A Convicted Felon's Son Built a $75 Billion Empire From Used Cars. The Stock Went to $3. Then It Came Back.

In Nashville, Tennessee, there is a five-story glass tower filled with used cars. You walk up to it, insert an oversized commemorative coin into a slot, and a car descends through the building on a conveyor system — like a vending machine, except instead of dispensing a bag of Doritos, it dispenses a 2021 Honda Accord.

The tower is beautiful. It is transparent. It is engineered to look like the future of automobile retail — clean, digital, frictionless. You can see every car inside. There are no dark corners, no hidden lots, no fast-talking salesmen in cheap ties trying to sell you on the undercoating package.

This is Carvana's brand promise: total transparency.

Here's the thing about glass towers. They let you see everything inside. But they don't show you who built them, or why, or what happened to the money.

Because the man who controls Carvana — who holds 80% of its voting power without any official role at the company — is a convicted felon who pled guilty to bank fraud during one of the largest financial scandals in American history. His son, the CEO, built the company with $100 million in seed money from dad's other business. The stock hit $377. Then it fell 99%, to $3.55. Four thousand people lost their jobs — some of them via Zoom. Analysts said it was dead. Competitors who tried the same model actually died. Then, in what might be the greatest comeback in the history of Wall Street, the stock climbed to $487. A 137x return from the bottom.

Two separate short-selling firms have published reports alleging that the father-son duo is running an accounting scheme. The SEC has issued a subpoena. The company says everything is fine.

The ticker is CVNA. The market cap is $75 billion. And the question of whether this is the most improbable redemption story in financial history or the most elaborate glass tower ever built — well, that depends on which side of the glass you're standing on.

Before we talk about Carvana, we need to talk about a $30,000 partnership that somehow obtained a $30 million line of credit. Because that's where this story actually begins.

Ernest Garcia II was a Tucson-based real estate developer in the 1980s. In February 1987, he obtained a $30 million line of credit from American Continental Corporation, the parent company of Charles Keating's Lincoln Savings & Loan Association. If that name rings a bell, it should — Lincoln Savings was at the center of one of the largest financial scandals of the 20th century, a collapse that cost American taxpayers approximately $3.4 billion and spawned the "Keating Five" political scandal that nearly ended the careers of five United States senators.

Garcia's role in this particular disaster was specific and creative. He was supposed to use the credit line to purchase 1,000 acres of land in Hidden Valley, outside Phoenix. The purchaser of record was an entity called West Continental — a limited partnership with a net worth of $30,000 and total assets of $87,000. Federal S&L examiners would later describe the transaction as a "sham," "concocted to allow Lincoln to report a phony profit on the land sale and camouflaged with a phony buyer."

Let me make sure you caught that. A partnership worth thirty thousand dollars was used to execute a thirty million dollar real estate transaction so that a savings & loan could book a fake $11 million profit. The ratio of actual net worth to transaction value was roughly 1:1,000. That's like using a Hot Wheels car as collateral for a Lamborghini dealership.

After the land purchase closed, Garcia drew $19.6 million of the credit line for his own purposes. Lincoln recorded an $11 million profit on the deal. Regulators called it what it was.

Garcia filed for personal and corporate bankruptcy. In a 2013 SEC filing, he stated that he had pled guilty "due to financial pressure and the chance to only receive a $50 fine in exchange for his cooperation."

And then he did something that makes the entire rest of this story possible.

He got back up.

In 1990 — the same year he pled guilty to fraud — Garcia formed a company called Duck Ventures, Inc. and purchased the assets of a bankrupt car rental company called Ugly Duckling Rent-A-Car. He bought it for under $1 million. Then he pivoted the business from renting cars to selling and financing used cars to subprime buyers — people with poor credit histories who couldn't get loans from traditional dealers.

In 1996, he took the company public on the NASDAQ under what might be the most honest ticker symbol in the history of financial markets: UGLY.

