Jun 16, 2025 · 1h 47m · 20vc
Larry Aschebrook, Founder & MP @GSquared: How We Lost Money on Uber and Made Millions on Lyft · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this deep-dive interview on the 20VC podcast, host Harry Stebbings sits down with G Squared Founder Larry Aschebrook to explore the firm's unconventional rise, high-concentration investment philosophy, and tactical mastery of secondary market liquidity. Larry candidly dissects G Squared's career-defining wins and painful losses, offering sharp advice on portfolio construction, navigating the AI boom, and the critical importance of prioritizing cash distributions (DPI) over paper gains.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 14.7% of the talking time here. How this is scored →
Harry as informed peer 3.2 Guest teaching 3.7 Guest disagreement 2.0 Harry pushing back 2.9
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:24:46 Rejection of dilution concerns in favor of DPI
Larry directly dismisses Harry's early-stage venture perspective on OpenAI, forcefully stating he does not care about dilution, only entry price and DPI.
Hardest push from Harry ▶ 38:57 Pushback on strict DPI guardrails as negative constraintsHarry directly challenges Larry on whether G Squared's strict fund lifecycles and DPI focus act as a negative constraint that cuts off multi-billion dollar gains in outlier positions like Palantir.
Biggest teaching moment ▶ 45:34 Explanation of structured equity IRR ratchetsLarry educates Harry on the mechanics of structured equity deals, explaining how minimum IRR hurdles and preferred ratchets extract value at the expense of common equity during market downturns.
Harry holds his own ▶ 1:19:07 Calling out guest's unexpected London winHarry demonstrates sharp market insight by humorously calling out Larry as a mid-American growth investor who somehow beat London insiders to Revolut back in 2018.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| The Rise of G Squared and Spotify | 1 | 2 | 1 | 1 | Harry asks open-ended conversational questions about Larry's background and entry into venture capital. Larry explains how he started buying secondary shares using a simple one-page form he created in business school. | |
| Divorce, Retirement Sacrifices, and Early Wins | 1 | 2 | 1 | 1 | Larry shares his personal sacrifices, including going through a divorce and cashing in retirement money to buy early Twitter and Alibaba shares. Harry asks straightforward follow-up questions about investment performance. | |
| The Foundation and Thesis of G Squared | 2 | 4 | 2 | 2 | Harry expresses surprise at Larry taking three years to raise his first $35M vehicle and asks about mechanics of fund closings. Larry educates Harry on the power of rolling closes and momentum in fundraising. | |
| Breaking Through with Alibaba and Outsourcing DD | 2 | 4 | 1 | 2 | Harry inquires how an unknown manager got access to Jack Ma's family office. Larry explains how he outsourced early investment committee memos and due diligence to analysts in India to save costs. | |
| Silicon Valley Syndicate Garbage and Secondary Discounts | 3 | 4 | 3 | 2 | Larry candidly critiques Silicon Valley syndicates for dumping bad deals onto new managers. Harry connects this to top-tier VC dynamics, while Larry highlights buying direct secondaries at 35 cents on the dollar. | |
| Unlocking Direct Secondaries with RIA Regulation | 4 | 5 | 3 | 5 | Harry challenges Larry on whether secondary buyers are detached from founder data and questions if true discounts still exist for premium assets. Larry reframes how full RIA registration and micro-transactions act as Trojan horses for data. | |
| The Stockholm Hustle: Snagging Spotify Secondary | 2 | 3 | 1 | 2 | Larry recounts flying unannounced to Stockholm to secure Spotify secondary stock from early executives. Harry listens and prompts the narrative forward. | |
| Tripling Down on Spotify and the Final Nine Million | 3 | 4 | 2 | 3 | Larry tells the dramatic story of borrowing $9 million from a heavy metal star seller on day 59 to complete a $150 million Spotify deal. Harry presses on diligence depth and execution pressure. | |
| The Lyft and Uber Divergent Outcomes | 4 | 4 | 2 | 4 | Harry expresses astonishment that G Squared made 3x on Lyft but lost money on Uber. Larry details selling Lyft pre-IPO while getting caught in Uber post-IPO valuation drops. | |
| DPI as G Squared's Unwavering North Star | 5 | 4 | 3 | 6 | Harry challenges Larry's strict DPI focus, asking if rigid fund lifecycles act as a negative constraint against holding outlier tech winners longer. Larry defends DPI as G Squared's non-negotiable metric. | |
