Three years ago the dealer body was writing Carvana's obituary. The stock had collapsed, the debt looked fatal, and every 20 Group meeting had someone confidently explaining that the car business can't be done online. Nobody's laughing now. Carvana just posted its tenth consecutive quarter of industry-leading growth: 197,325 retail units in Q2 2026, up 38% year over year, $7.4 billion in quarterly revenue, up 52%, and a record $513 million in quarterly net income. Full year 2025 came in at 596,641 units and $1.9 billion in net income, and the company is guiding to $2.7 to $3 billion in EBITDA for 2026. Wall Street values the company at roughly $87 billion, more than the largest publicly traded franchise dealer groups put together.

Here's the part that should get your attention: Ernie Garcia's own framing is that at a run rate of almost 800,000 units, Carvana still holds only about 1.5% of the U.S. automotive market. They're not done. And the reason they're winning has almost nothing to do with selling cars online and almost everything to do with operating like a technology company. That distinction is the lesson, and it's stealable.

Let me make this concrete with the world I know. I've worked inside CDJR retail, so run this experiment from the customer's chair: a buyer wants a used Grand Cherokee.

At Carvana, she filters to the exact trim, spins a 360-degree tour that flags every scratch, sees one price with fees itemized, gets real financing terms against her actual credit in about two minutes, values her trade from her couch, signs online, and picks a delivery window. Total transaction time, under half an hour, at 11pm, in pajamas.

At a typical CDJR store, that same buyer submits a lead on the same Grand Cherokee and receives an email asking when she can come in. The price online may or may not survive contact with the desk. Her trade can't be valued "until we see it." Financing terms don't exist until she's in the box, and the box comes after the ups system, the appraisal walk, the manager touch, and the four-square. Four hours if it goes well.

Same vehicle. Same buyer. One of these experiences was designed; the other one accumulated. That's the whole difference between a tech company and a traditional retailer: tech companies design the experience backward from the customer, then build systems to deliver it identically every time. Dealerships inherited a process built for 1985 and bolted tools onto it.

And before the objection arrives: this is not because Carvana's people are better than your people. It's because Carvana's system doesn't depend on which salesperson answered the lead. Consistency is an engineering output, not a personnel outcome.

The tech-company mindset shows up in Carvana's operating metrics, the kind of numbers most stores never compute. Carvana tracks and publicly reports SG&A per retail unit, now $3,568 and falling, down $157 year over year, while gross per unit runs around $7,014. They know their cost to recondition, transport, and sell a car to the dollar, and they attack those numbers with software the way your desk attacks a payment objection. Their debt-to-EBITDA now sits at 1.0, the strongest balance sheet in their history, three years after near-death.

Ask yourself honestly: do you know your store's all-in expense per unit sold, by department, this month? Most GMs know gross to the penny and expenses only at the statement level, thirty days later. The first tech-company habit costs nothing: instrument the store. You cannot engineer what you don't measure.

1. Sell the transaction time, not just the car (easy). The Cox Automotive data from our Issue #3 report said it plainly: buyers don't punish long shopping, they punish long transactions, and buyers who did half their steps online saved 42 minutes and reported 82% satisfaction. Move credit application, trade valuation, and F&I product selection online, the three steps buyers told Cox they want digital and mostly can't get. You already own tools that do most of this; they're turned off or buried.

2. Make the online price the real price (cultural, not technical). Carvana's entire trust advantage rests on one promise: the number you see is the number you pay. Every time a desk re-works an internet price, the store teaches the market that its website is fiction. You don't need one-price to fix this; you need the discipline that the advertised number survives contact with the tower.

3. Photograph and merchandise like inventory is software (moderate). A 360 tour with flagged imperfections outsells a walkaround of a wet car shot on a cloudy Tuesday. Carvana treats every VDP as a product page built to close. Your used car manager should own VDP quality with the same accountability as pricing.

4. Engineer the handoff, not the heroics (hard, and worth the most). Carvana wins on consistency: the 4 millionth customer gets the same experience as the first. In a store, that means the lead answered at 9pm gets the same speed and quality as the one at 10am, regardless of who's on. That's process design plus the AI coverage we covered in Issue #2, not a pep talk at the Saturday meeting.

Balance, because this publication deals in facts, not fear. Carvana has no service drive, and fixed ops is the profit engine of the franchise system. It can't do complex trades with negative equity conversations that need a human. It has no new-car franchise, no factory relationship, no local presence when something goes wrong, and its model concentrates in late-model used. Your moats are real. But a moat protects a castle; it doesn't excuse what happens inside it. The stores that pair franchise advantages with tech-company operations are the ones Carvana can't touch. The stores that lean on the moat alone are financing Carvana's next 1.5% of market share.

The obituary crowd was wrong three years ago. Don't be the second wave of it.

Nothing in this article is investment advice, and figures are from Carvana's published financial results.

Need help getting started? I offer consulting for stores that want to close the gap this article describes. Message us on LinkedIn: Rooftop Insider

Kamil Grzych works at the intersection of automotive retail, technology, and AI. His background spans dealership sales, BDC operations, digital retail, and technology implementation across Toyota, CDJR, and Lexus stores, along with building technology startups, one exit, and more than five years evaluating businesses in a family office. He is the founder of Rooftop Insider.

Rooftop Insider covers how dealership operators actually use technology and AI across every department of the store, written from the showroom side, not the vendor side. Forwarded this? Subscribe to get every issue.

Sources: Carvana Q2 2026 and FY2025 results (Business Wire, company investor relations); CNBC; Cox Automotive Car Buyer Journey Study (Jan 2026).