Mitchell Green: Why 50% of VCs Should Not Exist

A Money Maker in a World of Framework Thinkers

Mitchell Green is the founder of Lead Edge Capital, a growth equity firm with over $5 billion in AUM. His track record includes early or co-led investments in Alibaba, ByteDance, Grafana, Benchling, and Asana. He’s not a venture capitalist who deals in narratives; he’s a self-described “money maker” who targets 2-5x returns in 3-7 years. In a market gripped by panic over AI disruption to software, Green is buying.

This conversation with Harry Stebbings cuts through the noise of SaaS apocalypse headlines with a growth equity investor’s cold math. Green’s thesis is contrarian on multiple fronts: incumbents aren’t dying, China wins AI, half of VCs shouldn’t exist, and the best time to invest is approaching.

The SaaS Selloff Is a Setup, Not an Apocalypse

Green is actively buying public software stocks: Procore, Workday, Appian, Toast. His argument is straightforward. These incumbents have distribution, data, and balance sheets. They’re not going anywhere.

The selloff, he argues, is driven by Wall Street analysts’ numbers being too high heading into the year. Companies like Workday at 6.8% growth look terrible, but it’s a $10 billion revenue business generating $3 billion in free cash flow. The deceleration was predictable; analysts just didn’t model it.

“Give it a quarter or two. Companies will start to beat numbers. They’ll raise the numbers and the stocks will start to work, but it’s probably dead money for a little while.”

His playbook for individuals: don’t try to time the bottom. If you want $200,000 of a name, buy $50,000 every time it dips hard. Longitudinal data on the NASDAQ shows buying on big down days is nearly impossible to time but reliably profitable over the long term.

The 1999 Analogy: Who Survives Disruption

Green draws a sharp parallel to the dot-com era. In 1999, the debate was whether traditional retailers would be destroyed by e-commerce. Twenty-five years later, six or seven of the ten largest e-commerce companies in the US are traditional retailers: Walmart, Target, Home Depot, Lowe’s.

The companies that died (Sears, Kmart, Montgomery Ward, Bed Bath & Beyond) shared one trait: massive leverage that prevented them from innovating. Walmart had no leverage and went all in.

The same logic applies now. Any company with a mountain of debt, whether software, manufacturing, or services, is vulnerable in a technological disruption. The ones without leverage can invest aggressively and adapt.

“The mainframe business is still a $5.5 billion market. Mainframes came out in 1950. Most banks are run off of mainframes.”

Oracle, Microsoft, SAP are all “legacy” software companies. They’re also some of the biggest in the world. Legacy doesn’t mean dead.

ByteDance: The Most Underrated AI Company

Green’s most provocative position: ByteDance is the most advanced AI company in the world, dramatically underappreciated by the West. Lead Edge has been buying ByteDance stock aggressively, starting around the $200 billion implied valuation.

The math: Alibaba and Tencent trade at mid-teens earnings multiples with minimal growth. ByteDance grows 25-30% annually with massive profits. At Facebook’s 25-30x earnings multiple, or even at a discounted “China company” multiple, Green sees ByteDance doing $70-100 billion in earnings within five years. At 20x, that’s a $2 trillion company.

When Harry presses on the de-globalization risk, Green is dismissive. A US listing is “0% chance,” but a Hong Kong listing works fine. And despite years of threats to delist Chinese companies in the US, it never happened. Alibaba’s stock has doubled off its lows; China sentiment is improving.

Why China Wins AI

Green bets China wins the AI race. His reasoning:

• Power infrastructure: China can build nuclear and solar power plants in a couple of years. The US faces major power constraints and growing local pushback from communities dealing with data centers that triple electricity prices while employing only 50 people.
• Human capital: the number of PhDs, how much they value science and technology.
• Innovation track record: the biggest innovations in internet over the last decade or fifteen years have come out of China. E-commerce, social media, everything. DeepSeek was not a surprise to him.
• Cost efficiency: Chinese ingenuity in reverse engineering and building things at a fraction of American costs.

He also flags an underappreciated risk for US AI buildout: local communities pushing back against data centers. Small towns in Iowa, Kansas, or Virginia get giant ugly buildings, 50 local jobs, tripled power prices, and environmental concerns, while Silicon Valley makes hundreds of billions. Green expects real regulation to follow.

Buying Is Glamorous, Selling Is the Job

This is Green’s mantra. Lead Edge constantly rewrites its positions, asking: what’s the probability this doubles from here? If yes, hold. If a buyer offers a price that caps the upside, sell.

His “18-month rule” for investment discipline: if you invest today and the company hits your model for 18 months, are you in the money at a reasonable multiple? Not 50x revenues, a reasonable multiple. If the answer is no, don’t invest. This simple test has kept Lead Edge out of a lot of trouble.

On the ByteDance position specifically: at $550 billion (reported General Atlantic sale price), they’ve been offered higher. At $1.3 trillion, they’d sell “a bunch.” The math says the company will do $100 billion in earnings in five years; at 20x, that’s $2 trillion. So there’s room to run.

50% of VCs Should Not Exist

Green doesn’t soften this: 50% of people in the venture business should not be in it. Maybe 70%. It’s not just VCs; it’s private equity and alternative asset managers broadly.