| Sins of G Squared: Overvaluation, Hype, and Identity Crisis | 5 | 5 | 2 | 3 | Harry probes into G Squared's missteps during the 2020-2021 bubble. Larry delivers a detailed self-critique on overpaying for Toast, adopting Silicon Valley hype, and pivoting to structured equity with ratchets. | |
| Theranos and the Venture Capital 'Spider Sense' | 3 | 4 | 2 | 2 | Larry breaks down his near-miss with Theranos, explaining how a bad gut feeling led him to pay millions out of pocket to exit a binding agreement before public exposure. | |
| 23andMe and the Trap of Chasing Multiples | 2 | 4 | 1 | 2 | Larry recounts losing $70 million on 23andMe by failing to sell during the SPAC hype. Harry asks how LPs reacted to the loss. | |
| Secondary Liquidity, Bypassing ROFRs, and Ripe Avocados | 5 | 4 | 2 | 5 | Harry humorously connects selling Instacart at a 3x return to an un-ripe avocado delivered by an Uber driver. Harry asks technical questions about bypassing ROFRs and avoiding the tourist investor label. | |
| Getir Restructuring, Fighting Instincts, and Co-Investment Overhauls | 4 | 5 | 3 | 5 | Larry details the painful Getir restructuring and how his combative upbringing prevented him from walking away early. Harry pushes back on LP co-investment expectations during downturns. | |
| Spotting Early Fintech Winners in Revolut and Monzo | 5 | 4 | 3 | 4 | Harry praises Revolut as a generational outcome and jokingly asks how a mid-American growth investor spotted Revolut in London before local VCs in 2018. | |
| Portfolio Construction and Equal Weighting Across Megatrends | 6 | 5 | 5 | 5 | Harry argues he would put his whole fund into OpenAI at a $350B valuation. Larry aggressively rejects Harry's early-stage mindset, emphasizing that growth investors care about entry dollar price and DPI over equity dilution. | |
| AI Picks and Shovels and Legacy Tech Zombies | 4 | 3 | 1 | 2 | Larry describes legacy tech firms lacking native AI integration as vampires or zombies. Harry agrees and expands on why bottoms-up software re-architecture is hard for incumbents. | |
| The Venture Capital Liquidity Crisis and Selling Winners | 4 | 4 | 2 | 3 | Harry asks about private market illiquidity. Larry emphasizes that disciplined growth managers must be willing to sell winners to generate actual cash returns. | |
| The Complexities of Wealth and Fear of Poverty | 1 | 2 | 1 | 1 | Harry asks reflective personal questions about wealth. Larry speaks candidly about growing up on government cheese and how fear of poverty drives his wealth preservation mindset. | |
| The Psychology of Competitiveness and Building G Squared | 1 | 2 | 1 | 1 | Larry explains how his athletic background makes losses hurt far more than wins feel good. Harry asks if leaders should build immunity to losses. | |
| Quick-Fire Questions, Leadership Lessons, and Fund Metrics | 4 | 4 | 2 | 3 | In quick-fire questions, Larry cites Cal Ripken Jr. as his hero and dismisses TVPI and MOIC as fake paper metrics compared to true DPI. Harry steers through quick topics smoothly. |
Statements from this episode (54)
G Squared returned $800M to LPs from Coursera investment
“We were the largest shareholder of Coursera. Eight hundred million dollars back to LPs on Coursera.”
Aschebrook's 2010 one-page stock purchase agreement is still used today
“And still to this day, 15 years later, there's a form that I created at the Cary School of Business at Arizona State. One page form that then we sent to the company to buy stock. And it still floats around. I got it back from a broker not that long ago, and it…”
Aschebrook began investing by buying stock from Jack Ma's family office
“I just started buying Twitter shares and Alibaba stock from Jack Ma's family office, and then holy shit, it worked.”
Fund managers should close capital immediately to build portfolio momentum
“So with my mentality still to this day, raising 1,000,000,002 billion dollar funds is the money's there. You close. Start deploying it. Build a portfolio. Show some improvement in NAV. It is inertia that makes it easier.”
G Squared raised $1.4B and rejected $700M in 2021
“We raised a 1,000,000,004 and then turned down another seven hundred million. In very short period of time.”
G Squared outsourced early investment committee memos to analysts in India
“My early investment committee memos and the early vintages were outsourced to, you know, a group of really smart analysts in India, that you could pay a fraction of what you could pay talent in the US.”
G Squared concentrates 80% to 90% of its risk in 10 companies
“And still to this day, which some people are not comfortable with, 80, hopefully 90% of our risk are in 10 companies.”
