Why Now is the Best Time to Buy Public Software Companies
Why Now is the Best Time to Buy Public Software Companies
ELI5/TLDR
Mitchell Green built Lead Edge Capital into a $3.5 billion fund by doing something most investors consider beneath them: cold calling 9,000 companies a year and filling his investor base with 800 former CEOs who open doors, check references, and make introductions. The firm targets boring, profitable software companies, buys them through every creative side door imaginable, aims for steady 2-5x returns instead of moonshots, and almost never loses money. His core thesis right now: everyone hates software stocks because of AI fear, which means it is the best time in years to buy them.
The Full Story
The Machine
Lead Edge Capital is three partners — Mitchell Green, Brian Neider, and Nema Besharat — running an investment firm the way you’d run a software company. They met at Bessemer Venture Partners and Insight Partners, took the best operational ideas from both, and built something deliberately mechanical. Their north star metric is not IRR or multiple. It is gross dollar retention of their investors: they want 95% of their money to come back every fund. That number forces two things simultaneously — good returns and good service. Most firms pick one.
The machine starts at the bottom. Eighteen analysts, mostly in their early twenties, call roughly 9,000 companies a year. That is not a typo. Nine thousand. The firm’s eight-point checklist filters that down to about 900 worth looking at, then 150-175 get real diligence, and five to seven get funded. The checklist is not magic. Green freely admits there is zero correlation between a company scoring eight out of eight and actually performing well. The checklist exists to shrink the universe fast enough that the firm’s scarcest resource — time — does not get wasted on noise.
“If you want to know if it’s a good company, just call 10,000 of them. You’ll figure out really quick.”
The Eight Criteria (and Why They Don’t Predict Returns)
The checklist: $10 million-plus in revenue. Growing at least 25% a year. 70%-plus gross margins. Recurring revenue. Capital efficient — meaning total revenue exceeds total historical cash burn, a rough cousin of return on equity. Profitable at the bottom line. No customer concentration. And a reasonable price relative to forward returns.
Five of eight gets you into the conversation. All eight narrows 9,000 companies to 90 — too small a pool to do five deals a year. Green’s honest about this: the criteria are not predictive. They are a strike zone. He uses the Ted Williams analogy. Williams mapped the strike zone into 77 cells and knew his batting average in each one. You can hit a pitch two inches above the zone for a grand slam. But if you swing there all career, your career will be short.
“Our biggest mistakes have honestly been not swinging at the pitches when they were in our strike zone.”
The LP Weapon
Here is where Lead Edge gets strange. Ninety-five percent of their capital — about 800 investors — comes from former CEOs and executives, not institutions. The former CEO of General Motors. The former CEO of Pfizer. The former head of HR at Procter & Gamble. Green uses them at every stage of the investment cycle.
Sourcing: if a target company will not return their call, they have the former CEO of a Fortune 500 company in the same industry send a note. Diligence: they call an LP who ran a company in the same sector and ask them to back-channel a reference on a customer relationship. Post-investment: they email 800 executives asking who knows someone at a restaurant chain because Toast needs introductions.
“It turns out all these people invest in funds and never get asked for help.”
The trade-off is brutal on Green personally. He spends 60% of his time with LPs. It would be far simpler to raise from 20 institutions writing $300 million checks. But when Green started the firm, he had no track record, no brand, and no reason for anyone to take his money. What he could offer was a Rolodex of executives who made his cold calls warmer. Fifteen years later, that structural advantage has compounded.
Doubles, Not Home Runs
Lead Edge targets 2-5x returns in three to seven years. At the fund level, that translates to roughly 2-2.25x net with 20% net IRR. They have lost all their money in exactly one deal, ever. The baseball analogy Green reaches for is Cal Ripken — doubles and triples, every day, for decades. Not Sammy Sosa.
The reason they almost never hit zeros: 85-90% of their portfolio is recurring revenue, 50-60% is already profitable, 70% of the time they own preferred stock with downside protection, and almost none of their companies carry debt. When something starts dying, they sell. A third of their exits have been secondaries. They run a formal disposition committee — the same three partners, meeting once or twice a month, walking the portfolio and asking a single question about each company: what is the forward IRR from here?
“The fastest way to get fired at Lead Edge is to have a company and not tell us when there’s a liquidity opportunity.”
Their best-looking vintages — 2015 through 2018 — were not genius stock-picking. It was multiple expansion plus discipline to sell. They expected 2x in four years and got 4x in two, which does extraordinary things to IRR math. The reverse happened in 2021. Everyone’s 2020-2021 funds are ugly for the same reason in reverse: expected 4x in two years, getting 1.6x in eight.
The Basement Window
Seventy percent of Lead Edge’s deployed capital goes through what Green calls creative structures. His analogy: buying an apartment. You can go in the front door (lead a primary round), the side door (buy secondary shares from an early investor), or through the basement window with a pickaxe (buy a derivative position through an LP in another fund that holds the stock).
The Zoom investment is the textbook example. The company did not need money. Sequoia was not selling. Direct secondary was blocked. So Lead Edge went to a small early fund whose LPs had been locked up for ten years. They offered to buy those LPs out or step into their shoes in a new vehicle. If the original fund owned 2% of Zoom and half the LPs sold, Lead Edge now owned 1% of Zoom — with voting rights.
In a world where LPs and GPs are desperate for liquidity and distributions have dried up, this part of their business is, in Green’s word, “booming.”
The AI Thesis: Buy the Fear
This is the headline. Green believes the best risk-adjusted returns available right now are in public software stocks. His reasoning has several layers.
