The Twenty Minute VCLarry Aschebrook, Founder & MP @GSquared: How We Lost Money on Uber and Made Millions on Lyft
CHAPTERS
- 0:00 – 1:02
From big wins to painful misses: Coursera, Lyft, and losing on Uber
Larry opens with the core paradox of GSquared’s track record: enormous distributions to LPs alongside very real losses. He frames the firm’s strategy around concentration and liquidity—returning cash, not just paper gains.
- •$800M returned to LPs from Coursera; major gains in Lyft
- •Counterintuitive headline: made millions on Lyft, lost money on Uber
- •Concentration and timing drive outcomes more than “always pick winners”
- •Sets up the episode’s theme: wins are public, but mistakes are defining
- 1:02 – 5:43
Bootstrapping into venture via secondaries: buying private shares with nothing
Larry describes starting from poverty, working in fundraising for academic endowments, then deciding to build wealth by investing in private companies. In business school he begins buying private shares directly from classmates and contacts—before he even realizes how unusual it is.
- •Background: raised money for endowments; saw private-company investing create wealth
- •Business school thesis becomes the foundation of GSquared’s model
- •Early behavior: ask people directly to buy private shares
- •Creates a simple one-page stock purchase form that becomes binding (and risky)
- 5:43 – 6:51
The core thesis: companies stay private longer, creating a liquidity opportunity
Larry explains the macro shift post-financial crisis: fewer IPOs, more capital, and longer private timelines. That creates an opportunity to buy from employees and early holders who need liquidity—often at meaningful discounts.
- •Post-2008: IPO timelines lengthen dramatically, liquidity dries up
- •Direct secondary buying targets holders with no liquidity options
- •Early advantage: large discounts vs primary rounds due to limited secondary buyers
- •First-time manager problem: no pedigree, so access comes from hustle
- 6:51 – 9:39
Raising the first fund the hard way: three years to reach $35M
He recounts fundraising for the first vehicle—slow, multi-close, and deployed as capital arrived. Larry shares his philosophy for new managers: close when you can, start building NAV momentum, and use inertia to keep fundraising moving.
- •Three-year fundraising journey (2010–2013) to ~$35M
- •Multiple closes; deploying capital as it comes in
- •Advice: close money when available, deploy, show NAV progress
- •Being forced to raise slowly can improve discipline and process
- 9:39 – 20:20
Alibaba as the first real inflection point (and the “OPM” lesson)
Alibaba’s 2014 IPO becomes the first material proof that Larry’s approach can work. He explains how improbable access happened (networks and luck), how scrappy diligence looked early on, and why he prefers a “land and expand” approach with heavy concentration.
- •Bought Alibaba shares via Jack Ma’s family office connection; IPO validates the model
- •Early DD outsourced to analysts in India due to resource constraints
- •Investment style: start small, learn, then concentrate heavily
- •Concentration reality: a handful of names drive most outcomes
- 20:20 – 30:35
How GSquared got Spotify—and turned it into a billion-dollar outcome
Larry details the Spotify origin story: flying to Stockholm without an appointment, earning trust, and being asked to buy far more stock than the fund had available. He describes how access plus repeated “odd-lot” purchases built a massive position at a discount during peak fear.
- •“Show up” strategy: flew to Stockholm to get a meeting after no responses
- •Spotify’s approval process for shareholders; deep data access once inside
- •60 days to raise a massive allocation; even borrowed money to close
- •Bought amid artist/label tension; record labels’ behavior signaled conviction
- •Outcome: top-10 shareholder, continued buying, ~billion-dollar return to LPs
- 30:35 – 32:32
Lyft vs Uber: why one became a 3x and the other a $50M loss
Larry explains why GSquared sometimes backs competitors and how timing around liquidity matters more than brand narrative. Lyft was monetized before IPO at attractive levels, while Uber’s IPO pricing and their typical exit behavior left them underwater.