By 2002, Garcia and his team took Ugly Duckling private again, renamed it DriveTime Automotive Group, and set about building it into one of the largest used car dealer networks in the southwestern United States. At formation, DriveTime operated 76 dealerships across eight states.

Today, DriveTime generates approximately $3.9 billion in annual revenue. Ernest Garcia II owns 75% of it. He is the richest person in Arizona. He has been ranked as high as 95th on the global billionaire list and 13th in the United States.

From felony bank fraud to the richest man in Arizona. From a $50 fine to a multi-billion dollar empire. From ticker symbol UGLY to — well, we're getting there.

Ernest Garcia III graduated from Stanford with a degree in Management Science and Engineering in 2005. He did a stint at RBS Greenwich Capital in the principal transactions group — the kind of job where you learn how money moves through systems designed to be incomprehensible to outsiders. He then joined his father's company, DriveTime, in 2007 and worked there until 2012.

Five years inside the used car business gave Garcia III something that most Silicon Valley founders don't have: an intimate understanding of why the used car market is simultaneously enormous, essential, and universally despised.

The U.S. used car market is somewhere between $800 billion and $1.9 trillion, depending on how you measure it. Americans buy roughly 40 million used cars every year. It is one of the largest retail categories on Earth. And yet, in 2012, virtually none of it happened online. You still had to walk into a dealership, deal with a salesperson whose compensation structure was designed to extract the maximum amount from your wallet, haggle over a price that bore no relationship to any known reality, sit in a financing office while someone tried to sell you an extended warranty that would never cover the thing that actually breaks, and drive away wondering if you'd just been had.

Garcia III saw this and had what startup people call an insight, which is really just a fancy word for noticing something obvious that everyone else has been ignoring.

What if you could buy a car online? All of it. Financing, trade-in, paperwork, delivery. No dealership. No salesperson. No haggling. Just pick a car on a screen, click a button, and it shows up at your house.

On November 28, 2012, Garcia III co-founded Carvana with Ryan Keeton and Ben Huston. He was named president and CEO from day one.

The company was incubated as a subsidiary of DriveTime with $100 million in seed funding from his father's company. DriveTime provided the critical early infrastructure: existing inventory, logistics networks, and vehicle reconditioning capabilities. Carvana spun out as an independent entity in 2014.

Garcia III earned the Ernst & Young Entrepreneur Of The Year award in 2016 and made Fortune's 40 Under 40 in 2017. Carvana IPO'd on April 28, 2017, at $15 per share. It closed its first day of trading at $11.10 — a 26% decline. Wall Street was not immediately impressed by the son of a convicted felon selling used cars on the internet.

That would change.

In 2013, Carvana opened its first car vending machine in Atlanta. In 2015, it opened the first fully automated glass tower in Nashville — five stories tall, holding 20 cars, with three customer delivery bays and a coin-operated dispensing system. You buy a car online, show up at the tower, insert a comically oversized coin, and watch your vehicle descend through the building like something out of a Willy Wonka fever dream for adults with car payments.

It was, objectively, a stroke of marketing genius. The vending machines served a triple purpose: permanent billboard, customer fulfillment center, and PR magnet. Every local news station in every city where a tower went up ran a story about it. "The Future of Car Buying Is Here!" The towers are eight stories of glass and steel, visible from highways, Instagrammable from every angle. Texas alone has six of them.

Today, Carvana has over 40 vending machines across the United States. But the vending machines are the spectacle. The actual business is something different — and considerably more industrial.

Carvana buys used cars from consumers, wholesale auctions, and trade-ins. It reconditions them at Inspection and Reconditioning Centers using an AI system called CARLI, which the company says has reduced reconditioning costs by 45%. The cars are listed on the platform. A customer buys one entirely online — financing, trade-in, title transfer, everything digital. Then Carvana delivers it to your door or you pick it up at a vending machine. You get a seven-day return policy.

