# CarMax, Inc. (KMX): Q2 FY2027 Pre-Earnings Channel Check

Canonical URL: https://ai.marketcheck.com/i/s/kmx-q2-fy2027

**Free sample**: the post-print edition, with the accuracy review against the reported results. CarMax, Inc. reported on Sep 29, 2026. Investor Desk: https://ai.marketcheck.com/investor · PDF: https://ai.marketcheck.com/api/investor/pdf/s-kmx-q2-fy2027

Not investment advice. MarketCheck is a data and analytics company, not a registered investment adviser or broker-dealer.

## What the report says

North American Vehicle Transaction Signal

- Signal: Cautiously bullish. Confidence: Medium.

CarMax sold 215,133 used vehicles in our data in the June-to-August quarter, 14.3% more than a year ago, at an average last listed price of $27,855, up 6.4%. Its share of the used vehicles we track rose 0.60 percentage points to 4.09%. The typical car sold 1.6 days faster than a year ago, and hybrids, which clear faster than any other drivetrain at 31.5 days, grew 49.5%.

Realized pricing is firmer than a year ago: the average car sold 1.10% below its opening ask against 2.10% last year, with fewer cars discounted at all. The caution sits in the unsold stock rather than in what is selling. The asking premium on unsold inventory has gone 3.8%, 4.5%, 5.8% across the three periods, the lot has moved upmarket while the fastest sold growth stays at the affordable end, and sequential days on market rose on nine of the ten largest nameplates.

Inventory is flat on the quarter, 89,014 units against 89,182 in May, and the share of stock over 90 days sits below where it was a year ago at 8.2% against 8.7%. Growth, pricing and share are positive and material. The measure that would change this call is the asking premium on unsold stock, alongside the uptick in August's discount breadth.

---

**Post-Print Version: accuracy summary added September 30, 2026 (see Section 11C)**

# PART 1: SUMMARY

## Our Overall Take

**CAUTIOUSLY BULLISH | Confidence: MEDIUM**

CarMax reports its second fiscal quarter on 29 September. Our data shows a company selling more cars at higher prices than a year ago, while the stock on its lots drifts away from what is selling.

**Volume, price and share all moved up.** CarMax sold 215,133 used vehicles in our data between June and August, 14.3% more than the same quarter a year ago and 11.8% more on the same-store postal-code panel, generating $6.05B of revenue, up 21.4%. The average selling price was $27,855, up 6.4% on the year and 1.1% on the quarter. Its share of the used vehicles we capture rose 0.60 percentage points to 4.09%. The typical car sold in 33.6 days, 1.6 days faster than a year ago.

**Hybrids are the strongest part of the book.** They are 5.06% of units, up from 3.86% a year ago, and clear in 31.5 days, faster than any other drivetrain and 2.1 days inside the company average, at a $33,219 average price, a 19% premium. Volume is up 49.5% on the year. Fully electric is the weaker half of the electrified story: share has risen to 4.00% from 2.36%, but at 35.1 days these cars sell 1.5 days slower than the company average, and the whole of that drag is Tesla, which went from 29.5 days in May to 34.8 now while its volume rose 25%.

**Realized pricing is firmer, and the caution is in the unsold stock.** The average car sold 1.10% below the price CarMax first asked for it, against 1.03% in May and 2.10% a year ago, with 27.7% of cars discounted against 42.4% last year. What is selling is going out close to its ask. The concern is what is not selling: the asking premium on unsold inventory has widened from 3.8% a year ago to 4.5% in May to 5.8% now. Alongside it, the stock has moved upmarket, with sport-utility share going from 48.4% to 56.2% over the year and the $30,000-$40,000 band from 16.1% to 26.2%, while the fastest sold growth stays at the affordable end and Toyota, the largest and most expensive volume nameplate, fell 10.1% sequentially. Sequential days on market rose on nine of the ten largest nameplates.

**Inventory is flat and the aged tail is below last year.** Stock was 89,014 units at quarter-end against 89,182 in May, a change of 0.2%, while units sold fell 1.7%. Days of supply moved from 36.7 to 37.7. The share of stock listed over 90 days rose from 5.3% to 8.2%, but May was the lowest reading in two years and August sits below the 8.7% of a year ago. Year over year the picture is a lot 19.4% larger carrying 14.1% more sales, which is a build, but a proportionate one.

Cautiously Bullish follows. Growth, pricing and share are positive and material, realized discounting is narrower than a year ago, the hybrid book is strong, and inventory is flat sequentially with an aged tail below last year's. The asking premium on unsold stock is the measure that would change this call, and Section 6 explains why it and the realized discount point different ways.

## The Quarter in Numbers

| Measure | Q2 FY2027 | Q1 FY2027 | Q2 FY2026 | vs. Q1 | vs. year ago |
|---|---|---|---|---|---|
| Used vehicles sold | 215,133 | 218,772 | 188,279 | −1.7% | +14.3% |
| Used revenue | $6.05B | $6.08B | $4.99B | −0.5% | +21.4% |
| Average selling price | $27,855 | $27,544 | $26,189 | +1.1% | +6.4% |
| Used unit share of market | 4.09% | 3.97% | 3.48% | +0.12pp | +0.60pp |
| Used revenue share of market | 4.60% | 4.60% | 3.92% | −0.01pp | +0.67pp |
| Sold days on market (mean / median) | 33.6 / 26.0 | 32.7 / 24.0 | 35.2 / 27.0 | +0.9d | −1.6d |
| Active used inventory | 89,014 | 89,182 | 74,577 | −0.2% | +19.4% |
| Active days on market (mean / median) | 38.1 / 26.0 | 33.6 / 24.0 | 37.6 / 26.0 | +4.5d | +0.5d |
| Days of supply | 37.7 | 36.7 | 36.0 | +1.0d | +1.7d |
| Stock listed over 90 days | 8.2% | 5.3% | 8.7% | **+3.0pp** | −0.5pp |
| Average discount from first ask | −1.10% | −1.03% | −2.10% | −0.07pp | **+1.00pp** |
| Share of cars discounted | 27.7% | 26.2% | 42.4% | +1.4pp | **−14.8pp** |
| Asking premium on unsold stock | 5.8% | 4.5% | 3.8% | +1.2pp | **+1.9pp** |
| Fully electric share of units sold | 4.00% | 2.99% | 2.36% | +1.01pp | +1.64pp |
| Hybrid share of units sold | 5.06% | 4.79% | 3.86% | +0.27pp | +1.20pp |
| Plug-in hybrid share of units sold | 1.62% | 1.51% | 1.11% | +0.11pp | +0.51pp |

