins and what happens to Cursor
Everett predicts OpenAI could be a trillion-dollar company soon and compares investing in OpenAI vs Anthropic at their respective valuations. He then zooms into coding tools, arguing the market is exploding so quickly that multiple winners (including Cursor) can thrive even as share fragments.
- •OpenAI’s consumer gravity (ChatGPT) as a ‘locked in’ asset
- •Anthropic’s relative edge in some B2B commercialization and coding
- •Coding as an exploding category: market expansion beats share loss
- •“Golden categories” and why code generation may add $4–5B net new ARR/year
- 20:21 – 26:08
Why AI changes category economics: the need for a new taxonomy beyond SaaS
Everett argues AI app companies are being misjudged by SaaS metrics. He proposes shifting attention from gross margins to terminal margin structure, gross profit dollars per customer, and gross-profit-based multiples—because AI can capture labor budgets and create much larger contract values.
- •AI apps ≠ SaaS: inference in COGS changes the model
- •High margins can mean low usage (a red flag)
- •Focus on gross profit dollars per customer, not margin %
- •AWS as analogy: lower margins, massive dollars and strategic importance
- 26:08 – 28:04
AI inference cloud: changing his mind on ‘commodity’ businesses like CoreWeave
He explains why he reversed his skepticism toward AI inference cloud providers, which initially looked like low-margin compute resellers. With extreme demand curves and rapid scaling, he argues momentum and market pull can dominate traditional ‘business quality’ critiques—at least for a period.
- •Compute brokering can still compound at massive scale
- •Demand for inference resembles (or exceeds) early hyperscalers
- •Sometimes ‘shut your mind up’ and respect momentum
- •Public-market validation changed the perceived risk/reward
- 28:04 – 32:18
The labs are the new baseline: why they threaten AI app companies, and what moats are now
Everett warns that foundation model labs increasingly ship applications, setting a default customer experience that app-layer startups must beat. He argues moats haven’t fundamentally changed (seven powers still apply), but the pace and quality bar are higher; he also rejects the idea that distribution replaces technology as the moat.
- •Labs compete at the ‘base layer’ of UX and pricing ($20–$200/user)
- •App companies must deliver differentiated workflows beyond lab apps
- •Jasper as an early example of easy-come/easy-go without scaffolding
- •Moats still technological—especially talent and product-building know-how
- 32:18 – 37:37
Did Benchmark miss AI? Fund size, Conway’s Law, and why small funds can win
Everett argues venture firms ‘ship their fund size’: mega-funds must chase mega-rounds, while Benchmark can target higher cash-on-cash multiples. He claims Benchmark’s recent top holdings already show multiples that late-stage OpenAI rounds can’t match, reinforcing the small-fund advantage in net MOIC.
- •Conway’s Law applied to VC: structure dictates strategy
- •Mega-funds need mega-rounds to deploy capital efficiently
- •Benchmark’s edge: smaller fund, higher potential MOIC
- •Relevance vs returns: choosing the game you can win
- 37:37 – 41:15
Ownership discipline in an AI era: optimizing for impact and MOIC, not a fixed %
The discussion turns to Benchmark’s approach to ownership as round sizes grow and targets drift downward. Everett emphasizes Benchmark’s two north stars—being the most meaningful partner and delivering top cash-on-cash returns—arguing that rigid ownership heuristics confuse inputs with outputs.
- •Benchmark’s north stars: founder partnership + LP money-on-money
- •High ownership is not the goal; impact and returns are
- •AI outcomes are larger, enabling great funds without 20% stakes
- •Founders rarely regret ‘too much’ Benchmark ownership due to value-add
- 41:15 – 55:48
Benchmark and governance: the ‘fires founders’ critique vs fiduciary responsibility
Everett addresses public critiques (notably from Delian) about Benchmark being quick to replace founders. He argues governance norms have changed over decades, but boards still have ethical and fiduciary duties—especially when laws or ethics are breached—and that serious founders want real sparring partners, not sycophants.
- •Media narratives flatten nuanced board/founder situations
- •Historical norm: professional CEO searches were common post-investment
- •Governance obligations: protect shareholders/employees, not just NPS
- •The best founders want challenge and adult supervision in the room
- 55:48 – 57:21
People, product, market: ranking priorities—and why market is the most fungible
Everett ranks people first, product second, market third, arguing that great people and product capability are less ‘fixable’ than market selection. He notes that many iconic companies pivoted, making market the most malleable variable early on.
- •People are the upstream driver of everything else
- •Product is evidence of team quality and taste
- •Market matters, but can change through pivots early on
- •Inability to ‘upgrade’ mediocre teams makes people paramount
- 57:21 – 1:06:18
Why mega-funds won’t deliver great net returns: the post-Tiger world and capital velocity
Everett explains his 2021 ‘Playing Different Games’ thesis: Tiger popularized investment velocity as a core strategy, and many firms moved toward similar behavior. He argues mega-funds can make huge absolute dollars, but struggle to credibly promise venture-style net multiples (e.g., 5x net) due to physics of fund size and incentives.
- •Venture bifurcation: high-velocity vs high-touch craft
- •Tiger ‘died’ but spawned many Tiger-like strategies
- •Incentives cascade: juniors are promoted on deployment
- •Absolute gains ≠ venture-like net multiples for LPs at $8–$10B+ scale
- 1:06:18 – 1:26:42
Mega-fund life, first-deal pressure, and Benchmark’s biggest long-term risk: stasis
In the closing stretch, Everett critiques mega-fund career dynamics (narrow coverage, promotion pressure, lottery-like outcomes). He shares advice that a first deal failing can be liberating, then names stasis—failure to evolve while staying true to north stars—as Benchmark’s biggest threat in the next era.
- •Mega-funds can feel like banking/PE: coverage slices + promotion by deals
- •First deal psychology: failure can reduce pressure and improve rhythm
- •Benchmark’s threat: complacency and losing access to best founders
- •Need to stay dynamic without abandoning core principles