First, software’s competitive advantage was never about R&D. It was about distribution, sales, customer success, and switching costs. Any of Lead Edge’s portfolio companies could be rebuilt by 500 Microsoft engineers in a month. Microsoft just does not care about the chamber of commerce software market or niche tax products. The moat is the customer relationship, not the code.
“If you think Exxon or Procter & Gamble is going to start building their own HR software after spending three to five years implementing Workday, you’re out of your mind.”
Second, he sees the AI capex cycle as a repeat of the telecom bubble. Too much money chasing infrastructure buildout. The assumptions required to justify current AI spending — earnings, power generation, nuclear plants — do not hold up. Apple, by staying disciplined, may end up looking like the smartest player.
Third, and most contrarian: he believes AI models will commoditize. Google, Amazon, and Microsoft have more training data than any standalone model company. Chinese and European models cost a fraction to run. You can run them locally. The question he keeps asking: why would a company outside the US pay for OpenAI tokens when DeepSeek costs almost nothing?
“I’m convinced that people who invest in all these AI companies have to portray the view that software is dying because they have to justify how much money they’re going to spend.”
The opportunity, in his view, is that this fear has crushed software multiples to levels where boring, profitable, recurring-revenue companies are genuinely cheap. Constellation Software’s stock chart looks like a ski slope. He is not panicking. He is shopping.
Fourth, he thinks the real AI risk is not to the independent software companies but to the private-equity-owned ones. Firms like Thoma Bravo load companies with debt, cut R&D and sales staff to hit “Rule of 50” metrics, and then brag about the margins. An overleveraged, understaffed software company with no innovation budget is the one that gets disrupted — not the entrepreneur-led company reinvesting in growth.
He draws the parallel to e-commerce in 2000. Everyone thought big-box retail was dead. The top 10 e-commerce companies in the US today are mostly… the same incumbents. Walmart, Home Depot, Target. The ones that died — Montgomery Ward, Kmart, Sears — were overleveraged and stopped innovating.
Culture as Competitive Advantage
Green sits down with every single employee once a year. Every partner, every VP, every associate, every receptionist. Three questions: what do you love and hate about your job, what would you change if you ran the firm, and what would make your job easier. He got the idea from Tom Barnes at Excel KKR.
He sends handwritten thank-you notes to nearly everyone he meets. The 22-year-old analysts now do the same. They track it. He puts 23-year-old associates in front of LPs, which 99% of firms would never do. His logic: if they are smart enough to hire, they are smart enough to meet a client.
The three partners negotiate their carry economics in ten minutes. Some firms spend months fighting over it. Green attributes this to genuine mutual respect and the fact that all three have been together since fund one.
The Ski Racer
Green was a nationally ranked ski racer. He races cars competitively. He sleeps four to five hours a night. When the guy who hired him at his first job was asked why, he said: “Because when things get scary, you’re going to want to buy.” The logic being: a person who goes downhill at 80 miles per hour does not find a market crash particularly frightening.
“If you want to generate generational wealth, you need to be an entrepreneur. I actually think it’s easier to leave when you’re 27 than when you’re 45 and have three kids. I had nothing to lose if it failed.”
The kindest thing anyone ever did for him: Pete Wilmont, the former CEO of FedEx and a Williams College alum, backed Green’s failed college startup, then served as his reference for Bessemer. Wilmont told the interviewer they would be insane not to hire Green because he was the most persistent person he had ever met. Wilmont has since passed away. Green clearly has not forgotten.
Claude’s Take
This is a genuinely interesting interview because Green is doing something rare: describing an investment machine with enough specificity that you could actually try to replicate it. Most fund managers speak in platitudes. Green gives you the funnel numbers (9,000 to 900 to 175 to 5-7), the target returns (2-5x in 3-7 years), the LP count (800), the time allocation (60% LP, 25-30% investing, 15-20% ops), and the creative structure mix (70% non-traditional). That level of transparency is either very confident or very calculated — probably both.
The intellectual honesty about the eight criteria is the most telling moment. He admits there is no correlation between a company scoring eight out of eight and actual performance. Most fund managers would never say that out loud. It suggests Lead Edge has genuinely pressure-tested its own process rather than mythologizing it. The criteria are a filter, not a crystal ball, and Green knows the difference.
His AI thesis is contrarian in a way that has a decent chance of being right. The “software is dead” narrative is driven partly by VCs who need software to be dead to justify their AI fund sizes. That is a structural incentive bias worth noticing. The comparison to the telecom bubble is apt — dark fiber and burning GPUs are not identical, but the pattern of massive capex preceding a demand reality check has repeated before. His point about model commoditization is strong: if DeepSeek can match GPT-quality at a fraction of the cost, the ceiling on model pricing becomes real very quickly.
Where I would push back: the “incumbents always win” argument is selectively true. Yes, Walmart survived e-commerce. But the incumbents that survived were the ones that invested aggressively in the new paradigm. The question for software incumbents is not whether Workday’s customers will rip out their systems — they will not. It is whether Workday’s new customer pipeline gets disrupted by something that is 90% as good at 10% of the cost, built by three people with AI tools. Green’s own point about Coupa displacing SAP’s neglected Ariba works against him here: the vulnerable companies are the ones that stop investing. But some AI-native competitors might not need much investment to get started.
The firm’s structural advantages are real but personality-dependent. Green’s energy, persistence, and willingness to spend 60% of his time on LP relationships are not easily transferable. The machine works because the machinist is obsessive. Whether Lead Edge survives succession is an open question that the interview does not address.
The “just go do it” entrepreneurship advice is survivorship bias in its purest form, but Green at least earns the right to say it — he started with nothing, his first company failed, and he built something genuinely distinctive. The Pete Wilmont story at the end is the kind of detail that makes you believe the rest of it.