*Market share is computed against the retail volume MarketCheck captures, not against the true U.S. market, and is not comparable to a company-reported or industry-association figure. Shares are stated to two decimal places and changes below 0.3 percentage points are not interpreted. Section 12 sets out the basis of every figure above.*

## Eight Signals at a Glance

| Signal lens | Reading | Evidence |
|---|---|---|
| Volume and revenue | Positive | 215,133 used vehicles, +14.3% year over year; revenue $6.05B, +21.4%; unit share +0.60pp to 4.09% |
| Pricing power | Positive | Average selling price $27,855, +6.4% year over year and +1.1% sequentially |
| Sell-through | Positive | Sold days on market 33.6, 1.6 days faster than a year ago; median 26.0 days from 27.0 |
| Hybrid sell-through | Positive | Hybrids 5.06% of units from 3.86%, clearing in 31.5 days, faster than any drivetrain and 2.1 days inside the company average; volume +49.5% year over year |
| Fully electric sell-through | Mixed | Share 4.00% from 2.36%, but 35.1 days is 1.5 days slower than the company average and flat sequentially; Tesla, 65% of the cohort, slowed 5.3 days |
| Inventory build | Neutral | Active inventory flat sequentially, 89,014 against 89,182, with units −1.7%; days of supply 36.7 → 37.7; over-90-day stock 5.3% → 8.2%, below the 8.7% of a year ago |
| Realized discounting | Positive | Average discount from first ask −1.10%, against −2.10% a year ago; 27.7% of cars discounted against 42.4%; depth steady at −4.98% |
| Price gap on unsold stock | Negative | Asking premium on unsold inventory 3.8% → 4.5% → 5.8%; stock mix moving upmarket while sold growth sits at the affordable end; August discount breadth up to 32.6% from 25.7% |

---

# PART 2: FULL ANALYSIS

## 0. Data Scope Notice

> **What this report covers:** MarketCheck data on used and new vehicle sales and inventory at CarMax's U.S. stores. What we cannot see: CarMax Auto Finance, its income, loan loss provisions and interest margin; extended protection plan economics; wholesale auction volumes; reconditioning and other costs; and selling and administrative expense. CarMax Auto Finance contributed $140.2 million of income and wholesale sold 162,064 vehicles in the quarter ended 31 May 2026 (CarMax Form 10-Q for that quarter). A strong or weak signal here reads on the car-selling side only and does not by itself predict total company profitability. Treat it as one tile in the investment mosaic.

## 1. North American Vehicle Transaction Signal

**CAUTIOUSLY BULLISH | Confidence: MEDIUM**

Four of the seven lenses read positive and none reads negative on volume, price or sell-through. CarMax sold 14.3% more used vehicles than a year ago, 11.8% more on the same-store postal-code panel, at a 6.4% higher average price. It took 0.60 points of share in the volume we track and moved the typical car 1.6 days faster. Hybrids, its fastest-clearing drivetrain, grew 49.5%.

One lens reads negative, and it concerns the stock that has not sold rather than the cars that have. The asking premium on unsold inventory widened from 3.8% a year ago to 5.8% now, and the composition of the lot moved upmarket over the same period while the fastest sold growth sits at the affordable end. Realized discounting went the other way: the average car sold 1.10% below its opening ask against 2.10% a year ago, with 27.7% of cars discounted against 42.4%. The aged tail thickened through the quarter, from 5.3% of stock in May to 8.2% in August, though it remains below where it stood a year ago.

Confidence is Medium rather than higher because the measures carrying the caution, the asking premium and the aged tail, are the measures that turn first, and one quarter of movement in them is not yet a trend. Section 12 states what this signal is computed from and what it does not adjust for.

## 2. Signal Table

| Signal lens | Reading | Evidence (vehicle transaction data only) |
|---|---|---|
| Volume and revenue | Positive | 215,133 used vehicles, +14.3% year over year and +11.8% same-store, against −1.7% sequentially into a seasonally slower quarter; revenue $6.05B, +21.4%; unit share 4.09%, +0.60pp; revenue share 4.60%, +0.67pp |
| Pricing power | Positive | Average selling price $27,855, +6.4% year over year, +1.1% sequentially, achieved against a narrower discount from first ask, −1.10% versus −2.10% a year ago |
| Sell-through | Positive | Sold days on market 33.6 mean, 26.0 median, against 35.2 and 27.0 a year ago. Sequentially worse, 32.7 and 24.0 in May, and up on nine of the ten largest makes |
| Hybrid sell-through | Positive | 10,877 hybrid units, 5.06% from 4.79% last quarter and 3.86% a year ago; 31.5 days, the fastest of any drivetrain and 2.1 days inside the company average; $33,219 average price, a 19% premium. Plug-in hybrids 1.62% at 37.9 days, improving from 41.7 |
| Fully electric sell-through | Mixed | 8,597 units, 4.00% from 2.99% and 2.36%; $34,473 average price. But 35.1 days is 1.5 days slower than the company average and flat sequentially, and Tesla, 65% of the cohort, went from 29.5 to 34.8 days on volume up 25% |
| Inventory build | Neutral | Quarter-end used stock 89,014, flat sequentially against sales −1.7%; days of supply 37.7 from 36.7; over-90-day stock 8.2% from 5.3%, against 8.7% a year ago. Year over year the lot is 19.4% larger on 14.1% more sales |
| Realized discounting | Positive | Average discount from first ask −1.10% against −1.03% in May and −2.10% a year ago; breadth 27.7% of cars against 42.4%; depth steady at −4.98%; discount on total asking value −1.19%, in line with the per-car figure, so the concession is spread evenly across price bands |
| Price gap on unsold stock | Negative | Asking premium on unsold inventory 5.8% from 4.5% in May and 3.8% a year ago; sport-utility share of stock 56.2% from 48.4% and the $30,000-$40,000 band 26.2% from 16.1%, while the fastest sold growth is in the cheapest nameplates; August discount breadth up to 32.6% from 25.7% |
| Margin pressure | Not observable | Gross profit per unit, reconditioning cost and selling expense are outside MarketCheck scope. Management guided full-year gross profit per unit down approximately $200 |
| Management verification | Partly observable | One of the four claims from the June call has an observable counterpart in this data; the other three concern margin, extended protection plans and reconditioning cost. See Section 11A |

