Carvana Came Back From the Dead. Its Balance Sheet Didn't.

---

Begin, as honesty demands, with the genuine achievement, because Carvana's recovery is not a mirage and a
forensic piece that denied it would forfeit its credibility. In the first quarter of 2026 the company reported
revenue of $6.43 billion, up 52% from a year earlier; net income of $405 million; retail sales of 187,393
vehicles, up 40%; gross profit of $1.27 billion at roughly $6,783 per unit; and adjusted EBITDA of $672 million.
These are real numbers describing a real, profitable, fast-growing business. Management took a company that was
hemorrhaging cash and staring at insolvency, cut costs ferociously, renegotiated its debt, restored unit
economics, and returned it to growth and profitability. As an operational turnaround, it is genuinely
impressive, and the executives who pulled it off deserve the credit the market has given them.

So this is not the story of a fraud being exposed or a business that does not work. Carvana sells a lot of cars,
efficiently, and makes money doing it. The story is subtler and, for an investor, more important: it is about
the difference between fixing a stock and fixing the risks that the stock embodies — and about how a
resurrection narrative so emotionally satisfying that it has become Wall Street legend can cause the market to
stop pricing the very fragilities that nearly destroyed the company three years ago, and that have been reduced
but not removed.

This essay walks through what the $87 billion valuation assumes has been solved, and asks, item by item, whether
it actually has.

The debt is smaller, and it is still the story

Start with the thing that nearly killed Carvana the first time: debt. The 2022 near-death experience was a
balance-sheet event. Carvana had borrowed enormously to fund its growth, and when used-car prices fell and
interest rates rose and the credit market seized, the company could not service or refinance its obligations on
acceptable terms, and the equity nearly went to zero. The lesson the market took from the recovery is that this
problem has been solved. It has been reduced. That is not the same thing.

Carvana today carries roughly $5 billion of total debt — around $4.8 billion net of cash — and it is rated junk
by the credit agencies. That is a materially better position than the catastrophe of 2022, achieved through a
painful 2023 debt restructuring, and the bulls are right to credit the improvement. But $5 billion of junk-rated
debt is not a trivial encumbrance for a company whose underlying business is selling used cars, an industry with
thin structural margins and acute sensitivity to the economic cycle. More pointedly, analysts have flagged an
approaching "interest cliff" in mid-2026 — a point at which the favorable terms of the restructuring give way to
higher cash interest obligations, increasing the burden the business must carry. The debt did not disappear in
the resurrection. It was rescheduled, and some of the reckoning was pushed into the future, and the future has a
way of arriving.

The deeper point is structural: a business that nearly died of leverage and remains meaningfully levered has not
escaped the condition that nearly killed it. It has improved its odds of surviving the next downturn, not
removed its exposure to one. And the $87 billion equity valuation prices a company that has put leverage risk
firmly behind it, when the balance sheet says it has merely managed leverage risk down to a level that is
survivable in good conditions.

The profit engine is a subprime loan machine

Now to the part of Carvana that the "online used-car retailer" description obscures, and that matters more than
any other single fact about the company: a large and critical share of Carvana's profitability does not come
from selling cars. It comes from selling the loans that finance the cars — and many of those loans are
subprime.

Here is how the machine works. When Carvana sells a car, it frequently also originates the auto loan for the
buyer. It then sells those loans — packaged into securitizations and through arrangements with partners like
Ally Financial — to investors hungry for the yield. The gain on selling those loans is a significant component
of Carvana's gross profit per unit, the very metric the bulls celebrate. This means Carvana is not purely a
retailer; it is, in substantial part, a subprime auto-finance operation with a car-sales business attached. And
a subprime lender's fortunes are governed by two things that have nothing to do with how many cars it sells: the
credit quality of its borrowers, and the willingness of the market to keep buying its loans.

Both are flashing yellow. Subprime auto delinquencies across the industry have hit record levels — one figure
cited puts subprime delinquencies near 6.9%, the kind of reading associated with genuine consumer stress among
exactly the borrowers Carvana serves. Bearish analysts allege that Carvana's own borrower defaults run well
above industry norms, a claim the company disputes but which speaks to the central question: how good are the
loans Carvana is originating and selling? Because if the credit quality is deteriorating, then the gain-on-sale
profits flattering today's earnings are borrowing from tomorrow's losses, and the buyers of those loans will
eventually demand worse terms or stop buying altogether. A retailer's earnings are stable. A subprime lender's
earnings are a bet on the credit cycle, and the credit cycle is the one thing no management team controls.

