Carvana CEO & Co-Founder, Ernest Garcia: Building a $50B Company, Losing 99% and Coming Back

Summary

• Carvana’s 2022 collapse — stock down 99%, bonds at 40 cents on the dollar — was never the real near-death experience to Ernie Garcia; the real ones in his mind were two or three early capital crunches. His core lesson is that risk is less risky than people believe: “when you find yourself with your back against the wall, you can find action you might not have found otherwise… you have more moves than you think.”
• Every Silicon Valley fund wanted Carvana to be “a software layer on top of dealerships”; Garcia refused, believing vertical integration was essential to control the customer experience — a stubbornness that “led to at least one extra near-death experience” but that the team persisted in defending. His VC critique: when a business requires solving ten layered risks in a row, very few investors will take it on, so the most important input is judging the people.
• On AI and defensibility: software layers are getting easier to replicate, so the physical and business-process layers underneath remain the harder layers of defensibility — “if the software is easy to replicate, the next two layers are still hard.” Yet he names AI as Carvana’s relatively weakest layer today: “all businesses today are using AI at a small, small fraction of what’s possible,” and catching up to a horizon improving that fast “is a battle.”
• Garcia defends being public against Harry’s pushback that Greenoaks or Thrive boards could replicate the pressure: “When you’re public, it’s just cold. It’s ruthless. It’s just results” — relationships grade you on intentions and actions, the stock doesn’t. You learn who real long-term investors are only “when it’s not in their short-term interest to be a long-term investor”; in 2021, approximately zero said they weren’t long-term-focused.
• IPO pricing heresy: price it as high as you can — Carvana went out at $15, opened at $12, hit $8 within a week, and “no one cares” now; underpricing for a synthetic first-day pop just means less money, more dilution, and higher odds of dying. The post-IPO all-hands where everyone questioned the company built resilience a private company never gets: “there’s no forcing mechanism to make you talk about it.”
• Debt was “dangerous” pre-cash-flow but became an accidental pressure machine: it cost less dilution than equity would have, and huge cash flows that had to be covered plus “a market that’s telling you you’re stupid every day” generated pressure the team could never have self-created. The balance-sheet lesson he admits he’ll “never fully internalize”: carry more cushion than you think you need in build phases.
• Changed mind in the last 12 months: focus requires constant tending — “every 3 months you have to uncomfortably take things off your board.” After two years of pure profitability focus, Carvana is shifting finite focus back to growth; infrastructure “clearly supports 3 million” cars, alongside the roughly 400K figure discussed for a 40M-unit US used-car market, and the dealership purchases are “extremely early” experiments he won’t discuss.
• The founder psychology is unusually honest: the motivation wasn’t pleasing his father but beating him — “if he can do it, I can do it and I can do it better.” He’d chase expensive dopamine over cheap (“stock goes up, cheap dopamine”), thinks for every Steve Jobs “there’s probably a hundred people of similar quality” without the luck, and prescribes sport as his best answer for raising hungry kids in affluence.

Deep dive

1. Entrepreneurs are stubborn egomaniacs — and refusing to be a “software layer” proved it

• Garcia owns the label he once coined: “it’s really hard to sit in a room with a bunch of people who tell you you can’t and to believe that you can if you’re not stubborn and if you don’t have maybe a touch too much self-belief.”
• The defining early test: six months in, “we talked to every fund that you could possibly imagine,” and every one asked, “Can’t you just be a software layer on top of dealerships?” He concedes there was “a very good chance” they’d have raised money by bending — but the core belief was that you can’t deliver the customer experience without doing it all: logistics, reconditioning centers, finance connected directly into trade-ins, software on top.
• The cost was explicit: “making that choice led to at least one extra near-death experience.” His justification — most companies taking a big swing have multiple near-death experiences anyway, so sacrificing what you deeply believe to avoid one extra isn’t worth it.

2. The 99% drawdown wasn’t the near-death experience — risk is less risky than you think

• In 2022 the stock fell 99% and bonds traded at ~40 cents on the dollar — “to the world, that looked like a near-death experience. Honestly, from my perspective, I didn’t ever think it was that.” The real near-death moments were two or three early capital crunches, when “capital is the lifeblood” and they struggled to get it.
• His signature framing: everyone knows a story where an injury 1mm over would have killed someone. “Maybe everyone’s lucky, or when we reassess things in retrospect, we think it was closer than it was… you have more moves than you think.” With problem solvers around, back-against-the-wall moments yield fundraises, efficiencies, and stretched dollars you might not otherwise find.
• The conclusion he actually draws, hedged exactly this much: “I think you need to take risk to do something meaningful. And I think risk is less risky than most people believe.”