- •Strategy: sometimes back two companies attacking the same market
- •Lyft: sold mostly pre-IPO, realized ~3x
- •Uber: IPO timing/pricing plus their exit discipline led to losses
- •Approximate Uber loss cited: ~$50M
- •Chicago distance from “herd mentality” can help identify sell points
- 32:32 – 43:52
When rationality left the room (2020–2021): guardrails, overpaying, and the Toast wake-up call
The discussion shifts to the bubble period and Larry’s admission: the firm kept deploying like everyone else. Toast’s public trading levels become the “canary in the coal mine,” triggering a realization that they’d overpaid broadly and needed new risk controls.
- •2020–2021: public and private multiples detach from fundamentals
- •Guardrails mattered—but their framework had holes
- •Key failure: too much qualitative “gut” and social proof in decisions
- •Toast’s valuation spike signals overheating; triggers internal alarm
- •For growth-stage investing, entry price sensitivity is acute (small pricing errors matter)
- 43:52 – 53:53
“We fcked up”: pivoting mid-vintage with structure, selling losers, and lowering cost basis
Larry details the crisis response: admitting mistakes to LPs, raising additional capital to protect the vintage, and using structured equity and secondaries to repair portfolio math. He explains how IRR hurdles and ratchets can reshape outcomes—often at the expense of earlier holders.
- •Went back to LPs requesting additional capital to “protect” the vintage
- •Shift to structured equity: minimum IRR / multiple hurdles embedded in paper
- •Tactical response: cut losses, sell inflated assets, use secondaries to average down
- •Primary-heavy vintage with substantial structure changes the pref stack dynamics
- •Lesson: avoid “believing your own hype” and drifting from the firm’s true model
- 53:53 – 59:38
Dodging Theranos (mostly): the gut-check, the contract mistake, and the process lesson
Larry recounts almost buying Theranos via a secondary deal, walking away on instinct, then learning the hard way that sloppy paperwork can still be expensive. The episode becomes a case study in separating “good outcome” from “good process.”
- •Met Theranos management (not Elizabeth Holmes); something felt off
- •A binding one-page form created legal/financial exposure
- •Paid a settlement personally to protect LPs from the fallout
- •Key lesson: never sign before certainty; process must be airtight
- •“Spider sense” matters, but must be paired with disciplined execution
- 59:38 – 1:02:43
Losing $70M on 23andMe: the cost of chasing multiple instead of taking liquidity
Larry explains the 23andMe investment: genuine belief in the mission, a large position, and then a failure to sell during the SPAC window. It’s a blunt lesson in exit discipline—especially for a strategy optimized for DPI and capital velocity.
- •Entered 2016–2017; built a major position by 2018 vintage
- •SPAC boom offered an exit path (could’ve sold for ~2x)
- •Held too long; sold last shares near $0.70
- •Approximate loss cited: ~$70M on a ~$100M position
- •Fix: layered selling / dollar-cost averaging out to lock in liquidity
- 1:02:43 – 1:05:58
Getir/Gorillas: good first check, disastrous second check, and “fighting” past the exit
Larry describes why Getir became emotionally and operationally draining: restructuring, board involvement, and the temptation to throw good money after bad. The central mistake wasn’t the initial thesis—it was refusing to walk away after the first tranche.
- •Started with Gorillas in Berlin; later acquired by Getir
- •Initial exposure ~ $50M; follow-on restructuring check ~ $100M
- •Board-level involvement, heavy restructuring, founder battles (ongoing)
- •Core lesson: the “fight” mentality can become a curse
- •Total damage includes not just losses, but future fundraising/capital opportunity cost
- 1:05:58 – 1:18:37
The ‘janitor’ role in venture: ROFRs, microtransactions, and when co-invest goes wrong
Larry reframes GSquared as a point-in-time liquidity and cap-table problem solver rather than a long-term ‘tourist’ or traditional VC. He also dissects co-invest: why it was necessary for scale early, what broke in 2020-era bespoke LP requests, and the tighter rules they use today.