In 2022, Carvana spent $2.2 billion to acquire ADESA U.S., the second-largest wholesale vehicle auction business in America — 56 sites and 4,500 employees. The plan was to convert these auction facilities into "megasites" that handle reconditioning, wholesale, and retail logistics. It would put 78% of the U.S. population within 100 miles of a Carvana fulfillment center.

The timing of this acquisition was, in retrospect, the corporate equivalent of buying a beach house the day before a hurricane.

The pandemic was the best thing that ever happened to Carvana, right up until it was the worst.

When COVID-19 hit in early 2020, the stock was trading around $22. People were locked in their homes, terrified of in-person interaction, and suddenly the idea of buying a car without ever seeing another human being was not a novelty — it was a necessity. Used car prices exploded because new car production froze due to semiconductor shortages. If you wanted a car in 2020 or 2021, Carvana's website was the most frictionless place to get one.

Revenue went from $5.6 billion in 2020 to $12.8 billion in 2021. The stock hit $377 in August 2021. Garcia III was a visionary. Carvana was the future. The glass towers gleamed.

And then the Federal Reserve started raising interest rates.

Here's what happens to a company like Carvana when rates go up. First, the used car price bubble deflates because consumers can no longer afford the monthly payments that inflated prices created. Second, the company's $7 billion in debt — accumulated during the growth-at-all-costs era — becomes dramatically more expensive to service. Third, the customers who need financing (which is most of them) start defaulting or not buying at all. Fourth, the ADESA acquisition you just closed for $2.2 billion starts looking less like strategic genius and more like the worst-timed purchase since someone bought tulip bulbs in February 1637.

The stock went from $377 to $3.55 in sixteen months. A 99.06% decline.

Let me translate that into human terms. If you had invested $100,000 in Carvana at the peak, your investment was worth $940 at the bottom. Less than a thousand dollars. Your investment had lost more value, percentage-wise, than most things that physically catch fire.

In May 2022, Carvana laid off 2,500 employees — 12% of its workforce. Some of them were notified by Zoom call. CBS News quoted a worker saying, "It was so disrespectful." In November, another 1,500 were let go. Four thousand total jobs, eliminated in six months.

By December 2022, major creditors had signed a cooperation agreement for potential restructuring — the polite corporate term for "we're figuring out how to divide up the corpse." Analysts were convinced Carvana would be forced into bankruptcy by the end of the year.

Garcia III went on Bloomberg in July 2023 and said Carvana was "never at risk of bankruptcy."

This is either the statement of a CEO who genuinely believed his company would survive, or the kind of revisionist history that sounds better in a winner's press conference than it does in the wreckage of a stock that had lost 99% of its value while creditors circled.

For context: Vroom, the company that tried to do exactly what Carvana was doing, filed for Chapter 11 bankruptcy in November 2024. Its shareholders received 7% of the restructured company. Shift Technologies, another online car seller, also failed. The entire category of "buy cars online" was dying.

Carvana was supposed to die with it.

Here is where the story gets either inspirational or suspicious, depending on your disposition.

In July 2023, Carvana reached an agreement with its bondholders — led by Apollo Global Management — to restructure more than 96% of its $5.7 billion in unsecured notes. The details matter: they exchanged existing bonds for new secured debt, reduced total principal by $1.326 billion, eliminated 83% of their 2025 and 2027 maturities, and lowered cash interest expense by $430 million per year. Carvana also sold 35 million new shares to raise about $1 billion.

At the same time, management executed what can only be described as radical operational surgery. They slashed more than $1 billion from annual expenses. They shifted the entire company philosophy from "growth at all costs" to "gross profit per unit." Every car sold had to make money. No more subsidizing growth with debt.

And then something remarkable happened: it worked.