3. Venture struggles to underwrite ten risks in a row — so people matter most

• His “PC answer” for why Carvana couldn’t raise: capital and operational intensity were unpopular, marketplaces were the pattern, and “Phoenix was not viewed as a market where you could build a company.” His at-the-time answer: “Isn’t the way this is supposed to work that venture capital is supposed to fund risk in the economy?”
• The structural diagnosis: investors will take two or three stacked risks, “but when you’ve got like 10 in a row, it’s very hard to find that match.” Since you can’t foresee ten sequential problems, “the most important input” is “trying to understand the people.” You have to decide you believe in them to ultimately solve all the problems as they come.

4. AI raises the value of physical and business layers — and AI is Carvana’s relatively weakest layer

• To Harry’s question whether AI makes Carvana’s logistical “nightmare” more valuable: “the simplest answer I believe and I hope is yes.” A software layer sits on physical processes which sit on business processes — “if the software is easy to replicate, the next two layers are still hard. You want to have as many layers as you can.”
• Pressed on the layer he most lacks, he names AI itself: “all businesses today are using AI at a small, small, small fraction of what’s possible” given the quality of the technology today. “And the technology is improving so fast it’s unbelievable… trying to catch up to that horizon is a battle.” Relatively speaking, “we’re weakest there” — a new capability set Carvana hasn’t invested in for long.

5. Operators over strategists; abstraction is the enemy — go sit behind the person

• Forced to pick one, “you want operators… getting from zero to one is the operator, but the ceiling of your possible success is a function of the conceptual thinker.” The two archetypes usually repel each other — operators see strategists as head-in-the-clouds idealists; strategists find operational problems un-campfire-worthy — so pairing them is rare and valuable. His specimen of both: Jeff Wilke, “as fluent in frameworks and concepts” as he was obsessed with why Carvana’s car-parking shapes were slowing movement.
• His anti-abstraction doctrine: “the closer to the ground the better you are… the pressure of scale pushes everyone up the pyramid” toward one-on-ones and influence. The fix: in any readout, find the person everyone’s heads turn to when the question isn’t on the slide, then “go check in with that person every day… pull up behind them in a chair and look at a screen together.”
• Self-assessment worth keeping: seven direct reports, and “I would not consider myself a great manager. I consider myself much better at trying to solve problems and get involved than I am a natural manager.”

6. Hire on respect, not resume — and beware the charismatic storyteller

• His hiring rule inverts credentialism: “I would way rather take a person in any role that is respected by someone that I respect than someone who’s done it before… I don’t care that much if you did it at a company everyone thinks is great.”
• His biggest people mistake is an archetype he says is “even more common in Silicon Valley”: smart, conceptual, articulate, high-energy storytellers who “lack the ability to find the occasions where they’re wrong and own that” — instead shifting the explanation and bouncing to the next problem, always with followers in tow. “When you see those people, you’ve got to address it quickly,” or they produce “a ton of wasted effort.”
• Why Carvana’s exec team never bounced: they’re “not motivated by status as much as most are.” Departures happen when difficulty hits and “blame starts to show up… it starts to feel like this is no longer great for my resume.”

7. Public markets are cold, ruthless, and just results — which is exactly why you want them

• Harry’s pushback — Greenoaks or Thrive board members could provide the same pressure privately, without the costs — gets a flat “I respectfully disagree”: “the stock does not care about who you are. It doesn’t care about your relationships… When you have a relationship, some portion of how you’re evaluated is based on intentions and actions. When you’re public, it’s just cold. It’s ruthless. It’s just results.”
• His advice to anyone considering an IPO is a values question: “Is it important to succeed or is it important to have the best personal life that you can have? If success is more important, you want to be public.” He doesn’t love it himself — “it’s a cost of the goal.”
• Even earnings prep, widely seen as a waste, forces him out of his functional view into an accounting lens: “the reason we’re going to get a bunch of questions from a bunch of smart people is because they’re good questions… productivity is found by confronting the things that you don’t want to confront.”
• On short-termism: “In 2021, what percentage of investors said they weren’t long-term focused? Approximately zero… you find out who the real long-term investors are when it’s not in your short-term interest to be a long-term investor.” He endorses Benjamin Graham’s voting/weighing machine “100%” — high potential necessarily means high volatility, “and you just got to deal with it.”

