CVNA · Forward model · Retail vehicle sales · Garcia 3M case

Assumptions & reasoning

• Every quarter of this history is the disclosed 'Retail vehicle sales, net' line from a Carvana shareholder letter filed as an 8-K exhibit. Nothing here is apportioned, and the fourteen quarters sum to the reported consolidated revenue with zero difference.
• Seasonality [Q1 1.00, Q2 1.02, Q3 1.03, Q4 0.95] is derived from this line's own printed quarters by centred 2x4 ratio-to-trend, normalised to mean 1.0, using the post-turnaround window from 2024 Q1 so the 2022-23 contraction does not contaminate it. Within-quarter spreads are 0.024 to 0.046, comfortably under the 0.08 amplitude, and the direction matches the 10-K's own statement that used-vehicle sales peak late in the first quarter with tax refunds and trough in the holiday quarter.
• Because the seasonal index carries the shape, the unit count in the driver is the DESEASONALISED 193,456, not the printed 197,325: the engine reapplies the 1.02 second-quarter factor itself, and entering the printed figure would count that 2% twice. The volume growth rate is likewise a deseasonalised trend, not a printed sequential step.
• Revenue per retail unit rose 17.4% year over year to $27,908 for a disclosed one-off reason - FTC guidance requiring dealers to include mandatory fees in advertised prices lifted industry retail prices, and Carvana 'followed the market on retail pricing'. That is a level shift, not a growth rate, so the price drift here is 0.4% a quarter, about 1.6% a year. A model that extrapolated 17% would be compounding a regulatory price step.
• The volume ceiling is set at 750,000 a quarter, the disclosed real estate for 3 million annual retail units, not at the 375,000 a quarter that the current built-out footprint supports. The base path's retail volume passes 1.5 million units a year during 2030 - 1.43 million in calendar 2029, 1.67 million in calendar 2030 - which is the point at which the buildout management has already begun, the first full ADESA buildout producing from early 2027, has to have delivered. Neither the cost nor the schedule of the rest of that buildout is disclosed.
• Retail's 13.22% basis margin is disclosed retail gross profit of $700M plus the $28M of depreciation and share-based compensation inside cost of sales, over $5,507M of retail revenue. The glide toward 18.0%, which reaches 16.29% by 2031 Q2, is the corporate SG&A leverage this engine cannot put in the overhead line. Read literally as gross profit it would be $4,923 a retail unit against $3,689 today, and that is not the claim: the claim is that the same unit carries less overhead.
• Capital expenditure is assigned entirely to this vertical because reconditioning capacity is what Carvana is building. 1.0% of retail revenue matches the disclosed $102M of property and equipment purchases in the first half of 2026 against $10,335M of retail revenue; the glide to 2.0% follows management's statement that it will 'continue investing in our infrastructure ahead of future growth'.