## 3. Sold Volume Summary and Market Share

| Period | Used units | Used revenue | Average price | Sold days on market (mean / median) |
|---|---|---|---|---|
| Q2 FY2027 (Jun–Aug 2026) | 215,133 | $6.05B | $27,855 | 33.6 / 26.0 |
| Q1 FY2027 (Mar–May 2026) | 218,772 | $6.08B | $27,544 | 32.7 / 24.0 |
| Q2 FY2026 (Jun–Aug 2025) | 188,279 | $4.99B | $26,189 | 35.2 / 27.0 |

**Volume and revenue are up year over year and flat to slightly down on the quarter.** Used units rose 14.3% and revenue 21.4% against the same quarter last year. Sequentially units fell 1.7% and revenue 0.5%, into a quarter that is seasonally slower than March-to-May: the Q1-to-Q2 move has run −16.1%, +1.5%, −8.1% and −1.7% on units across the last four fiscal years, so this year's is the mildest of the four.

New vehicles are immaterial to CarMax and are reported here only for completeness: 559 units in the reporting quarter against 330 in May and 382 a year ago. CarMax is a used-only retailer and the new-vehicle line carries no signal.

**Market share**

| Period | Used unit share | Used revenue share |
|---|---|---|
| Q2 FY2027 | 4.09% | 4.60% |
| Q1 FY2027 | 3.97% | 4.60% |
| Q2 FY2026 | 3.48% | 3.92% |
| Change vs. last quarter | +0.12pp | −0.01pp |
| Change vs. last year | +0.60pp | +0.67pp |

CarMax took 0.60 points of used unit share and 0.67 points of revenue share over the year, reaching 4.09% and 4.60% of the volume we track. Both readings sit inside the 3.00-4.40% and 3.50-4.90% bands this ticker has run in. The sequential moves, +0.12pp and −0.01pp, are below the 0.3 percentage point floor at which we interpret a share change and are not read as movement.

*Market share is computed against the retail volume MarketCheck captures, not against the true U.S. market, and is not comparable to a company-reported or industry-association figure. Section 12 sets out the denominator and its limits.*

**Same-store approximation.** On a postal-code panel of the 244 zip codes where CarMax sold in both Jun–Aug 2026 and Jun–Aug 2025, units rose 11.8% year over year against 14.3% across all zip codes. Twelve zip codes appeared and seven disappeared. CarMax does not routinely acquire dealerships, so the panel is read as same-store rather than as store openings and closings. This is a zip-code approximation, not a certified same-store comparison.

## 4. Top Selling Makes

| Make | Units | vs. last quarter | vs. last year | Average price (last quarter) | Sold DOM (last quarter) |
|---|---|---|---|---|---|
| Toyota | 25,470 | **−10.1%** | +8.1% | $30,150 ($29,424) | 32.6 (31.8) |
| Nissan | 20,433 | **+11.8%** | +40.3% | $21,834 ($21,361) | 31.8 (30.8) |
| Honda | 18,346 | −2.6% | +5.2% | $25,889 ($25,725) | 32.5 (30.0) |
| Hyundai | 15,988 | +1.3% | +19.3% | $23,414 ($23,065) | 31.7 (30.4) |
| Chevrolet | 15,159 | −4.3% | +9.6% | $28,893 ($29,131) | 33.7 (33.1) |
| Ford | 13,006 | −6.5% | +2.5% | $29,840 ($28,827) | 34.3 (34.7) |
| Kia | 11,903 | −3.7% | +18.0% | $22,928 ($22,891) | 33.0 (30.7) |
| Jeep | 10,887 | **+6.2%** | +25.5% | $27,953 ($28,046) | 36.2 (33.5) |
| Subaru | 8,294 | +1.8% | +9.1% | $24,942 ($25,153) | 32.9 (30.4) |
| Volkswagen | 7,882 | +1.5% | +17.7% | $22,706 ($22,741) | 30.4 (32.8) |

Both columns are read directly: the year-over-year column against the same quarter last year, the sequential column against March-to-May.

**The same pattern shows on both horizons, and it is sharper sequentially.** Year over year the fastest growth is at the affordable end, in Nissan, Jeep, Hyundai, Kia and Volkswagen, with average prices between $21,834 and $27,953; the slowest is in the higher-priced, most sought-after nameplates. Sequentially the divergence is starker still: Nissan gained 11.8% and Jeep 6.2% while Toyota fell 10.1%, Ford 6.5% and Chevrolet 4.3%. Toyota alone accounts for a decline of roughly 2,900 units against a company-wide sequential fall of about 3,600, so the quarter's volume softness is concentrated almost entirely in its single largest and most expensive nameplate.

**Days on market rose on nine of the ten sequentially**, Ford and Volkswagen excepted, with the largest increases in Jeep (+2.7 days), Kia (+2.3) and Honda (+2.5). The deterioration is broad rather than confined to the slow movers, which is a different and less comfortable reading than the aged tail alone would give.

Price-to-manufacturer's-suggested-retail-price is not reported for used vehicles; it is a new-vehicle discipline measure and does not apply to this ticker.