8. The IPO from hell built resilience — and price it as high as you can anyway

• The IPO as told: a roadshow of ten 45-minute grillings a day for a week (“I lost like 10 pounds”), out at $15 as “an unknown entity out of Phoenix with none of the hallmarks of success,” opening at $12, with no one wanting to buy it, down to $8, TV anchors asking “Why do investors not believe in you?” Then the text message: “we have to pull everyone into a room” — and the all-hands where employees decided whether they believed the answers. “You’re all now a little tighter because you just fought something together.”
• The structural point: “As a private company, what do you do when you go through a tough time? You don’t talk about it because there’s no forcing mechanism… you get to avoid building resilience.”
• His pricing heresy, delivered knowingly: “price it as high as you can… We went out at 15 bucks. We opened at 12. It’s a complete disaster. We’re down to eight a week later. Do I remember? Who cares? No one cares.” But if we’d priced it at 10 or 8 to get some magical bump, we would have gotten a lot less money, been more likely to not make it across the line, and taken more dilution.

9. Debt was dangerous — and an accidental pressure machine

• He doesn’t sanitize it: “debt is dangerous in general and more dangerous for a company that’s not positively cash flowing.” But in Carvana’s case it worked out — equity raised at the time of the debt would have meant “quite a bit more dilution” than taking debt and paying it off later.
• The unplanned benefit: debt “was almost like being super public.” Once the market shut off funding, they had to cover interest, not just reach break-even — “there’s no substitute for huge cash flows that have to be covered and a market that’s telling you you’re stupid every day.” He’s careful about the credit: “that wasn’t our story. We don’t get to take credit for that.”
• The Hindenburg-era lesson, stated as a permanent weakness: “the market can be more volatile than I ever thought,” so build in resilience — people with reps from hard days survived together, and “a little bit stronger balance sheet than you think you need… a lesson that I’ll never fully internalize,” because his instinct is to take as little dilution as possible. And the mindset that goes with it: “no one’s ever going to clap for you. If you allow yourself to need outside validation, you’re putting yourself in a weak spot.”

10. Beat dad, chase expensive dopamine, raise kids on sport — then telescope to 3 million cars

• He prefaces every self-analysis with a warning: “any questions you ask any person about themselves are subject to the story they tell themselves.” That said: his father’s personal bankruptcy (hidden from him until 10 or 12) taught him “he was successful and he was human” — and the honest motive was not to please or impress but “I’m positive that it was beat. If he can do it, I can do it and I can do it better.” The competitiveness is visceral: after high school football losses he’d sit alone in the back of the bus and cry “a tough cry.”
• Against Harry’s cynicism about the Steve Jobs everyone-can-do-it quote, Garcia sides with luck: “If I got to live a thousand lifetimes, I don’t think I’d be part of something as big as Carvana in very many of those lifetimes… for every Steve Jobs there’s probably a hundred people of similar quality that haven’t had anywhere near the same success.”
• His happiness framework: “There’s cheap dopamine and there’s expensive dopamine. Stock goes up, cheap dopamine… Put yourself in a fight and lose and put yourself back in a fight and win” — that’s the deeper contentment worth chasing. On raising hungry kids in wealth (“brutally hard”): children “have never listened to their parents, but they’ve never failed to observe what they do,” and sport is his best answer — “someone wins and someone loses every single time and it doesn’t matter where you started.”
• The forward look, via his changed mind on focus: “every 3 months you have to uncomfortably take things off of your board” — after two years focused on nothing but profitability, Carvana is shifting finite focus toward growth and foundation-laying, wary of “and”-ing into a million projects. On dealership purchases: “extremely early… we should save our words.” On scale: roughly 400,000 cars against a 40-million-unit US used-car market, infrastructure that “clearly supports 3 million,” and hopes for “many millions” — because “what looks possible is always a multiple of where you are today,” and where you are keeps changing.