Finding the Next Figma, Wiz, & Stripe Before It's Obvious | Neil Mehta Interview

ELI5/TLDR

Neil Mehta runs Green Oaks, a 55-company, $15B AUM firm that ignores the industry’s drift into matrix-style coverage investing. He thinks 10–15 founders a year matter, so his small team stalks them, writes $500M–$1B checks when asked, and stays for 15 years. His test for a great company: ask everyone inside whether their best days are ahead or behind. Everything else — JDCE, moats, growth persistence — flows from that single signal.

The Full Story

The one question that matters

Mehta says if you held a gun to his head and made him pick one thing to evaluate a company on, it wouldn’t be growth rate, margins, moat, or NPS. It’s the answer employees give to: are your best days ahead of you or behind you?

“If you were running a company, that’s the one stat you should care about more than anything else in the world.”

He extends the lens to Europe — a continent where, in his read, most people have believed for forty years that the best years are behind them. He doesn’t say it unkindly. He says it like a diagnosis.

JDCE — the tattooed acronym

Jaw-Dropping Customer Experience. The phrase was born watching Bom Kim build Coupang in Korea. The Green Oaks team would play focus-group tapes and watch mothers cry because diapers arrived at their door in the morning instead of being hauled back from a store. That is not an NPS score of 9. That is someone begging you not to take the product away.

A JDCE requires breaking trade-offs — doing something technically or operationally impossible, across a long enough timeline that competitors give up. Coupang built warehouse management software, routing algorithms, delivery camps, localized distribution points, packaging optimized for low waste, and the decision of where exactly to leave a box at 6am so the customer isn’t woken but the package stays safe. Retention went from market average in the 30s to Coupang’s 60s.

“Capitalism is basically full of a sea of businesses that are not really doing anything that difficult. They’re just kind of me-too products swimming in the river of beta.”

The laws of great businesses don’t change

Mehta keeps coming back to a Wright Brothers analogy. Everyone chasing flight stared at birds flapping their wings. The Wrights realized the laws of aerodynamics were the laws of aerodynamics — you don’t beat them, you work with them on a fixed wing.

“The laws of great businesses are the laws of great businesses. It is not complicated. It is the discipline of only looking for those types of businesses and those types of founders.”

Every time Green Oaks has screwed up, he says, it’s because they let someone convince them the laws were different this time. Crypto was the big one — “no, no, you don’t need a board, you don’t need an auditor, you don’t need to delight customers.” The laws are never different.

Concentration over coverage

The rest of the industry, he argues, has become a matrix — sectors on one axis, geographies or stages on the other, junior people chasing every square to brag about “92% coverage of Series Bs.” He thinks this is correct for 90% of the industry. The returns were too high; more capital should compress them.

But for the 10–15 best founders each year, coverage is precisely the wrong posture. Those founders want speed, fidelity, differentiated insight, and a partner, not a junior associate rolling up to a Series A. Green Oaks does 10–12 investments per fund with 55 active portfolio companies across $15B AUM. Mehta says he can talk to every founder he backs in half a day.

A corollary: he takes the first meeting almost every time. The first meeting should feel to the founder like the fifth or sixth, because the prep has been absurd.

The Coupang case study

Green Oaks led 5 of 8 rounds. Invested “a little under a billion” across 10 years. Joined the board at seed stage and has been on it for 15 years. Was buying more shares last quarter.

Bom Kim’s calendar on a Sunday night would have Monday through Friday blocked out on a single problem — say, negotiating diaper COGS in Q2 2014 — for five hours a day, with everything else allowed to burn.

“Focus means saying no to everything else, everything else at the cost of doing what the single most important thing is.”

Mehta talked to Bom more than anyone other than his wife and his partner Benny. 2am, 3am, 4am — Bom’s cycle time was 24 hours.

Unreasonable growth is a feature, not a bug

When Hamas attacked Israel, Wiz had a meaningful share of its go-to-market and engineering team drafted. Green Oaks told them to forget Q4, don’t even report numbers. Wiz had one of its best quarters.

“Having unreasonable expectations is a competitive advantage.”

Coupang had a year where growth dropped to 18%. Advisors told Bom, “that’s fine, Amazon never grew more than 30, you’re good.” Bom refused and later said convincing the whole organization to become a high-growth culture again was one of the hardest things he’d ever done.

Green Oaks is happy with volatility — most of their 5+ year holdings have been through patches where the fundamental premise of the business was being questioned. The trick is knowing when questioning is right (sometimes it is) versus when you can see the other side.

The Masa defense and the Yuri case for GOAT

Asked who the greatest investor of all time is, Mehta skips Masayoshi Son because he thinks the obvious answer is Masa. He mounts a defense of Masa — the PowerPoint mockery, the xenophobic undertone in Valley gossip, the ARM bet everyone called stupid that printed money, and his underrated willingness to stand up for entrepreneurs when things get hard (Tony at DoorDash, Bom at Coupang).

But his pick is Yuri Milner. When Mehta was at D.E. Shaw and watched Milner invest in Facebook at a $10B valuation in a loss-making private internet company, it broke his mental model of what growth investing even was. DST’s impairment ratio has stayed remarkably low while Milner has personally caught ByteDance, Xiaomi, and others — any one of which is a fund’s worth of returns.