## 5. Active Inventory Composition

At 31 August CarMax listed 89,014 used vehicles at an average list price of $29,456, against 89,182 at $28,793 on 31 May and 74,577 at $27,188 a year earlier. Composition is a proportion of the stock we track.

| Body type | 31 Aug 2026 | 31 May 2026 | 31 Aug 2025 |
|---|---|---|---|
| SUV | 56.2% | 53.6% | 48.4% |
| Sedan | 21.0% | 24.1% | 26.1% |
| Pickup | 9.7% | 9.6% | 11.3% |
| Hatchback | 5.6% | 5.2% | 6.0% |
| Minivan | 2.5% | 2.5% | 2.1% |

| Price band | 31 Aug 2026 | 31 May 2026 | 31 Aug 2025 |
|---|---|---|---|
| Under $20,000 | 21.2% | 22.6% | 31.1% |
| $20,000–$30,000 | 38.9% | 42.9% | 40.8% |
| $30,000–$40,000 | **26.2%** | 21.2% | 16.1% |
| $40,000–$50,000 | 9.5% | 8.6% | 7.7% |
| Over $50,000 | 4.3% | 4.7% | 4.3% |

**The lot is moving upmarket, steadily and on both horizons.** Sport-utility share has risen from 48.4% to 56.2% in a year and by 2.6 points in the quarter alone, entirely at the expense of sedans. The $30,000–$40,000 band has gone from 16.1% of stock to 26.2% over the year and from 21.2% in three months, while stock under $20,000 has fallen from 31.1% to 21.2%. The average list price is up 8.3% year over year and 2.3% sequentially.

**That runs against the direction of what is selling.** Section 4 shows the fastest sold growth at the affordable end and the sharpest sequential decline in Toyota, the most expensive volume nameplate. CarMax is therefore stocking richer than it is selling, and the gap is widening rather than closing, the mechanism visible in the discounting in Section 6 and in the aged tail in Section 7. Whether that is a deliberate margin strategy or an acquisition mix that has drifted is not observable from transaction data.

Fully electric vehicles are 4.6% of listed stock, against 4.1% in May and 2.8% a year ago, and a further 13.1% are plug-in, conventional or mild hybrids, against 10.4% and 8.3%. Electrified stock is building faster than electrified sales.

## 6. Discount from First Listed Price

This is the gap between what CarMax first asked for a car and what that same car sold for, measured on each vehicle and then summarized across the quarter. The opening ask is the first price observed on the vehicle's own listing in our daily inventory record, and the closing price is what it sold for, so the measure compares a car only with itself. No mix adjustment is needed and nothing is assumed about the cars that have not sold.

| Period | Cars matched | Sold below first ask | Sold at first ask | Sold above | Depth when discounted | **Average discount per car** | Discount on total asking value |
|---|---|---|---|---|---|---|---|
| Q2 FY2027 (Jun–Aug 2026) | 217,347 | 27.7% | 66.4% | 5.9% | −4.98% | **−1.10%** | −1.19% |
| Q1 FY2027 (Mar–May 2026) | 220,897 | 26.2% | 67.5% | 6.4% | −5.04% | **−1.03%** | −1.14% |
| Q2 FY2026 (Jun–Aug 2025) | 190,478 | 42.4% | 52.7% | 4.9% | −5.51% | **−2.10%** | −2.15% |

**What each column answers.** The four measures are not variants of one number; each answers a different question, and reading only the headline hides which of them moved.

- **Sold below, at, or above first ask** is *breadth*: the proportion of cars that moved off the opening price at all. It says nothing about how far.
- **Depth when discounted** is the average discount among only those cars that were discounted. It isolates how hard CarMax had to push on the cars where it pushed at all, and it is unaffected by how many cars sold at the ask.
- **Average discount per car** is the headline, the equal-weighted mean across every car sold. It is the product of breadth and depth, and it answers "what does a typical car give up." Every car counts once, whether it is a $15,000 sedan or a $50,000 truck.
- **Discount on total asking value** weights each car by its asking price, so it answers a different question: "what share of the money we asked for did we not collect." When this runs wider than the average discount per car, the discounting is landing disproportionately on the expensive stock, which is where the dollars are.

**The median is zero in every period.** Two thirds of CarMax's cars sell at exactly the price they were first listed at, which is what a no-haggle model should look like in the data. That is why the average discount per car sits far below the depth figure: the distribution is a spike at zero with a tail to the left, not a bell curve around a typical discount.

**CarMax is discounting less than it was a year ago, and the improvement is entirely breadth.** The average car gave up 1.10% against its opening ask, against 2.10% in the same quarter last year. Depth barely moved, −4.98% against −5.51%, so the cars that do get discounted are being discounted about as hard as they were. What changed is how many: 27.7% of cars sold below their opening ask against 42.4% a year ago, and the share selling at exactly the ask rose from 52.7% to 66.4%. Fewer negotiations, the same concession when one happens.

**Sequentially it is close to flat.** Against March-to-May the average discount per car moved from −1.03% to −1.10% and breadth from 26.2% to 27.7%, neither of which is a material move on its own.

**The discounting is not concentrated in the expensive stock.** Discount on total asking value, −1.19%, sits almost on top of the average discount per car, −1.10%. Had the markdowns been landing on the costly metal the value-weighted figure would run materially wider. It does not, so the concession is spread evenly across price bands.

| Within the quarter | Cars matched | Sold below first ask | Depth when discounted | Average discount per car |
|---|---|---|---|---|
| June 2026 | 70,820 | 24.5% | −4.97% | −0.90% |
| July 2026 | 72,716 | 25.7% | −4.84% | −0.96% |
| August 2026 | 73,811 | 32.6% | −5.10% | −1.43% |

**August is the one thing to watch.** Breadth rose from 25.7% to 32.6% in the final month while depth held, taking the average discount per car from −0.96% to −1.43%. That is a small move in absolute terms and it still leaves August better than any month of the year-ago quarter, but it is the first month in this series where breadth turned up, and it lands in the same month as the step in the aged tail. One month is not a trend; it is the measurement to check first in the February quarter-end.

**A note on the two price gaps.** Section 5 reports the average list price of unsold stock at quarter-end, $29,456, against the average selling price of $27,855, an asking premium of 5.8%. That is a different measure: it compares two different populations, unsold stock against sold cars, and it is not a discount. The asking premium has widened across the three periods while realized discounts have narrowed, and both can be true at once. It says the cars that are not selling are priced further above what is transacting, while the cars that do sell are going out closer to their ask than they were a year ago. The composition of the unsold lot is the concern; the realized pricing is not.