Kleiner Perkins sold G Squared terrible early cleantech deals
“And there were some dogs in there, some cleantech shit that Kleiner Perkins sold me that was just a flaming bag of turds on your front porch.”
G Squared acquired early secondary shares at 35 cents on the dollar
“In early vintages, we got like 35 cents on the dollar by being a secondary direct buyer over primary buyers because there was no other secondary buyers.”
G Squared made a 3x return on Wiz in 18 months
“It starts to provide you that Trojan horse moment to know when to triple down on a whiz, which we did in our 20, 22 vintage fund, and have the outcome that we're gonna have. Those founders are amazing, but to make three X in 18 months, Is because of what they …”
G Squared agreed to a $150M Spotify deal with a $125M fund
“Before we left the meeting, they were like, do you think you could buy a hundred and fifty million dollars worth of stock? And I said, sure. And I didn't have the money, because we were in the process of raising the fund, and I think we had like closed on a hu…”
G Squared uses initial small checks as Trojan horses for company data
“So in that land and expand strategy that we run, as it's developed often, that first check is, is kind of that Trojan horse, because you want to, you like it, you've done some mosaic theory work, you think it works, but you don't really know, you go in.”
Sweden had 25% Spotify penetration when G Squared invested in 2014
“At that time, the penetration of the usership in, in, in Sweden was like, 25% of the population used it. I mean, that's amazing. 25% of the population used the product.”
G Squared generated $1 billion in LP returns from Spotify
“We then went on to become a top 10 global shareholder of Spotify. We went as far as to putting up signs in their break room. We'll buy your shares, buying all these odd lot transactions along the way and built a huge position. And it ended up being like a bill…”
G Squared bought Spotify secondary stock at a 50% discount
“We bought it like a 50% discount to the current financing round. And we just kept buying it there for the next two years.”
G Squared raised $700 million in Spotify co-investment vehicles
“And in that you take a lot of concentration in Spotify, but on top of the concentration we had, we did another seven hundred million in co-invest.”
G Squared rarely holds post-IPO stock, preferring pre-IPO sales
“Rarely do we hold posts public. Oftentimes we sell before they're public, and as it went public at the price it did, on a converted basis we were underwater.”
G Squared lost 20% on Uber and made 3x on Lyft
“We lost I think we made, we lost like 20 cents on the dollar, and Lyft we made like a three X.”
Silicon Valley exhibits sheep-like herd mentality in secondary trading
“One of the issues with Silicon Valley, one of the benefits of being in Chicago is we isolate ourselves a bit from the herd mentality. And it's like, sometimes it's like sheep. They just jump one on after another and they buy more stock and things that you see …”
G Squared eliminated gut feel from its investment process after 2021
“I think we had to really adjust ours after 20, 21, because I think it had holes in it that were Exacerbated. Too much qualitative. Too much my gut, my feel. Too much who else is in the room. Too much, yeah, just the soft stuff that venture capitalists and grow…”
Growth VCs miss 2.5x net targets if entry valuation is off
“Because at our stage, unlike yours, which it doesn't, you know, from your seat, in a lot of ways I envy your seat, because if you pay 50 pre or a hundred pre, if it's a good business, you're still gonna make a five X. A 10 pre, 50 pre, does it really matter fo…”
G Squared sold massive holdings during the 2020-2021 peak
“Oh, just massive amounts. I mean, we made ridiculous sums of money for LPs. In that period, which set our business up to be able to survive a bad vintage, and we're working our tails off, our asses off on our twenty-twenty vintage, but it's hard.”
G Squared lost $400 million of LP capital on investments
“Yeah, it's a big win, but hey, we also lost. Four hundred million dollars of their money over here.”
G Squared made a 3x return selling Palantir stock at $9
“I sold it at like, I think we made a three X on it and sold it at like nine dollars a share.”
G Squared's 2020 fund comprised 70% primaries and 40% structured equity
“And so that vintage of all of our vintages has 70% primary. With, and 40% of that has structure.”
Aschebrook personally paid millions to settle a Theranos secondary deal dispute
“And then actually I don't think it was formally sued. It was threatened to be sued. And I agreed to settle it out of court and gave him a couple million dollars to go away.”
G Squared agreed to buy $50M in Theranos secondary stock
“The agreement was to buy about fifty million dollars of stock over, like, four months, because, you know Because I, because I'm still raising the fund.”