The numbers from fiscal year 2025 are staggering. Revenue: $20.32 billion, up 49% year-over-year. Net income: $1.9 billion. Adjusted EBITDA: $2.24 billion, an 11% margin. They sold 596,641 cars. Gross profit per unit hit $7,362 in Q3 2025 — nearly triple the industry average.

The net debt-to-EBITDA ratio dropped from 17x in 2023 to 1.3x in 2025. If you don't speak finance, that means they went from drowning in debt to comfortably treading water in roughly two years.

On December 22, 2025, Carvana was added to the S&P 500.

From $3.55 to $487. A 13,617% gain. If you had invested $10,000 at the absolute bottom in December 2022, it would have been worth $1.37 million at the January 2026 peak. That is, by almost any measure, the greatest stock comeback in modern Wall Street history.

Garcia III, reflecting on the near-death experience, told CNBC: "One of our values has always been, 'We're in this together.' But that's an easy thing to write on a wall. It's a hard thing to make true. In moments like that, you get tested."

He also said: "And you've got to get a little lucky along the way."

On January 2, 2025, Hindenburg Research published a report titled "Carvana: A Father-Son Accounting Grift For The Ages."

Hindenburg, for those unfamiliar, was the short-selling research firm that had previously taken down or severely damaged companies like Nikola, Adani Group, and Block. When Hindenburg published, people paid attention. The title alone was a missile aimed directly at the Garcia family's glass tower.

The core allegations:

The last point is the one that lingers. In 1990, Ernest Garcia II pled guilty to bank fraud for helping a savings & loan report fake profits through sham real estate transactions. In 2025, his son's company is accused by short sellers of inflating profits through questionable related-party transactions with entities controlled by the father.

Pattern recognition is not proof. But it is uncomfortable.

The market's reaction to the Hindenburg report was surprisingly muted — shares fell only 1.9%. Bulls shrugged. Carvana called the allegations "intentionally misleading and inaccurate."

Hindenburg subsequently shut down. But the story wasn't over.

On January 28, 2026, Gotham City Research published its own report: "Carvana: Bridgecrest and the Undisclosed Transactions and Debts." This one hit harder. Gotham alleged that Carvana had overstated earnings for 2023-2024 by more than $1 billion; that DriveTime "fuels over 73% of Carvana's EBIT" through undisclosed subsidies; and that Carvana sells loans to Bridgecrest — a Garcia II entity — at inflated values, booking immediate profits while Bridgecrest marks those same assets down by up to 15%.

The stock fell 14.2% in a single day. It hit an intraday low representing a 22% drop.

In June 2025, Carvana had disclosed receiving a subpoena from the SEC, which the company believed "primarily relates to allegations raised by a short-seller report." They said they were "fully cooperating." Class-action investigations were triggered.

Carvana's CFO responded to the accounting concerns with: "There is no ambiguity in our accounting practices."

There is no ambiguity. Everything is transparent. Just like the glass tower.

Let's set the accusations aside for a moment and look at what is measurably, verifiably true about Carvana as a business.

Carvana operates in over 300 U.S. markets with same-day delivery in 50+. It has approximately 23,100 employees, 40+ vending machines, and 56 ADESA sites (27 of which have been converted into megasites). The company has processed over 4 million total car transactions since founding.

The revenue trajectory tells a story of breakneck growth, near-death, and recovery:

The profitability transformation is real, regardless of how you feel about the governance. In fiscal 2024, Carvana reported its first full year of GAAP profitability: $404 million. In 2025, that jumped to $1.9 billion — though it's worth noting that this figure includes a $685 million non-cash tax benefit, making underlying earnings closer to $1.2 billion.

Adjusted EBITDA hit $2.24 billion in 2025, an 11% margin. Gross profit per unit reached $7,362 — a metric that matters enormously in the used car business because it tells you how much money the company makes on each car after buying, reconditioning, and delivering it. Carvana's GPU is nearly triple the industry average.