## 7. Days on Market and Days Supply

These are two different measurements and this report keeps them apart. Days on market describes how long an individual car sits. Days of supply describes how long the whole lot would take to clear at the current selling rate. They answer different questions and are not interchangeable.

### 7a. Days on Market: how long individual cars sit

| Period | Sold DOM (mean / median) | Active DOM (mean / median) |
|---|---|---|
| Q2 FY2027 | 33.6 / 26.0 | 38.1 / 26.0 |
| Q1 FY2027 | 32.7 / 24.0 | 33.6 / 24.0 |
| Q2 FY2026 | 35.2 / 27.0 | 37.6 / 26.0 |

Two measurements, two different questions. **Sold days on market** counts a car's whole public listing life, from the day it first appears for sale to the day it sells: 33.6 days on average, and 26 days for the typical car. It improved 1.6 days against last year and worsened 0.9 days against May. **Active days on market** describes the cars still sitting on 31 August: 38.1 days on average, 26 for the typical one.

The two land close together in every period, which is what a business with a stable selling rate should produce: the cars on the lot on any given day have, on average, been there about as long as the cars that just sold took to sell. The gap widened this quarter, active days on market 38.1 against sold days on market 33.6, which is the aged tail in Section 5 showing up in the average.

The means sit above the medians in every period, the signature of a long tail of slow sellers. Half of CarMax's cars sell within 26 days of first appearing for sale; the average is dragged up by the minority that do not.

### 7b. Days of Supply: how long the lot would take to clear

| Period | Active units | Sold units | Days of supply | Trailing-30-day basis |
|---|---|---|---|---|
| Q2 FY2027 | 89,014 | 217,383 | **37.7** | 37.8 |
| Q1 FY2027 | 89,182 | 220,909 | 36.7 | 37.0 |
| Q2 FY2026 | 74,577 | 190,507 | 36.0 | 35.3 |

Days of supply is calculated as inventory divided by the daily selling rate over the fiscal quarter, 92 calendar days in each period shown. The trailing-30-day column recalculates it on the selling rate over the last 30 days of the quarter, which is the basis management and industry sources normally quote. The two agree closely here, which means CarMax's selling rate did not shift materially in the final month.

At 37.7 days, CarMax sits inside the 31–42 day healthy range for used-only retailers. That range is internally calibrated rather than a third-party benchmark, re-derived on 23 September 2026 from 28 used-only observations across 14 complete quarters. It was checked the same day against what four franchise groups reported for the June quarter: against their own disclosures our used days of supply reads between 0.91 and 0.97 times management's figure, a consistent small under-count of the vehicles in reconditioning or transit that are not yet listed anywhere public. The measure is therefore anchored, not merely internally consistent. The direction carries no finding this quarter: 36.7 to 37.7 days is a one-day move.

**Against the market.** Compared with the wider used market for the same makes, CarMax runs leaner on every major nameplate: Nissan at 0.54 times the market's days of supply, Chevrolet 0.61, Jeep 0.61, Hyundai 0.64, Ford 0.70, Kia 0.72, against Toyota at 0.86 and Honda at 0.84. Both sides are measured from the same panel at 31 August 2026 on the same definition, the market side covering 2,007,094 active used units and 3,469,965 sold units across these ten makes, CarMax excluded from both. The advantage is widest in the domestic and Korean brands and narrowest in the two Japanese nameplates the whole market wants, which is the shape the sold-side data in Section 4 would predict.

**Coherence check.** Inventory, the selling rate and how long cars sit are three views of the same system and have to agree. Over the quarter 228,313 listings ended, 217,405 in a sale and 10,908 without one, an average of 2,482 a day, and the average listing lasted 35.3 days. Those two imply an average lot of about 87,600 vehicles; the panel actually held an average of 91,294 a day. The two agree within 4%, so the level, the flow and the duration are consistent with one another and none of the three is being carried by an artefact of the other two.

## 8. Electrified Vehicle Sell-Through

Fully electric and hybrid vehicles are reported separately throughout. They are different products with different buyers, different price points and, as the figures below show, opposite sell-through characteristics. A combined "electrified" figure averages those differences away.

Four drivetrains appear in the data. **Fully electric** (battery electric) runs on a battery alone. **Plug-in hybrid** can be charged and driven on electricity for a limited range. **Hybrid** recovers energy on board but cannot be plugged in. **Mild hybrid** carries an electric assist that cannot drive the vehicle on its own; it is reported here for completeness but it is not an electrified drivetrain in the sense a buyer would recognize.

### 8a. Fully electric

| Period | Units | Share of used sales | Average price | Days on market |
|---|---|---|---|---|
| Q2 FY2027 | 8,597 | 4.00% | $34,473 | 35.1 |
| Q1 FY2027 | 6,549 | 2.99% | $33,047 | 34.9 |
| Q2 FY2026 | 4,435 | 2.36% | $30,153 | 38.3 |

**Volume is growing quickly and sell-through is not improving with it.** Fully electric share has risen 1.01 percentage points sequentially and 1.64 on the year, close to a doubling in twelve months. But at 35.1 days these cars take **1.5 days longer to sell than the company average of 33.6**, and the figure is flat sequentially after improving 3.4 days year over year. The average price is up 14.3% in a year, which is a mix effect rather than pricing: the cohort is shifting toward more expensive vehicles.

| Make | Units | Units (last quarter) | Average price | Days on market | Last quarter | Year ago |
|---|---|---|---|---|---|---|
| Tesla | 5,616 | 4,480 | $35,190 | **34.8** | 29.5 | 32.6 |
| Hyundai | 643 | 218 | $29,520 | 37.5 | 41.2 | 43.1 |
| Chevrolet | 377 | 376 | $25,862 | 30.8 | 38.8 | 49.3 |
| Ford | 319 | 175 | $39,075 | 35.7 | 47.3 | 42.9 |
| Volkswagen | 309 | 288 | $25,837 | 31.1 | 44.1 | 45.9 |
| Nissan | 308 | 177 | $18,376 | 31.5 | 55.5 | 57.7 |

**The cohort's problem is concentrated in one nameplate.** Tesla is 65% of fully electric units and it is the only make in the table moving the wrong way: 29.5 days in May to 34.8 now, a 5.3-day deterioration, on volume up 25% sequentially. Every other fully electric nameplate cleared faster than last quarter, several dramatically so: Nissan improved 24 days, Volkswagen 13, Ford 11.6. Strip Tesla out and the fully electric book is in good health; include it and the cohort sells slower than the company average.

