20VC: Inside Sequoia's Investment Committee: Lessons from Don Valentine, Doug Leone and Alfred Lin | How the SpaceX and Citadel Deals Went Down | What Sequoia Specifically Looks for in Founders with Julien Bek

What It Covers

Sequoia partner Julien Bek gives a behind-the-scenes account of how Sequoia's investment committee operates, how deals like SpaceX and Citadel were sourced, how the firm reads founders across cultures and backgrounds, and why agents becoming the primary customer represents the next structural shift in software economics.

Key Questions Answered

• •IC Conviction Over Consensus: When SpaceX was first presented to Sequoia's IC, at least one partner voted it a "1" — the lowest possible score. The sponsor pressed forward anyway, organized a partnership site visit, secured a smaller initial check, and built it into one of the firm's best-ever investments. Historical fund data consistently shows the highest-returning deals correlate with the sponsor's highest personal conviction, not with consensus votes.
• •Founder Reading via Vulnerability: To surface authentic founder signal in 30-minute meetings, Bek opens by sharing personal hardship — parental illness, financial instability, split households — before asking about the company. This lowers defenses and prompts founders to reveal character rather than rehearsed narrative. The technique bypasses polished pitches and exposes how founders handle ambiguity, pressure, and self-awareness under conditions that feel conversational rather than evaluative.
• •Worst Reference as Diagnostic Tool: Doug Leone's interview framework asks candidates first who their best reference is, then immediately follows with who their worst reference would be and why. The second question reveals self-awareness, honesty, and blind spots. Bek extends this to founder diligence: calling the named worst reference often yields more texture and useful signal than any positive reference, regardless of what that reference actually says.
• •Distance Traveled Over Pattern Matching: Rather than filtering for Ivy League credentials or specific childhood trauma, Bek maps the trajectory a founder has traveled relative to their starting point. Two Polytechnique graduates can represent radically different levels of drive depending on origin. The operative question is whether that trajectory will continue, not whether the founder fits a predetermined profile. Early life history — before professional experience — provides the richest data on slope.
• •Services-to-Software as the Next Trillion-Dollar Model: The 1-to-6 ratio between software spend and services spend in categories like accounting, legal, and customer support represents the addressable expansion. Companies that sell outcomes — closed books, resolved tickets — rather than tools capture the $6, not just the $1. Customer support has already crossed into this "autopilot" phase, with over $1 billion ARR in outcome-based AI contracts. The prerequisite is reaching human parity in task completion.
• •Agents as the New Customer Segment: Cloudflare data suggests agent-to-human traffic is already at parity and projects a 1,000x agent traffic advantage within five years. Agents carry pre-training and post-training biases — defaulting to Cloudflare and Vercel for hosting — making brand positioning in agent decision-making as consequential as SEO was for human search. Founders should optimize for "answer engine optimization" (AEO) and treat agent preferences as a distinct, measurable customer acquisition channel.