The Rwanda insurance disaster

Green Oaks’ biggest mistake wasn’t a company. It was a holding company called GGH, launched at the firm’s founding, that bought 51–100% stakes in P&C insurers in Pakistan, Rwanda, Nigeria. The thesis was clean: insurance penetration as a % of GDP follows an S-curve against GDP per capita; if you believe a country’s GDP grows, owning the best P&C insurer is a levered play.

In Rwanda, they bought the leading insurer and discovered the unit economics were catastrophic — $1.80 of loss for every $1 of premium, and that was the best performer in the country. The explanation: nobody audits, and as long as you write more premiums next year, you can roll the losses. Then a claim came in from a politically-connected family where, per Mehta’s telling, the claimant — the person allegedly killed in the car accident — showed up in court as the defendant. Green Oaks wound down GGH. It cost years.

SVB weekend and Carvana’s near-death

Two stories for fast conviction at scale.

Rippling, SVB weekend: Parker Conrad called Friday morning — Rippling wasn’t in trouble, but its plumbing (SVB) was, and Monday payroll for Rippling’s customers was at risk. 30 minutes of conversation, then a handshake on $500M. By Sunday it looked like the crisis would resolve. Conrad said: we’re doing it anyway. On Monday, every Rippling customer got paid on time.

Carvana, 2022: The stock went from $300+ to $5. Green Oaks started buying at $100 and kept buying all the way down. Ernie Garcia was losing $3,000 per unit in EBITDA terms plus $2,000 in interest per unit. Everyone said bankruptcy was inevitable. Mehta flew to Phoenix expecting a fog-of-war CEO. Instead Garcia spent ten minutes at dinner hunting for his Marriott points card. He wasn’t cutting frantically to placate the market — he was A/B testing internally to find the right cuts. Mehta saw a man slowing everything down in a crisis and calibrated his own conviction accordingly.

“What’s the difference between being down 95% and 97.5%? Just half.” — Benny Peretz, on doubling down at the bottom

The Ayn Rand part

Mehta deleted his personal email when he started Green Oaks. Still doesn’t have one. Has no hobbies. No beach house. Works 80 hours a week. Says zero interest in selling the firm, zero interest in the enterprise value of Green Oaks, zero interest in the Hall of Fame of AUM.

“I think you have to decide whether you want to be in the hall of fame of returns or the hall of fame of AUM.”

He also funds restaurants and a theater on Fillmore Street in San Francisco through a nonprofit — buying commercial real estate at 5.25% cap rates and renting it to mom-and-pops at 3% caps. The math is terrible on purpose. A politician named Aaron Peskin once led picket lines with Mehta’s face on a sign, calling him a billionaire taking over the city.

Key Takeaways

The one employee-poll question — best days ahead or behind — beats every quantitative KPI for judging a company’s health
JDCE (Jaw-Dropping Customer Experience) requires breaking trade-offs that competitors won’t touch; it’s rarely about features, usually about operational or technical violence over years
“Focus” is the discipline of letting everything else burn while you attack the single most important problem — most people don’t actually do this
10–15 founders a year matter; the industrialized matrix model of VC is correct for 90% of the industry but wrong for those 10–15
Volatility is welcome — most great businesses have 1–2 patches where the thesis looks broken; you can sometimes see through
The difference between the best and worst companies at Series B is enormous in outcome, tiny in pricing — that spread is where differentiated insight pays
High growth is healthy for bits businesses; “growth persistence” (growing 80–90% of last year’s growth) compounds into category ownership
The archetype of the great founder looks the same every time: focus, credible aggression, determination, divergent thinking; bad founders look different in a thousand ways
Red flags are often green flags — Musk being called unbackable taught Green Oaks to never outsource primary diligence again
Concentration in LPs, concentration in portfolio, concentration in attention — same philosophy at every layer of the firm
Critics’ most valid gripe: Green Oaks was early but “what have they done lately” — and some recent rounds look priced insanely

Claude’s Take

Mehta is the real thing. The podcast is long, Patrick O’Shaughnessy lets him ramble, and the rambling is the value — this is a guy articulating a coherent worldview at 40 that he clearly started building at 24 in Hong Kong watching Chinese teenagers glow-in-the-dark with QQ on their BlackBerries.

The BS filter finds little to flag. He’s honest about GGH (“single biggest mistake we’ve made”), honest about Carvana being a near-death experience, honest about the critique that some of his rounds look crazy. He doesn’t oversell the methodology — he explicitly says you could drop 40–50 IQ points and still do this if you have the discipline. That’s either radically humble or a bit of false modesty, depending on mood, but it’s not salesmanship.

The one soft spot is the “less competitive today than 10 years ago” claim. His argument is that the matrix firms are playing a different game — they want coverage, not ownership of the vital few. That’s plausible. But it’s also self-serving to say you have no real competitors, and the evidence for his framing is largely vibes and anecdote. There are probably 8–12 firms that would show up to the Rippling SVB call with $500M and a handshake.

The Yuri-over-Masa pick is a tell. It signals he values category-defining conviction at scale over volume of bets. It’s also a slightly contrarian flex in a room full of people who’d say Mike Moritz. The Moritz note about the handwritten-op-ed coaching is the warmest human moment in the interview.

Worth 8/10. One notch down from a 9 because the growth-is-good gospel is genuinely worth listening to but not exactly earth-shattering, and because the best-days-ahead metric, while elegant, is closer to an aphorism than a framework. One notch up from a 7 because the JDCE operational detail (Coupang package placement, Rippling plumbing, Carvana gift-certificate scene) is the kind of texture that travels.