The lifeline with a renewal cliff

The willingness-to-buy-the-loans problem deserves its own scrutiny, because it is the exact mechanism that
nearly killed Carvana in 2022 and it has not been eliminated — it has been outsourced to a partner whose
commitment expires.

Ally Financial has been a crucial funding lifeline, committing to purchase billions of dollars of Carvana's
automotive finance receivables — a recent arrangement covered up to $4.0 billion of loans. This is what allows
Carvana to keep originating and offloading loans at scale: someone has to be willing to buy them. But that
commitment is not permanent, and analysts have flagged an Ally renewal cliff around October 2026 — a point at
which the arrangement must be renewed, renegotiated, or replaced. If Ally were to reduce its commitment, demand
better terms, or walk away — for reasons related to Carvana's loan quality or simply to Ally's own appetite for
subprime auto risk in a deteriorating cycle — Carvana would face the same fundamental problem it faced in 2022:
a loan-origination machine that depends on someone downstream being willing to take the loans, and a sudden
narrowing of that willingness.

This is the recurring vulnerability of any business built on originating assets to sell rather than to hold. It
runs beautifully as long as the buyers show up, and it stops the instant they hesitate. Carvana's 2022 crisis
was, at its core, a moment when the buyers hesitated. The recovery has restored the buyers' appetite, but it has
not changed the structural dependence on it — and the calendar contains a specific date on which that dependence
will be tested.

The family that keeps selling

There is a pattern in Carvana's ownership that any forensic reader is trained to notice: the people who know the
company best have been persistent, large-scale sellers of its stock. Carvana was founded and is run by Ernest
Garcia III, and its largest shareholder is his father, Ernest Garcia II. Across the company's history, the
Garcias have sold extraordinary quantities of stock — by various accounts, billions of dollars' worth during the
share-price surges, including a reported $3.6 billion sold between 2020 and 2021, and further large sales as the
stock recovered to its current heights.

Insider selling is not, by itself, proof of anything; founders diversify, and large holders sell for many
legitimate reasons. But the scale and persistence of the Garcia family's selling, concentrated into the periods
of greatest share-price euphoria, is the kind of pattern that warrants attention rather than dismissal. When the
people with the most complete information about a company's true condition are converting their paper wealth into
billions of dollars of cash at every peak, an outside investor paying the peak price is, at minimum, on the
opposite side of a very well-informed trade. The bulls explain this away as ordinary diversification by a family
whose wealth is overwhelmingly concentrated in one stock. The skeptics note that "the insiders are selling
billions at the top" is a sentence that has preceded a great many disappointments.

The allegations, clearly labeled

It would be incomplete to discuss Carvana without noting that it has been the target of detailed bearish
research, and intellectually honest to label those claims precisely as what they are: allegations, disputed by
the company, not adjudicated findings. The short-seller Hindenburg Research published a report alleging, among
other things, accounting manipulation, lax loan underwriting, and approximately $800 million in loan sales to a
suspected undisclosed related party — framing the company, in its characteristically incendiary style, as a
"father-son accounting grift." Carvana has rejected these allegations.

A reader should weigh such reports carefully and skeptically — short-sellers are talking their book, and their
most dramatic claims are not always borne out. But the existence of detailed, specific allegations about related-
party loan sales and underwriting quality, layered on top of the documented facts of heavy insider selling, a
subprime-dependent profit model, and junk-rated leverage, is itself part of the risk profile. The point is not
that the allegations are true; it is that an $87 billion valuation prices the company as though such questions
do not exist, when they manifestly do, and when the burden of disproving them falls on a company whose largest
shareholder has a complicated financial history of his own. The market is entitled to give Carvana the benefit
of the doubt. It is not entitled to forget that the doubt exists.

Eighty-seven billion dollars for a used-car dealer

Put the valuation in perspective, because the number is the whole argument and it is easy to lose its scale amid
the resurrection romance. Carvana commands a market capitalization of roughly $87 billion. That is a staggering
sum for a used-car retailer, and the cleanest way to feel it is by comparison to CarMax — the largest and
most established used-car retailer in America, a profitable, decades-old operator with its own enormous sales
and financing business. Carvana, younger and far more leveraged, trades at a valuation that dwarfs CarMax's
on most measures, and at an earnings multiple that would be aggressive for a software company, let alone for a
business whose core product is selling depreciating metal at thin margins.

