The Statistician Who Bets on Founders
Here's a thing Jason Risch does every week that most venture capitalists don't: he calls Chief Information Security Officers at growth-stage companies, unprompted, just to listen. Not to pitch a portfolio company. Not to mine for deal flow. Just to hear what keeps them up at night. By the time a security startup walks into Greylock with a deck, Risch has already built a map of the problem from the people who live it.
That's the edge. At Greylock - a firm that helped shape the careers of Reid Hoffman, Peter Thiel, and countless others - Risch focuses his bets on the two categories he believes are genuinely reshaping the decade: cybersecurity and artificial intelligence. His portfolio reflects a pattern. He backed Wiz early, before cloud security had become the obvious arms race it is today. He backed LlamaIndex when most engineers were still arguing about whether RAG was a real technique. He backed Cribl before data observability was a category anyone could Google.
"There's no better feeling than seeing someone that I've built a close relationship with succeed and build a company with a global impact."
- Jason Risch, Partner at GreylockBut the CISO calls are only one thread. Before investing, Risch typically meets founders several times, in person, across months. He watches how they recruit. He listens to how their thesis evolves under pressure. He's looking for the specific signal that a leader can convince brilliant people to take a risk on them - because if they can't do that, nothing else follows.
He thinks about investing the way a statistician thinks about experimental design. "I try to shift the distribution to increase a founder's odds of success," he has said. Not guarantee the outcome. Shift the distribution. The distinction matters: it's the language of someone who understands uncertainty rather than someone pretending to eliminate it.
Marin County, the Oakland A's, and Michael Lewis
Jason Risch grew up in Marin County watching the Oakland A's, catching 150 or more of 162 games in a season. Then he read Moneyball - Michael Lewis's account of how Oakland's front office exploited statistical inefficiencies to compete against teams spending three times as much. The book didn't just explain baseball. It handed Risch a worldview.
Statistics, he decided, was a tool for finding what others missed. The A's weren't smarter than the Yankees. They just looked at different numbers - on-base percentage when everyone else chased batting average. The lesson was that the market was consistently wrong in predictable ways, and a disciplined probabilistic approach could find the gaps.
That intellectual orientation followed him to Stanford, where he studied Mathematical and Computational Science as an undergraduate (with, of all things, a Classical Studies minor) and then pursued a graduate degree in statistics. It followed him to McKinsey, where he worked in the Bay Area practice advising large companies on disruptive technology - and where he also saw, firsthand, the gap between what enterprise data management looked like in consulting frameworks and what it looked like in the real world. Messier. Much messier.
It followed him to Opendoor, where he joined during a period of rapid growth in a cross-functional role bridging management consulting and data science. And it followed him, in 2018, to AI Fund - the startup studio built by Andrew Ng, the deep learning pioneer whose courses had introduced an entire generation to neural networks. At AI Fund, Risch wasn't investing in AI companies. He was building them - from scratch, from the first prototype, hiring the founding teams before spinning them out as independent ventures.
Four Stops That Made the VC
Where the Bets Go
Investment range: $500K - $200M • Sweet spot: $25M • Fund: $1B
The Companies He Bet On
The pattern across Risch's portfolio is early conviction in founders attacking technically hard problems in categories that most investors still underweight. Security infrastructure that incumbents are too comfortable to rethink. AI tooling that developers need before most enterprises have even admitted they need it.
When Voice Authentication Breaks
Risch doesn't just invest in security companies - he thinks publicly about the problems they're solving. His 2024 essay on deepfakes is a case study in how he approaches a new attack surface: map the threat model, find the gaps in the current defensive stack, then figure out what kinds of companies can fill them.
The Deepfake Problem
Risch has documented a cascading failure in traditional authentication: as AI-generated voice and video become indistinguishable from real, the identity verification systems that financial institutions have relied on for decades start to break. OpenAI recommends eliminating voice authentication entirely. That recommendation isn't paranoia - it's an acknowledgment that a foundational assumption (your voice is uniquely yours) no longer holds.
His framework for the defensive opportunity covers three layers: detection (ensemble AI models that flag synthetic content), provenance and watermarking (technologies like C2PA that embed traceable metadata), and hygiene (behavioral analysis and protocol-based verification that doesn't rely on biometrics at all).
The cybersecurity stakes, in Risch's view, aren't just corporate. They're geopolitical. Nation-state actors with AI-powered offensive capabilities are a qualitatively different threat from script-kiddie attacks, and the infrastructure to defend against them is being built by the same kinds of early-stage startups he backs.
"Cybersecurity is a critical battleground not only for enterprises but geopolitically as well - and increasingly sophisticated AI-powered offensive cyber from nation-states makes continued defensive development all the more important."
- Jason RischHow He Actually Decides
Risch's investment philosophy is not about being first to a deck. It's about the depth of relationship before the deck exists. He tries to meet founders multiple times before a company has even launched - building a mutual information base that makes the eventual investment decision less of a bet and more of a natural continuation of an ongoing collaboration.
What he's evaluating in those early conversations: communication ability, because founders who can't articulate why they're right will struggle to recruit. Recruitment ability itself, because the quality of the team a founder can assemble in the first six months is one of the clearest leading indicators of trajectory. And the evolution of the idea - how the thesis sharpens under interrogation, how the founder incorporates new information and adjusts.
Risch describes his decision-making as explicitly Bayesian: "incorporates new information and adjusts prior beliefs." Each founder meeting updates the prior. The weekly CISO calls update the market map. The investment thesis isn't a fixed document - it's a probability distribution that shifts as evidence accumulates.
That same approach shapes how he supports companies post-investment. He's not parachuting in with advice; he's building the same kind of relationship with founders that he started before they had a product, sustained through the unglamorous middle of a startup's life, and present when the company becomes something that matters.
On and Off the Cap Table
He married Victoria, who was also his classmate at Stanford. He grew up in Marin County watching the Oakland A's - not 80 games, not even 120. More than 150 out of 162, in a year. He named his Twitter handle @rischter_scale, a pun on his surname and the Richter seismic scale, because apparently when you study math and classics at Stanford you develop a very specific sense of humor about tectonic metaphors.
He reads Neal Stephenson - Snow Crash and Cryptonomicon, both of which are essentially user manuals for the security-and-cryptography-and-digital-money future Risch is now actively funding. He reads Ray Dalio and Ian Morris for systems thinking. His Classical Studies minor sits in quiet, productive tension with his mathematical background: the same person who models probability distributions read Thucydides for fun.
He came to San Francisco by way of South Carolina (birth) and Marin County (upbringing), but he's as Bay Area as anyone who went to Stanford for seven years and then built three careers in the startup ecosystem before landing at one of the valley's oldest firms.