I'll research CVNA before writing the preview. Let me gather recent earnings, filings, news, and stock price data.Let me read the most recent earnings materials and check recent news/filings.Let me get the year-ago Q2 2025 numbers for comparison, recent stock price, and any recent news.Let me check for recent news and analyst commentary via document search.I have what I need. Let me write the preview.

Carvana (CVNA) — Q2 2026 Earnings Preview

Report date: After market close, Wednesday, July 29, 2026 · Conference call to follow · Event: 2026Q2 Earnings Call

The one-line setup

Carvana enters this print with an enviable operating track record — nine straight quarters as (by its own framing) the fastest-growing and most profitable auto retailer — but with a stock that has drifted lower all year and a set of guided margin headwinds that make Q2 more about the quality of the beat than whether the headline records get set. Management has already told investors that both retail units and Adjusted EBITDA should hit all-time records in Q2; the debate is around GPU direction, the reconditioning recovery, and whether the multiple can hold.

What management already promised for Q2

On the Q1 call (reported April 29), CFO Mark Jenkins and CEO Ernie Garcia set a fairly specific bar:

Records on both units and EBITDA. "Assuming the environment remains stable, we expect a sequential increase in both retail units sold and Adjusted EBITDA, leading to all-time company records on both metrics."
Full-year 2026: "significant growth in both retail units sold and Adjusted EBITDA."

Because Q1 2026 units were 187,393 and Adjusted EBITDA was $672M, the guidance implies Q2 units above ~187k and EBITDA above $672M. For context, the year-ago Q2 2025 delivered 143,280 units (+41% YoY), $4.84B revenue, and $601M Adjusted EBITDA (12.4% margin). Simply matching the recent ~40% YoY unit cadence would put Q2 2026 around ~200k units.

Metric	Q2 2025 (yr ago)	Q1 2026 (last qtr)	Q2 2026 (implied bar)
Retail units	143,280 (+41%)	187,393 (+40%)	>187k; ~200k at ~40% YoY
Revenue	$4.84B	$6.43B	Record
Adjusted EBITDA	$601M (12.4%)	$672M (10.4%)	New record (>$672M)
Net income margin	6.4%	6.3%	—
GAAP operating income	$511M	$581M	—

The most important thing to watch: GPU, and specifically the guided year-over-year decline

This is where a "record EBITDA" quarter could still disappoint. Management explicitly guided retail GPU to fall year-over-year in Q2 (while rising sequentially), driven by three things:

~$100/unit of tariff-related benefit that boosted the year-ago Q2 and won't repeat.
Lower shipping fees and higher non-vehicle (reconditioning) costs this year — Carvana has been deliberately passing logistics savings back to customers.
~$100–$200/unit hit from narrower industry-wide wholesale-to-retail spreads, after an unusually hot Q1 wholesale market where retail prices lagged.

Garcia characterized the spread compression as transitory/seasonal ("catching up on a little bit longer lag"), so listen for whether that reversed as expected. Non-GAAP total GPU has been drifting: $7,140 (Q1'25) → $6,562 (Q4'25) → $6,911 (Q1'26). Any sequential improvement in retail GPU (Q1'26 non-GAAP retail GPU was $3,250) would be read as the wholesale/retail spread and recon costs normalizing.

The reconditioning story — the swing factor on margins

The key operational narrative: Carvana hit a "bump in recon" in Q4 2025 that pressured GPU and dinged the stock. Management spent most of the Q1 call on the recovery — centralized labor scheduling, new "Roll Call" tooling under its Carli software, and AI-driven line optimization. Crucially, they said that by early Q2, labor hours per unit were "just shy of our all-time best."

Two things to note: - Timing lag: recon cost is capitalized into inventory and only hits the P&L when the car sells ("cars carry the cost of reconditioning at the time they were produced, not when sold"). So the efficiency gains achieved in Q2 should be the quarter where the improvement actually flows through gross profit — a potential positive surprise. - Management tempered expectations, saying the full benefit "will take time" and there is still a several-hundred-dollar spread between top- and bottom-quartile facilities.