The bull justification is growth: Carvana is taking share and growing units at 40%, so it deserves a growth
multiple. But notice what that argument quietly assumes — that the growth is durable, that the unit economics
hold, that the financing gains persist, and that the credit cycle stays benign — every one of which is a
contingency this essay has questioned. A used-car retailer is, at the end of the day, a capital-intensive,
cyclical, low-margin business operating in a fragmented and competitive market. Dressing it in a technology-stock
multiple does not change its underlying nature; it merely raises the price of being wrong about it. At $87
billion, there is no margin of safety: the valuation already credits years of flawless execution, benign credit,
continued funding access, and durable hyper-growth. Anything less than all four, and the multiple has a very
long way to fall — as it demonstrated, in the other direction, in 2022.

What a recession does to this machine

The single most important question for Carvana is one no management team can answer: what happens to this
business in a genuine economic downturn? Because every element examined here — the leverage, the subprime loan
book, the funding dependence, the cyclical product — is a fair-weather strength that becomes a foul-weather
vulnerability, and they are all correlated, meaning they would deteriorate together.

Trace the chain. In a recession, used-car prices fall, compressing the gross profit on every vehicle and
impairing the value of inventory. Simultaneously, unemployment rises, and the subprime borrowers who make up a
large share of Carvana's loan book begin to default at higher rates — defaults that are already, by some
accounts, running hot in a still-healthy economy. As loan losses mount, the investors and partners who buy
Carvana's securitized loans demand higher yields or pull back entirely, exactly as they did in 2022, choking the
gain-on-sale profits that flatter earnings and straining the company's ability to fund new originations. And all
of this lands on a balance sheet still carrying roughly $5 billion of junk-rated debt with a cash-interest
burden set to rise. Falling margins, rising loan losses, retreating funding, and a heavy debt load — arriving at
once, because a recession triggers all of them simultaneously — is a precise description of the conditions that
nearly bankrupted Carvana three years ago.

The company is unquestionably better prepared for that scenario than it was in 2022: more profitable, somewhat
less leveraged, operationally sharper. But "better prepared for the thing that nearly killed us" is not the same
as "no longer vulnerable to it," and an $87 billion valuation embeds the latter while the balance sheet supports
only the former. The resurrection proved Carvana can survive a near-death experience. It did not prove the
company has stopped being the kind of business that has them.

What the bulls get right

In fairness, the bull case has real substance beyond mere momentum, and the skeptic must concede it. Carvana's
operational improvement is genuine and large: it has demonstrably lowered its cost to sell a car, improved its
logistics, grown units at 40% while expanding gross profit per unit, and generated real adjusted EBITDA and
positive net income — this is not a cash-burning story like the EV startups elsewhere in this series. The
used-car market is enormous and still largely offline, giving Carvana a genuine long runway to take share from
fragmented traditional dealers. Management has proven, under maximum duress, that it can execute and survive. And
the deleveraging, while incomplete, is real progress from the abyss of 2022. A bull can reasonably argue that a
proven operator with a structural advantage in a vast market deserves a premium, and that the subprime and debt
risks, while real, are manageable in a non-recessionary environment.

The honest synthesis is that Carvana is a genuinely good operating business wrapped around a genuinely fragile
financial structure, trading at a price that credits the former and discounts the latter. The turnaround is
real; the risks are also real; and the $87 billion valuation has decided that the turnaround has retired the
risks. It has not. It has subordinated them to good conditions — and good conditions, for a levered, subprime-
exposed, funding-dependent business, are precisely the thing that does not last forever.

The kicker

The most dangerous stories in markets are not the frauds, which eventually collapse, but the genuine comebacks,
which seduce by being partly true. Carvana really did come back from the dead, and that fact is so emotionally
powerful — the left-for-dead stock that rose a hundredfold, the management team that refused to die — that it
has overwhelmed the market's memory of why the company nearly died in the first place. The debt is smaller but
still junk. The profit still leans on subprime loans sold to buyers who can stop buying. The funding lifeline
still has a renewal date. The founding family still sells. None of these is a prediction of doom; all of them are
the same vulnerabilities that, in a worse environment, took the stock to four dollars. The resurrection fixed the
price. It did not fix the machine.

The stock came back from four dollars to eighty-seven billion, and somewhere in that ascent the market decided
the near-death had been a bad dream rather than a warning. The debt is still there. The loans are still subprime.
The lifeline still expires. And the family, as ever, is still selling into the cheering — which is the one detail
from the first act that the second act has not bothered to change.