Anne Wojcicki was a visionary founder but not a great public CEO
“You got Ann Wojciechowski, who's a fantastic founder, in my opinion to not maybe a great public CEO, but awesome founder and visionary of what she wanted to build and the ability to go do it.”
G Squared sold its final 23andMe shares at 70 cents
“We can start selling at seven, which would have meant I'm about a two X. I think I sold the last share at 70 cents.”
VCs should dollar-cost average out of positions post-IPO
“Dollar cost average out just like your dollar cost average in and just start getting liquidity when you can privately and continuing it through when they list.”
G Squared generated a 3x net return on Toast
“Toast, awesome business. It's like, three X net, and started in 2018.”
G Squared stopped share distributions to LPs to avoid blame
“We stopped distributing shares because LPs blame you if they hold it.”
G Squared put $200M total risk into Gorillas and Getir
“The total first was, it was like, fifty million bucks in total between secondary and primary, kind of, we got that plus sum and value of Gatir, and then the next check was a hundred million bucks. To restructure that equity, to pull it forward. So you get two …”
Getir's main mistake was expanding beyond low-labor-cost Turkey
“I think the mistake the company made there was it left Turkey and didn't focus maybe on other markets that were similar to Turkey from a employment cost perspective.”
Getir investment cost G Squared $500M in capital and LP churn
“In that case, from that fund, my guess is those decisions we made cost our partnership probably a half a billion dollars of capital. Between capital actually lost, and investors that you can't bring back.”
VC funds should only offer co-investments on core fund deals
“The way we run it today is the way I think people should run their, only in conviction, when your funds are investing alongside of it, that's when you do the co-investment. You don't do it the way we did it in twenty-twenty.”
Rapid startup growth prevents building scalable positions via dollar-cost averaging
“You can't, over time, dollar cost averaging in through lots of transactions, you can't build a scalable position if the company goes from zero to a hundred and No time.”
The venture downturn allowed G Squared to build a large Wiz position
“Normalized scenario, we would have never been able to build this position in Wiz because it would, it was a hockey stick from day one, but the capital availability, even to the best assets in the world, wasn't great for 18 months.”
Fanatics generates billions in revenue and hundreds of millions in EBITDA
“I mean, the business is fantastic, and it's at scale, massive scale, billions of dollars of revenue, and hundreds of millions of dollars of EBITDA, like a massive company.”
Growth investors should focus on absolute dollar prices and DPI, not dilution
“I don't give a shit about the dilution. I care about the price I pay. In dollars. And the price I'm going to sell it at, in dollars. I am focused on DPI, not MOIC. I don't care about the dilution.”
G Squared is actively buying secondary shares in Anthropic, OpenAI, and Databricks
“A lot. A lot. And I've got appetite for more. Same with OpenAI. Same with Databricks. Wiz.”
G Squared acquired a large stake in Anthropic via FTX bankruptcy auction
“You know, we were fortunate enough during, you mentioned FTX earlier, to be awarded during the bankruptcy a bunch of stock. We bid on it. We went through the whole process, bid on the stock, bought a big chunk, and now the company, look, the company is, in my …”
G Squared holds investments in GPU cloud providers Lambda and CoreWeave
“That's why we have Lambda. We also had CoreWeave.”
Hundreds of legacy private tech companies are 'zombies' doomed by AI
“You have incumbents that are large. We call them, there's two types of companies that were, that are big and private that aren't in AI. They're kind of, we talk internally, we call them vampires and zombies. Because they're massive. Many of them are profitable…”
Secondary shares and aggressive cost cuts can yield 5x in 3 years
“No, and then, by the way, layer some secondary on, lower the cost basis, and you have, you can make a five X in three years.”
Offloading $1B of Anthropic stock in private markets is extremely difficult
“Today I think there's a handful of liquid private companies basically. They're quasi-liquid. But at scale. Small positions, yes. But try to go move a billion dollars of Anthropic.”
VC funds must sell winning companies to build a sustainable business model
“Sell, you know, sell your dogs and keep your winners, and I think it's actually the, it's somewhat of the inverse, is you have to be able to, willing to sell your winners. To have a sustainable business model. And to generate the return that will actually driv…”
G Squared bought Bolt at $20 per share before Sequoia paid $260
“I think it was like, 20 bucks a share or something, and Sequoia did it at two 60.”
G Squared made 8-10x on Spotify and 6-7x on Bolt
“Spotify, early investment that he trusted our business model in, let's say eight to 10 X. Bolt, you have to be determined, but let's just assume it's somewhere between, you know, 50% and face value of what Sequoia paid. It's, you know, six, seven X.”