The problems he identifies:

• Too much money, too many tourists: people who don’t show discipline on price.
• Napkin-valuation lunacy: founders spinning out of Anthropic or OpenAI raising at $2 billion for nothing more than an idea. Green questions whether any company that raised at a multi-billion valuation on just an idea has ever worked. (Harry counters with Anthropic; Green notes the original seed was much lower.)
• Negative value investors: 50-60% of people in the industry probably add negative value to companies. The worst offender is someone who went to Stanford Business School, worked 18 months at a startup, becomes a venture investor, and acts like an expert.

What investors should actually do: help founders recruit. Connect them with people who’ve built businesses from $20 million to $200 million. Get out of the way. This is why Sequoia, Benchmark, and Index are great: they help portfolio companies recruit amazing talent.

DPI Is Math, Marks Are Opinions

Green is evangelical about returning cash to LPs. His advice to younger fund managers:

“Liquidity windows open and close. When they are open, take advantage of them. You should be selling, even your winners. Sell 20%, sell 30%, sell 5%. Your job is to return money.”

The industry’s DPI problem is real. In 2030 or 2032, LPs will wake up and realize they’re still sitting on positions from 2012 and 2015. The fund managers who survive 10-20 years are those who consistently return capital.

He praises FJ Labs’ Jose Marin and Fabrice Grinda as examples of investors who play the DPI game well: they understand not every company goes to the moon, take chips off the table, and rinse and repeat.

On mega-funds ($15 billion a16z, $10 billion Thrive): Green thinks they’re “insane.” A $9 billion growth fund needs to write investments into $150 billion companies. At steady state, companies trade at 10x earnings. To make a double with dilution, you’re betting a company will do $25 billion in earnings. Very few companies achieve that. But he acknowledges these funds might be the ones to find the next social-media-scale platform, so there’s an argument both ways.

Gross Dollar Retention: The Only Metric That Matters

For the SaaS founders DMing Harry daily about hitting the “triple triple double double” growth targets but getting no investor interest at $20 million ARR, Green says: call us. Lead Edge finds these interesting because they can get in at good prices.

But the single most important number: gross dollar retention. Not net (which includes upsells), but gross. You ended 2024 with $20 million in revenue. Those same customers, no upsells, just downsells: what did you end 2025 with?

• 90% (= $18M): good
• 95%: great
• 98%: amazing
• Below 90%: won’t touch it

The dead wood problem in venture: companies with 60-70% gross dollar retention. It’s manageable at $10 million ARR, but at $150 million, they’re just churning through customers. The company with 95% retention grows fast without burning cash because it’s not spending on sales and marketing to refill a leaky bucket.

The Meme-ified Stock Market and the Liquidity Trap

The casinoization of public markets creates a paradox. Green notes a random research report (the Catrini piece) can get 25 million views and move markets, while people like Stan Druckenmiller, Howard Marks, Ken Griffin, and Mark Benioff say software isn’t dead, and nobody listens.

This actually creates opportunity for long-term investors. Buy good businesses at multiples of earnings when everyone’s panicking. But it also makes the liquidity problem worse: if you’re Canva or Stripe, why go public in this environment? Stripe feels vindicated staying private. But that hurts LPs who can’t get distributions.

Green’s solution: the pressure will come when LPs tell the biggest venture and PE funds “we’re not investing in your next fund until you get liquidity in these names.”

One underappreciated issue he flags: stock-based compensation dilution in big Silicon Valley companies. The amount of equity dilution to shareholders is enormous, yet few people talk about it. Companies that respect share count (like Ellison’s leveraged recap of Oracle, buying back massive amounts of stock without selling his own) create real shareholder value.

The Coming Downturn Is the Best Opportunity

Green’s most counterintuitive excitement: he’s looking forward to a major downturn in the next 10 years. Markets don’t go up forever. Combined with AI productivity gains, it creates the setup for a generational investing opportunity.

His parallel: just as the best internet companies (the ones that really scaled) were started in 2003-2006, after the dot-com bust, the best AI companies may emerge after the current hype cycle burns out. “Gen one” AI companies are the equivalent of Pets.com. The real winners haven’t been started yet.

“Always have money to play the game. If you don’t have money, you’re out of the game.”

Some Thoughts

Mitchell Green’s worldview is refreshingly mechanical in an industry addicted to narrative. A few observations worth sitting with:

• The leverage test is the simplest diagnostic for who survives disruption. Not whether a company is “AI-native” or “legacy,” but whether it has the balance sheet to invest through the transition. This applied in 1999 and applies now.
• The 18-month rule is an underrated framework. Most bad growth investments fail this test at the point of entry but get funded anyway because “the company is great.” Good company and good investment are two fundamentally different things.
• His China thesis is genuinely contrarian for a Western investor. Not hedging, not “China will be competitive.” He’s saying China wins, primarily because of energy infrastructure advantages most people ignore.
• The DPI sermon is aimed at a specific audience: emerging managers in their first few funds who confuse marks with returns. “Marks are opinions, DPI is math” is the kind of line that sounds obvious but is practiced by maybe 20% of the industry.
• His critique of mega-funds is math-based, not emotional. At $10 billion, you need to find companies that will do $25 billion in annual earnings. There are perhaps ten such companies in the world today.