Carvana's Record Quarter Meets a De-Rating Market

Record units and revenue, a thinner EBITDA margin, and an agentic-commerce story the tape has not yet rewarded.

CVNA · Earnings Call · 2026-10-28

The machine is bigger; the multiple is smaller

Carvana's Q3 2026 was, by operating measures, the best quarter in company history: “This is our first quarter with over 200,000 sales. Our run rate is now 870,000, double what it was just 2 years ago.” — Ernie Garcia, President, Chief Executive Officer · 2026-10-28 Revenue crossed $8 billion, retail units rose 39% to 217,385, and adjusted EBITDA hit a record $800 million. Yet the stock has been a de-rating story: shares are roughly 38% below the January 2026 peak and down 8.9% over the last 90 trading days. The market is not disputing the machine. It is disputing the price it should pay for it. The price-to-revenue multiple sits near 1.9x, down about 87% year over year; price-to-operating income is about 21x, a fraction of last year's peak. That is the central tension of this report: record operating results, compressed equity multiple. The margin line explains why. “Adjusted EBITDA margin was 10%, a decrease from 11.3%, primarily driven by selling a mix of higher-priced vehicles and giving back customers in other GPU.” — Mark Jenkins, Chief Financial Officer · 2026-10-28 Revenue grew 42%, but gross margin in the prior quarter was already 18.8%, down 3.2 percentage points year over year, and operating margin was 9.2%, down 1.3 points. The company is scaling; the P&L is also showing the cost of scaling.

Rates, fuel, and the cost of giving gains back

The fresh company keywords this quarter were rising rates, gains to customers, and fuel costs—a very different mix from the wholesale-spread and tax-season vocabulary of earlier calls. Management was explicit that the macro moved against them. “This quarter ... gas prices and in interest rates ... impacted us to the tune of about $200 per unit.” — Ernie Garcia, President, Chief Executive Officer · 2026-10-28 Mark Jenkins described a deliberate pass-through of benchmark-rate increases to customers, with lags; that is why other GPU was flat sequentially despite higher benchmark rates but down $299 per unit year over year. In plain terms, Carvana chose to protect the customer offer rather than the near-term margin line. That is consistent with years of rhetoric about fundamental gains, but it is also why the EBITDA margin slipped even as dollars hit a record. The macro is shared—Fed rate hike was a top global keyword this quarter—but Carvana's transmission channel is consumer APR, not deposit costs. This is also a reversal from the prior quarter's cost tailwind. In April, Jenkins could say “operations expenses down slightly year over year” — Mark Jenkins, Chief Financial Officer (CFO) · 2026-04-29; in Q3, operations expense per unit rose $109 on higher fuel and logistics costs. Overhead leverage of $159 per unit and retail GPU gains partly offset it, but advertising expense also rose $45 per unit as Carvana pushed demand ahead of inventory. The company is deliberately applying pressure to its own system. The question for investors is whether this is temporary macro noise or a new cost baseline. Management argues the former: “those things tend to zero out,” Ernie Garcia said, with the path to 13.5% EBITDA margin still clear.

The agentic-commerce option: new label, old ambition

The most interesting new language was Agentic commerce, which jumped into the company's top keywords this quarter. Garcia framed AI agents as a way to reduce digital friction to zero while leaving physical friction unchanged—a structural advantage for a vertically integrated retailer with deterministic APIs and its own logistics.

We think that we’ve been positioning the business for that kind of a shift for 14 years ... it pushes people more in the direction of digital transactions ... reduce digital friction to zero while leaving physical friction largely unchanged.
Ernie Garcia, President, Chief Executive Officer · 2026-10-28

The phrasing is new; the ambition is not. In February, Garcia already called Carvana “positioned to be an AI winner and not something that is disrupted by AI,” and in October 2025 he described “ambient agents” writing and deploying code. So this is an escalation of a multi-quarter narrative, not a fresh strategic pivot. The market has not yet put a separate value on it—there is no monetization guidance, and analysts are still asking what incremental investment is required.

What management stopped quantifying

Two silences matter. First, reconditioning centers remain the binding constraint. Carvana produced a record 200,000-plus units in the quarter and slightly outproduced sales, but immediately available inventory still grew much slower than sales. The language is almost identical to prior calls: reconditioning is the hardest operational problem, and scaling it at 39% growth is difficult. Second, the franchise-dealer/Stellantis experiment got another polite non-answer. A prior call said “It is still early. Stay tuned. We will share more when it is time to share more.” — Ernie Garcia, Chief Executive Officer (CEO) · 2026-04-29 This quarter: “we don’t have a ton more detail to add.” That is not a negative by itself, but it explains why the “new cars” keyword faded after being a hot topic earlier in the year. The other signal is the 2027 reporting change. “As we become more profitable on a GAAP basis, I think the value of adjusted EBITDA goes down.” — Mark Jenkins, Chief Financial Officer · 2026-10-28 That is a confident thing to say if you believe the earnings are durable. It also shifts the market's attention to GAAP operating income and net income, where Carvana's Q3 print was strong: $714 million of operating income and $483 million of net income. The balance sheet is improving too: net debt to trailing EBITDA of 0.8x, EBITDA to interest of 6.9x, and interest coverage around 6.7x in the latest fundamentals. The refinancing of $1.7 billion of senior secured notes into a Term Loan B saves about $45 million a year. Carvana is executing the operating plan, but the equity is no longer paid for execution alone. The stock's drawdown says investors want proof that the margin compression is investment, not deterioration, and that agentic commerce is more than a compelling narrative. Q4 guidance—sequential unit and EBITDA growth, with the high end of the $2.7–3.0 billion EBITDA range—gives them a near-term test. If rates and fuel normalize and reconditioning catches up, the “machine” story wins. If not, the multiple stays compressed.