Carvana Drops 20% After FOIA Documents Reveal Family Affiliate Burned $1 Billion While Public Company Reported Profits

Carvana didn’t just wobble on January 28, 2026. It fell off a cliff. The stock dropped about 20%, finished at $410.04, and even dipped to $375 intraday. And no, this wasn’t your typical “short seller screams, stock bleeds” moment.

What set it off was a Gotham City Research report that claims it obtained audited financial statements for DriveTime and Bridgecrest. Both are private companies controlled by Ernie Garcia II. Gotham says it pulled the documents through FOIA requests, then laid them side-by-side with Carvana’s public numbers. The headline clash: the private affiliate shows roughly $1 billion in cash burn while the public company talks up record profitability. That mismatch creates what forensic-minded investors call a “reverse pyramid,” where the public roof looks shiny while the private foundation crumbles.

The Mechanism: Where Value Goes to Die

Here’s the core accusation, and it’s not abstract. Carvana booked $755 million of gains on loan sales in 2024, which juiced EBITDA right away. Meanwhile, Bridgecrest, the Garcia family’s loan-servicing arm, allegedly charges an implied servicing fee of about 0.12%. Market rates run closer to 2.0% to 2.9%. Think of it like selling a car at “full price” after quietly giving away the warranty, the wheels, and the engine service for pennies.

At the same time, DriveTime marked down its loan portfolio by about $900 million. Yet Carvana booked gains tied to those same assets. If you’re wondering how both things can be true, that’s the point. Gotham says they shouldn’t be, at least not without airtight explanations.

The report also points to specific reconciliation gaps. Carvana reports $193 million in commissions earned, but DriveTime shows $205 million paid. Deferred acquisition costs differ by $580 million across schedules. These aren’t rounding errors. In plain terms, they suggest the public-facing business may be recognizing income the broader Garcia-controlled ecosystem hasn’t truly earned.

Then there’s GoFi LLC, a conduit entity Gotham says is 100% owned by Garcia II. It reportedly generates 97% of revenue from “gains on sale” to related parties. If that sounds like a device meant to scrub fingerprints, you’re not alone. The allegation is that loans move Carvana → GoFi → Bridgecrest. Carvana can book a “sale,” but the credit risk ultimately circles back onto the family balance sheet. Like passing a hot potato around the dinner table, then pretending it’s gone because you switched hands.

The Forensic Question That Changes Everything

Most short reports argue about interpretation. This one goes straight for the throat: where does the value actually get created, and who’s footing the bill?

Gotham claims DriveTime burned more than $1 billion in operating cash flow from 2022 through 2024. It also says leverage ran 20x to 40x EBITDA, a level that makes an entity feel less “highly levered” and more “functionally insolvent.” Interest coverage reportedly sank to about 0.5x to 1.0x. Translation: operations don’t comfortably pay the interest. So how does it survive? Gotham points to $1.4 billion of new debt issuance.

And while that’s happening, Gotham says Garcia II pulled $352 million in cash distributions. Picture a sinking ship where the captain grabs the lifeboats first. That’s the dividend recapitalization argument in a distressed-looking structure.

Liquidity comparisons add more bite. Gotham highlights DriveTime at roughly 0.7% of revenue in available liquidity versus CarMax at 9.7%, plus the private entity has no corporate revolver. If loan quality slips, funding can vanish fast.

This architecture matters because Carvana’s gain-on-sale allegedly makes up about 55% of EBITDA. If related-party pricing gets reset by auditors, regulators, or counterparties simply refusing to play along, that earnings stream doesn’t fade. It disappears.

Three Regimes, Not a Cartoon “Bull vs Bear”

If you’re trying to trade this like a morality play, you’ll get whiplash. The cleaner way is to think in regimes:

Regime 1: Cosmetic. Disclosures tighten, presentation changes, but the economics hold. Multiples compress because investors slap on a quality discount. The stock bounces, though not all the way back.

Regime 2: Structural. Loan-sale gains reset lower after auditors or counterparties force more conservative pricing. The story shifts from “turnaround” to “structured finance machine under scrutiny.” Multiples get cut in half even before any restatement. Gotham frames this as the base case.

Regime 3: Break. Funding closes at DriveTime or Bridgecrest, or a legal or accounting event hits. At that point, the market stops caring about EBITDA and starts pricing liquidity and governance risk. Gotham’s $0 to $100 target range lives here.

What Would Confirm or Crush Gotham’s Case

If you want high-signal tells, Gotham points to a few: changes in Carvana’s related-party disclosures, shifts in auditor posture (Grant Thornton audits all three entities), and timing risk around the February 18 earnings report. Gotham even calls out the possibility of 10-K delays.

If you’re bullish, the bar is simple but brutal. Don’t just wave your hands. Bring a numbers-first rebuttal that explains why DriveTime looks catastrophic while Carvana looks brilliant, using counterparty-level and pricing-level evidence that the loan sales are real and occur at arm’s length.

Why the Market Could Get Nasty Fast

The setup can turn violent. Short interest sits around 9% of float. Options implied volatility is above 70%. Yet borrow costs are still under 1% annualized, so shorts aren’t getting mechanically squeezed out. That mix tends to create chop, air pockets, and sudden drops, not neat V-shaped reversals.

The edge here isn’t guessing direction. It’s recognizing the next 2 to 6 weeks are loaded with catalysts: Carvana’s response, the February 18 earnings release, and the possibility of regulatory attention. The FOIA angle makes this feel less like opinion and more like a paper trail, because Gotham claims it’s working off audited third-party statements that contradict the public narrative.

If you manage long-only money, treat this like a governance and earnings-quality problem. Size accordingly until an audited rebuttal shows up. The real question isn’t whether Gotham is “good” or “bad.” It’s whether Carvana’s earnings engine leans on related-party architecture that markets won’t reward anymore without extreme transparency.

NOT INVESTMENT ADVICE