The debt situation has improved dramatically but isn't resolved. Total debt remains approximately $4.6 to $5.1 billion. The PIK interest period — a feature of the 2023 restructuring that allowed Carvana to pay interest with more debt instead of cash — expires in mid-2026, at which point cash interest payments will jump to over $500 million annually.

Here's the number that should make you either salivate or sweat: Carvana currently has approximately 1.6% of the U.S. used car market. Management has publicly stated a "3 Million Unit North Star" — a target to sell 3 million units per year, roughly five times their current volume. If they get there with anything resembling their current GPU, the revenue math gets genuinely enormous.

Carvana's long-term target is 4% market share at a $76 billion gross profit opportunity, aiming for a 13.5% adjusted EBITDA margin by 2030-2035.

That is either a roadmap or a hallucination. The difference depends on execution, market conditions, and whether the SEC investigation turns up anything that changes the math.

The bull case for Carvana rests on three pillars, and each of them is genuinely compelling.

First: the market is absurdly large and absurdly offline. Americans spend somewhere north of $800 billion a year on used cars. Online penetration is approximately 2-3%. If you believe that used car retail will follow the same trajectory as virtually every other retail category — from books to clothing to groceries — then the digital share will grow from 3% to 15% to 30% over the next decade. And Carvana is the only scaled player left standing. Vroom is bankrupt. Shift is dead. CarMax is trying to go digital but is fundamentally an omnichannel retailer with 200+ physical stores and all the overhead that implies.

Second: operational leverage. Carvana spent $2.2 billion on 56 ADESA sites. Only 27 have been converted into megasites. Only 15 are fully integrated into the retail network. Current capacity utilization is approximately 70%. This means Carvana can dramatically increase volume without proportional increases in capital expenditure. They've already built the machine; they just need to feed it more cars. Every incremental unit sold on existing infrastructure drops more margin to the bottom line.

The AI angle is real, too. CARLI, their reconditioning AI, has cut costs by 45%. Sebastian, their customer-facing AI agent, handles 60% of vehicle acquisition interactions and 30% of retail purchases. These aren't buzzword deployments; they're measurable cost reductions.

Third: the comeback itself is evidence of operational capability. Carvana was sixteen months from peak to bottom. It restructured $5.7 billion in debt, slashed a billion dollars in annual expenses, tripled its GPU, achieved GAAP profitability, joined the S&P 500, and produced a 137x return from the low. Whatever you think about the Garcias, the operational turnaround is among the most impressive in corporate history. The team that did this — whether through skill, desperation, or both — clearly knows how to run a car business.

JPMorgan and Wells Fargo both rate the stock "Overweight." The median analyst price target is $445. The narrative is simple: it's Amazon for cars, the TAM is enormous, the competitors are dead, and the worst is behind them.

The bear case for Carvana is not about whether the business is real. Revenue is real. Cars are real. Customers are real. The bear case is about whether the profits are what they appear to be, and whether the governance structure should make you nervous even if everything else checks out.

Start with the family. Ernest Garcia II — convicted felon, no official role at Carvana — controls approximately 80% of voting power through dual-class Class B shares that carry 10x the voting weight of public Class A shares. He cannot be fired. He cannot be overruled. Public shareholders have, for all practical purposes, no say in how the company is governed. And since 2020, he has sold more than $5 billion worth of Carvana stock.

Read that again. The controlling shareholder, who has no job at the company, has extracted five billion dollars from the stock while retaining voting control. Between August 2020 and August 2021, he sold $3.6 billion. In the year after he stopped selling, the stock plunged 99%. Then it recovered, and he sold another $1.4 billion.

Hindenburg pointed out a symmetry that is hard to ignore: "Ernest Garcia II pled guilty to felony bank fraud, helping a savings & loan generate fake profits through sham real estate transactions. Decades later, short sellers allege his son's company is generating inflated profits through questionable transactions with entities controlled by the father."