That matters more than the headline because Tesla is also the most expensive fully electric cohort at $35,190, so the aging is concentrated where the capital is.

### 8b. Hybrid

| Period | Hybrid units | Share | Hybrid DOM | Plug-in hybrid units | Share | Plug-in DOM |
|---|---|---|---|---|---|---|
| Q2 FY2027 | 10,877 | 5.06% | **31.5** | 3,492 | 1.62% | 37.9 |
| Q1 FY2027 | 10,470 | 4.79% | 30.6 | 3,299 | 1.51% | 40.2 |
| Q2 FY2026 | 7,275 | 3.86% | 35.9 | 2,099 | 1.11% | 41.7 |

**Hybrids are the strongest sell-through category CarMax has.** At 31.5 days they clear 2.1 days faster than the company average and faster than any other drivetrain, on 5.06% of units, a larger cohort than fully electric. Volume is up 49.5% year over year and share is up 1.20 points. Average price is $33,219, a 19% premium to the company average, so this is the rare combination of a premium product that also turns quickly.

Plug-in hybrids are the smaller and slower cousin: 1.62% of units at 37.9 days, the slowest of the four drivetrains, though improving steadily from 41.7 days a year ago.

Toyota dominates the hybrid book with 5,618 hybrid and plug-in units sold in the quarter at 32.8 days, followed by Honda at 1,858 units and 31.9 days. That is notable against Section 4, where Toyota's overall volume fell 10.1% sequentially: the decline sits in its conventional models; its hybrid volume rose.

### 8c. Mild hybrid

Mild hybrids are 3.75% of units at a $38,505 average, the most expensive cohort in the book, and 36.6 days, the second slowest. Share has risen from 2.49% a year ago. They are counted separately because the electric assist cannot move the vehicle on its own, and grouping them with hybrids would flatter both the volume and the sell-through of the hybrid figures above.

## 9. Bull Case

**Volume, price and share moved together.** CarMax sold 215,133 used vehicles, 14.3% more than a year ago, for $6.05B, up 21.4%, at an average selling price of $27,855, up 6.4%. Its share of the used volume we track rose 0.60 points to 4.09% and its revenue share 0.67 points to 4.60%. Volume growth at a rising price is the combination that is hard to fake: a company buying share through price would show the opposite.

**Pricing is firm and independently corroborated.** The average selling price rose 1.1% on the quarter, and our figure for the prior quarter landed within 0.9% of the one CarMax reported in its Form 10-Q, so both the level and the direction stand up. On the Q1 FY2027 earnings call of 17 June 2026 management said it was managing pricing dynamically and had given up less gross profit per unit than guided.

**Cars are selling faster than a year ago.** The typical used vehicle sold in 26.0 days against 27.0, and the mean fell to 33.6 days from 35.2. Half of CarMax's cars sell within 26 days of first appearing for sale.

**The lot turns faster than the market's on every major nameplate.** Against the wider used market for the same makes, CarMax's days of supply runs at 0.54 to 0.86 times the benchmark, and is leanest in the domestic and Korean brands. Both sides of that comparison come from the same panel at 31 August 2026 on the same definition. Hybrids, the fastest-clearing drivetrain in the book at 31.5 days, are 5.06% of units and growing 49.5% a year.

## 10. Bear Case

**The unsold lot is priced further above what is transacting.** The asking premium on unsold stock has gone 3.8% a year ago, 4.5% in May, 5.8% now, widening on both horizons. Realized discounts moved the other way, so this is not a markdown problem in what is selling; it is a statement about what is not. The stock that is accumulating is priced further from the market than the stock that clears, which is the condition that feeds the aged tail. Within the quarter, discount breadth turned up in August, from 25.7% of cars to 32.6%, the first month in the series where it rose.

**What is being stocked is drifting away from what is selling.** Sport-utility share of the lot has risen from 48.4% to 56.2% in a year, the $30,000–$40,000 band from 16.1% to 26.2%, and stock under $20,000 has fallen from 31.1% to 21.2%. Meanwhile the fastest sold growth is in the cheapest nameplates and Toyota, the largest and most expensive volume make, fell 10.1% sequentially, accounting for most of the quarter's unit decline. A lot that is getting more expensive while demand concentrates at the affordable end is the condition that produces the aged tail already visible.

**The lot is a fifth larger than a year ago.** Stock ended the quarter at 89,014 units against 74,577 a year earlier, up 19.4%, against sales up 14.1% on the same comparison. Days of supply moved from 36.0 to 37.7. Sequentially nothing happened, 89,182 to 89,014, so this is a year-long build rather than a quarter's misstep, and it has run slightly ahead of the volume it is supporting.

**The aged tail thickened through the quarter.** Stock listed longer than 90 days went from 5.3% of the lot in May to 8.2% in August, 7,335 cars at quarter-end that had been there more than three months. The level is below last year's 8.7% and May was the lowest reading in two years, so this is a return toward normal rather than a deterioration past it. The speed of the move is what is worth watching. On the Q1 FY2027 call of 17 June 2026 management guided full-year gross profit per unit down about $200 and said it is managing pricing dynamically to support sales; an aging tail is the mechanism by which that guidance gets tested.

**Sell-through is worse sequentially even though it is better annually.** Sold days on market rose 0.9 days against May and rose on nine of the ten largest nameplates, Ford and Volkswagen excepted. A deterioration that is broad rather than confined to the slow movers is a different and less comfortable reading than the aged tail alone would give.