- •Positioning: solves liquidity/cap-table cleanup for companies and funds
- •ROFR navigation depends on trust, value-add, and repeated small transactions
- •Operational edge: many small deals build data access and a path to concentration
- •Co-invest early: sometimes 4x co-invest to fund due to access/scale needs
- •2020 mistake: co-invest outside core conviction positions; LPs blame you when it fails
- •Current model: co-invest only alongside top positions in the fund, often equal-weighted
- 1:18:37 – 1:22:52
Getting to Revolut early and building megatrend portfolios (fintech, SaaS, consumer, mobility)
Harry presses on how GSquared identified Revolut in 2018 despite being US-based. Larry attributes it to megatrend focus, prior fintech wins, and the practical constraints of building meaningful positions before companies hockey-stick too quickly.
- •Fintech as a core megatrend; learnings from SoFi and N26 informed Revolut interest
- •Revolut’s rapid trajectory makes position-building hard via slow accumulation
- •Monzo framed as a later “value play” relative to Revolut
- •Portfolio construction: ~10 high-conviction positions balanced across megatrends
- •Risk management: balance high-momentum bets (e.g., AI) with scaled cash-flowing businesses (e.g., Fanatics)
- 1:22:52 – 1:27:53
Why buy Anthropic at $61B: DPI mindset, leaders-only AI strategy, and picks-and-shovels
Larry argues that in AI, concentration belongs in category leaders because time and capital barriers limit credible new entrants. He explains why he ignores dilution (focuses on dollars in/dollars out and DPI) and complements foundation-model exposure with infrastructure and enabling platforms.
- •Thesis: foundation model leadership is durable; few winners (OpenAI/Anthropic)
- •Would buy Anthropic at ~$61B; prioritizes price path and DPI over dilution math
- •“Go to the winners” because the fund’s horizon is shorter
- •AI approach: leaders + picks-and-shovels (e.g., infrastructure) + embedded AI exposure across the rest of the portfolio
- •Liquidity nuance: small positions can move; moving huge blocks at scale is different
- 1:27:53 – 1:33:55
Vampires vs zombies: the coming purge, liquidity reality, and what LPs misunderstand
They discuss the looming shakeout among large private companies that lack AI-native DNA and/or can’t adapt. Larry emphasizes that liquidity is harder than most LPs appreciate—and argues DPI is the only metric that truly matters for allocators.
- •“Vampires vs zombies”: big private companies that can adapt vs those that are effectively dead
- •AI transition is hard; can’t just “slap AI on” products
- •Liquidity is often harder than access; timing exits is brutally difficult
- •Critique: TVPI/MOIC can mislead; DPI is what funds can actually spend
- •Industry response: evergreen/interval/continuation structures as symptoms of liquidity strain
- 1:33:55 – 1:39:19
Does money make you happy? Complexity, edge, and the psychology of wins vs losses
Larry reflects on how wealth changes life: it removes friction but adds complexity and social drift. His motivation remains the chase and competition—losses imprint more deeply than wins, and maintaining “edge” becomes harder as comfort grows.
- •Money makes life easier, not necessarily happier; complexity increases
- •Personal driver: fear of returning to poverty and desire to protect future generations
- •Staying grounded gets harder as your environment changes
- •Edge comes from competition and pursuit of the next win—not from status
- •He doesn’t seek immunity to losses; the fight is part of his identity
- 1:39:19 – 1:47:28
Quick-fire: admired grit, leadership growth, and why paranoia can be productive
In the closing rapid round, Larry highlights durability and grind (Cal Ripken Jr.) as an ideal. He also shares what he’d change about his leadership style and how productive paranoia helps him close, invest, and stay focused on what can break next.
- •Admires longevity and discipline: Cal Ripken Jr. and the ‘20-mile march’ mindset
- •Would change: be ‘chiller’ without losing intensity; reduce collateral damage
- •Memorable LP moment: on-the-spot commitment in London
- •Advice to endowments: stop chasing TVPI/MOIC; focus on DPI
- •Paranoia helps with discipline and urgency; can hurt if it fuels over-fighting