The related-party transactions are the crux. Carvana and DriveTime/Bridgecrest — entities controlled by Garcia II — are deeply intertwined. Carvana sells loans to Bridgecrest. Bridgecrest services Carvana loans. Extended warranty reimbursements flow between the companies. Gotham City Research alleged that DriveTime "fuels 73% of Carvana's EBIT" and that Bridgecrest marks down the loan assets it buys from Carvana by up to 15% after purchase — meaning Carvana books inflated profits on the sale that Bridgecrest subsequently writes down.

If true, a significant portion of Carvana's reported profitability is an optical illusion created by transactions between the son's public company and the father's private one.

Carvana says all related-party transactions are "properly disclosed." The SEC subpoena is active. There is, as the CFO says, "no ambiguity."

Beyond the governance, the financials have genuine risks. Total debt is still $4.6-5.1 billion. The PIK interest holiday ends mid-2026, at which point Carvana will need to start paying $500+ million annually in cash interest. The 2025 net income of $1.9 billion included a $685 million non-cash tax benefit — strip that out and you're closer to $1.2 billion, which is great but makes the $75 billion market cap look considerably more aggressive. The stock's 90-day historical volatility is 157%, compared to 20% for the S&P 500. This is not a sleep-well-at-night stock.

And then there's the cyclicality. Used car prices are sensitive to interest rates, consumer credit conditions, new car production levels, and tariff policy. The pandemic created a once-in-a-generation tailwind. The rate-hike cycle created a once-in-a-generation headwind. The next cycle will create something, and the only certainty is that it will be different from the last two.

Let me take you back to the vending machine.

You're standing in front of a glass tower in Nashville, or Phoenix, or Austin. Eight stories of transparent glass, filled with shiny used cars, each one inspected by an AI called CARLI and ready for delivery by an AI called Sebastian. You're holding an oversized commemorative coin — the kind of object that feels simultaneously futuristic and absurd, like something from a theme park ride about the year 2050 as imagined in 1985.

You insert the coin. A car descends.

The glass tower was designed to signal transparency. You can see every car. There are no back rooms, no hidden ledgers, no side deals in parking lots. It's the architectural equivalent of "we have nothing to hide."

But transparency is a strange thing. A glass building lets light in, but it also reflects it. And depending on where you're standing and how the light hits, what you see through the glass and what you see reflected in it can be very different things.

Here is what I can tell you with certainty. Carvana sells real cars to real customers for real money. The operational turnaround from near-bankruptcy to $20 billion in revenue is genuinely extraordinary. The competitive landscape — with Vroom dead, Shift dead, and CarMax still figuring out digital — has never been better. The AI-driven efficiency gains are measurable and meaningful. The total addressable market is genuinely enormous.

I can also tell you this. The controlling shareholder is a convicted felon who has sold $5 billion in stock. Two separate short-selling firms have published detailed reports alleging accounting manipulation through related-party transactions between the son's public company and the father's private empire. The SEC has issued a subpoena. The dual-class share structure ensures that minority shareholders cannot do anything about any of it.

Garcia III told CNBC: "In those moments, unfortunately, the reality is people will line up to kick you when you're down."

He also said the company was "never at risk of bankruptcy" — about a stock that had fallen 99%.

From $377 to $3.55 to $487. From UGLY to Carvana. From a $50 fine to the S&P 500. From a $30,000 partnership that obtained $30 million in credit to a $75 billion market cap. The Garcia dynasty has built something that is either a generational transformation of the American used car market or the most elaborate glass structure in financial history.

The coin is in your hand.

Whether you insert it or walk away is a question that no analyst report, no short seller, no SEC investigation, and certainly no article can answer for you. Because the answer requires something that Carvana's glass towers, for all their engineered transparency, cannot provide: a view of what's happening inside.

This article is for educational and entertainment purposes only and does not constitute investment advice. The author holds no position in Carvana (CVNA) or any related securities.