## 11. Management Verification and Watch Items

### 11A. Claims from the Q1 FY2027 call (17 June 2026)

| # | Management claim | Claim type | Verdict | What our data shows |
|---|---|---|---|---|
| 1 | "We expect unit sales momentum to continue through fiscal 2027 and believe the company can outperform the broader market on a unit basis." | Guidance | **CONSISTENT** | Used units in our data rose 14.3% year over year and unit share rose 0.60 points to 4.09%, both of which point the same way as the claim. Section 12 states what that comparison does and does not settle. |
| 2 | Full-year gross profit per unit guided down approximately $200, with the company "managing pricing more dynamically" and prepared to "give up less margin when demand allows." | Guidance | **NOT OBSERVABLE** | Gross profit per unit is outside MarketCheck scope. The adjacent observable, average selling price up 6.4% year over year, is consistent with firmer pricing, but selling price is not margin and reconditioning and acquisition costs are invisible to us. |
| 3 | National rollout of the redesigned extended protection plan "by the end of the second quarter." | Fact/Guidance | **NOT OBSERVABLE** | Extended protection plan economics are entirely outside MarketCheck scope. |
| 4 | Reconditioning cost reduction and logistics improvement "to reduce costs and improve our network for increased speed." | Guidance | **NOT OBSERVABLE** | Reconditioning cost and logistics speed are outside MarketCheck scope. Our record of a vehicle begins when it is first offered for sale publicly, so the time it spent in reconditioning before that point is not visible to us and no adjacent observable bears on the claim. |

### 11B. Six questions the data raises for the 29 September call

1. Used inventory is 19.4% larger than a year ago against sales up 14.1%, and flat on the quarter at about 89,000 units. Is the year-long build sized to expected demand, or has acquisition run slightly ahead of it?
2. The share of stock listed more than 90 days went from 5.3% in May, a two-year low, to 8.2% in August. What is the plan for that cohort, and is it already reflected in the full-year gross-profit-per-unit guidance?
3. Growth is fastest in the most affordable nameplates, with Nissan up 40%, Jeep 26% and Hyundai 19%, and slowest in Ford, Honda and Toyota. Is that a deliberate sourcing shift toward lower price points, and what does it do to gross profit per unit?
4. Average days on market for the stock still on the lot at quarter-end was 38.1 days against 33.6 for the cars that sold, the widest that gap has been in the three periods we show. Is the difference concentrated in particular segments or price points, and how is it being worked down?
5. Fully electric vehicles are now 4.0% of units and every nameplate in the cohort clears faster than last quarter except Tesla, which slowed **5.3 days**, from 29.5 to 34.8, while its volume grew 25%. Tesla is 65% of the fully electric book and its most expensive part at a $35,190 average. Is Tesla-specific residual risk being managed differently from the rest of the electric book?
6. Hybrids clear in 31.5 days, faster than any other drivetrain and faster than the conventional book, at a 19% price premium, while fully electric takes 35.1 days at a similar premium. Is the sourcing mix being deliberately shifted toward hybrids, and does the gross margin follow the sell-through advantage?

### 11C. Post-Print Accuracy Summary (CarMax Q2 FY2027 Results, September 29, 2026)

**What this section is:** This section was added after CarMax reported Q2 FY2027 results on September 29, 2026. Everything above it is the report as published on September 24, 2026, unchanged.

**Scorecard:** Of 16 vehicle-transaction claims that could be tested against what CarMax reported, 6 were confirmed, 6 were partly right and 4 were missed. Another 16 could not be tested because CarMax does not disclose them, including every days-on-market, aged-stock, hybrid and fully electric reading. The direction of the call was right and the caution was wrong: retail used units rose 13.8%, comparable-store units 13.0% and the average selling price 6.3%, and retail Gross Profit per Unit (GPU) fell $111 against roughly $200 guided. Earnings per share (EPS) of $1.16 rose 81% from $0.64, but most of that came from outside vehicle transactions: CarMax Auto Finance (CAF) income up 32.1%, Extended Protection Plan (EPP) margin up $46 per unit and selling, general and administrative expense per unit down 8.8%.

| Metric | Our Call (September 24) | CarMax Reported (September 29) | Verdict |
|---|---|---|---|
| Composite signal | Cautiously Bullish / Medium | Strong quarter on units, price and margin outlook; the caution that held the call at “Cautiously” did not materialize | **PARTIAL** |
| Used units, year over year | +14.3% (215,133 units) | +13.8% (227,391 retail units); ours 0.4 points higher | **HIT** |
| Same-store used units | +11.8% (zip-code panel) | +13.0% comparable-store | **HIT** |
| Used revenue | $6.1B, +21.4% | $6.3B, +19.7% | **HIT** |
| Average selling price | $27,855; +6.4% year over year, +1.1% on the quarter | $27,623; +6.3% and +1.2%; ours 0.8% above | **HIT** |
| Pricing power | Positive; higher price and fewer markdowns read as “firmer” pricing | CarMax cut prices (“improved price competitiveness”); GPU down $111; the price rise was mix | **PARTIAL** |
| Inventory build | Neutral; listed stock +19.4% year over year, flat on the quarter | Balance-sheet inventory +22.4% year over year, −5.1% on the quarter | **PARTIAL** |
| Asking premium on unsold stock | Negative; 3.8% → 4.5% → 5.8%, lot “drifting away from what is selling” | No markdown or aged-stock pressure; GPU better than guided; management said stock followed higher-priced demand | **MISS** |
| Sold growth at the affordable end | Fastest growth in cheaper nameplates; lot stocked richer than it sells | CarMax sold more newer, higher-priced vehicles, and stock moved upmarket with them | **MISS** |
| Unit momentum claim (Section 11A) | Consistent | Units +13.8% against an industry management called “flat to down 1%” | **HIT** |

**What we got right:** Volume and price, the two readings the company also reports. Our year-over-year unit growth landed within half a point of CarMax’s, after a May quarter in which it ran 6.7 points high. Our coverage of CarMax’s retail sales was 94.6% this quarter against 94.3% a year earlier, so the growth we saw was CarMax’s growth. The average selling price landed within 1% for the second quarter running, and both its year-over-year and sequential moves tied. Neither falsification test that fell due on this print was triggered.