Other model watch-items

Overhead reset. Q1 overhead stepped up (seasonal SBC vesting, winter snow-removal, and AI/tech + facility investments). Jenkins framed Q1 as a "new level" rather than a one-off — so don't expect a big sequential drop; watch for creep.
Advertising. Ad spend/unit rose ~$92 YoY in Q1 ($630/unit). Management intends to keep spending given only ~2% used-car market share vs. ~20% e-commerce penetration in other retail. A step-up pressures near-term SG&A but is the deliberate growth lever.
Fuel/diesel. With fuel prices elevated (Iran-conflict backdrop cited in Q1), expect a modest logistics/cost-of-sales headwind in Q2 — management flagged it as within "normal quarter-to-quarter fluctuations."
Other GPU / financing. Carvana has been giving rate concessions back to customers (lower other GPU) offset by higher finance & VSC attach. Watch the loan-sale gain dynamics and any change in prime/subprime mix.
GAAP noise from Root warrants. GAAP net income swings with warrant fair value (a -$42M drag in Q1 vs. a +$158M gain in the year-ago period). This makes GAAP EPS volatile and less comparable — Adjusted EBITDA and operating income are the cleaner reads.

Balance sheet & capacity — the structural bull case

Deleveraging: net debt / TTM Adjusted EBITDA fell to 1.1x in Q1, described as the strongest ever, with the stated goal of reaching investment-grade credit ratios. Liquidity was ~$6.9B (cash $2.41B).
Capacity runway: current footprint (18 IRCs + 16 integrated ADESA sites) supports ~1.5M units of built-out annual capacity (only ~0.75M run-rate used), with real estate for 3.0M — the backbone of the 2030–2035 goal of 3M units at 13.5% Adjusted EBITDA margin. Expect updates on the 6–8 ADESA integrations planned for 2026 and the first full ADESA build-outs.
Optionality: continued (deliberately vague) commentary on Stellantis/new-car dealership pilots and ADESA Clear wholesale platform — Garcia has repeatedly said "it's still early."

Stock setup & sentiment

Shares closed around $66 on July 28, and the setup into the print is cautious: - Down ~17% YTD (from ~$80 to start January) and ~31% off the late-January high (~$96). - The stock derated after the Q4 2025 recon miss (mid-$70s in mid-February down toward the mid-$60s) and has been range-bound ~$60–73 for months, unable to reclaim the Q1-report level (~$79). - The recent slide from ~$70 (mid-July) to the low-$60s suggests some de-risking into the number.

Implication: Sentiment is not euphoric, which lowers the bar somewhat. But CVNA remains a premium-multiple, high-expectations name — a record EBITDA print alone likely isn't enough. The reaction will hinge on (1) units relative to the ~40% growth cadence, (2) whether retail GPU rose sequentially and the wholesale-spread pressure proved transitory as promised, (3) evidence the recon efficiency gains are flowing through margins, and (4) any raise/color on full-year 2026 EBITDA.

Bottom line — what to focus on

Units: Is ~40% YoY growth intact (~200k)? Deceleration below ~35% would be a red flag given the "demand is available" narrative.
Retail GPU trajectory: Guided down YoY — did it beat the feared ~$100–200 spread headwind and rise sequentially?
Recon flow-through: Proof that Q4's bump is fully behind them and labor-hours gains are hitting gross profit.
EBITDA margin: Q1 was 10.4% (down from 11.5%, partly optics from gross-revenue accounting). Watch the underlying margin path back toward the 13.5% long-term target.
Full-year framing & deleveraging: Any FY26 EBITDA anchor and continued progress toward investment-grade.

Net: A near-lock on headline records, but a "prove-the-margins" quarter. The asymmetry favors the stock if GPU surprises to the upside on recon/spread normalization; downside risk is a "records but soft GPU/guidance" print that confirms the market's caution.

Note: I found no company-specific news items in the last two weeks in the available news feeds, so this preview is built primarily from the Q1 2026 shareholder letter/call, the year-ago Q2 2025 release, and recent price action. Consensus estimates from sell-side were not available in the tools; the "implied bar" figures above are derived from management's own guidance and historical growth cadence.