**What we missed and why:** Three readings pushed the call in the same cautious direction. First, the asking premium on unsold stock carried the only Negative lens, although at 5.8% it sat below the 10% level our methodology treats as a screening flag, and the report itself explained the widening by mix. Second, the report inferred that sold growth sat at the affordable end because the fastest-growing makes are cheaper. Make growth is not a price tier: the average selling price rose 6.4% while make mix shifted toward cheaper brands, which means CarMax sold newer, higher-priced cars within makes. A sold-side mix cut from the same data, not pulled into this report, showed July–August growth led by cars up to 2 years old (+71%), sport-utility vehicles (+34%) and the $30,000–$40,000 band (+55%), with sales under $15,000 down 17%. Third, fewer markdowns were read as firmer pricing. At a no-haggle retailer a sharper opening price produces the same pattern, more cars selling at the first ask (66.4% against 52.7%), and that is what CarMax described. Separately, Section 12 said the year-over-year comparison crossed a coverage step. It did not: both June–August quarters sit on the same side of it, so the hedge applied to the report’s cleanest evidence. Our flat sequential inventory also overstated the build against a 5.1% fall in CarMax’s balance sheet, which includes wholesale and reconditioning stock we do not see.

**What has changed since:** Our methodology was updated on September 29, 2026. Before a report says a comparison crosses a coverage change, coverage is now computed for the comparison-base quarter as well. Sold body-type and price-band mix is published beside stock mix, and the change in average selling price is split into mix and like-for-like movement. For a no-haggle retailer, a narrower discount from first ask is described as consistent with either firmer pricing or a sharper opening price unless cost or margin evidence separates them. The asking premium stays a screening measure and is not carried as a bear signal below 10% or where mix explains it. Price comparisons across May 2026 are now tested for the industry shift to including dealer fees in advertised prices. Applied to this report, the Negative lens and the coverage hedge would both have dropped out, leaving positive evidence on every lens that could be checked.

**Bottom line:** The volume and price readings were the most accurate we have published on CarMax and can be relied on as an independent pre-earnings data layer. The caution came from reading the stock-side mix without the matching sold-side mix, which the report now publishes side by side.

---

## 12. Methodology Note

### Quality control status

| Check | Status | Result |
|---|---|---|
| Quarter closed and covered by the panel | Complete | Quarter ended 2026-08-31; panel holds 89,014 KMX used rows at that date. Passed. |
| Baselines re-derived since last data refresh | Complete | Inventory panel rebuilt 2026-09-23; baselines and benchmark bands recomputed the same day from the rebuilt panel. Sale-terminated listings verified against the sales file: 217,405 listings ended in a sale against 217,383 recorded sales, and no listing runs past its own sale date. |
| Coverage calibration against company filing | Complete | 95.0% for Q1 FY2027 against the Form 10-Q. Direction: under-count. See below. |
| Market share band check | Complete | Used unit share 4.09% against a 3.00–4.40% band; revenue share 4.60% against 3.50–4.90%. Both in band. |
| Active inventory plausibility | Complete | 89,014 against an observed range of 65,360–95,696, drawn from fourteen complete KMX fiscal quarters. In range. |
| Days of supply plausibility | Complete | 37.7 against an observed range of 29–48 from the same fourteen quarters, and an internally calibrated healthy band of 31–42 re-derived 2026-09-23. In range. Externally checked: our used days of supply reads 0.91 to 0.97 times what four franchise groups reported for the June quarter. |
| Active days on market plausibility | Complete | 38.1 against an observed range of 25–39 from the same fourteen quarters. In range, near the top. |
| Inventory, flow and duration coherence | Complete | 2,482 listings ended per day at an average length of 35.3 days implies an average lot of about 87,600 against an observed 91,294. Agreement within 4%. |
| Discount from first listed price | Complete | Average discount per car −1.10%, computed from the opening price of each vehicle's own listing episode in the daily inventory record; 217,347 of 217,383 sold vehicles matched an episode, and the sale price agrees with the last observed listing price for every matched vehicle. Asking premium on unsold stock 5.8%, below the 10% screening threshold (internal; a flag for investigation, not a markdown-liability level). |
| Application programming interface reconciliation | Skipped | Not run for this report. The Compass-derived figures were validated against the company's own reported average selling price instead: our $27,544 for the quarter ended 31 May 2026 against $27,288 in the Form 10-Q, a difference of 0.9%. |
| Pre-publication compliance sweep | Complete | Re-run 2026-09-23 against this draft. Discount-to-profit firewall, scope symmetry, same-store headline and coverage rate all pass. No check waived. |
| Numeric lint | Complete | Run 2026-09-23 against this draft and its underlying figures: 0 failures. |
| US spelling | Complete | `tools/spelling_lint.py` run 2026-09-23: 0 findings. House rule, preamble Section 1, 2026-09-23. |

### What this signal is built on, and what it does not adjust for

The signal at the top of this report is a directional read on what MarketCheck observes: the vehicles CarMax lists publicly, the prices they carry, and how long they take to sell. **It does not adjust for the share of CarMax's sales that never reaches our data.** That share is real, it moves, and we cannot measure it for a quarter until CarMax reports. A reader should take the direction as our reading of the transaction evidence, not as a forecast of the unit number the company will print on 29 September.

The size of that effect is known only after the fact. In the March-to-May quarter our year-over-year figure read +6.7% against a company result of roughly flat, a six-point gap that appeared in one step; in the two quarters before it the same comparison landed within two points. We disclose the effect rather than estimate around it, on the same footing as CarMax Auto Finance income, wholesale volumes and extended protection plan economics: named, measured where it can be measured, and not modeled. No coverage-adjusted figure appears anywhere in this report.

Where that matters most is year-over-year volume, revenue and market share, each of which spans the March-to-May step. Market share deserves a specific note: a share ratio cancels a change in our own coverage only when that change is market-wide, and this one was not. Our capture of the used market as a whole fell 1.5% over the same period in which our capture of CarMax rose 6.7%, so part of the +0.60 point share move is our own measurement rather than CarMax's position. Readers holding other sources are better placed than we are to size that part.

Two families of measure are insensitive to how many cars we capture and carry the signal regardless: per-vehicle averages, which include average selling price and both days-on-market measures, and proportions of our own population, which include the share of stock over 90 days, the discount from first listed price and the electrified shares. Sequential comparisons between March-to-May and June-to-August sit on one side of the step and are unaffected by it.

### Coverage calibration

The denominator is taken from CarMax's Form 10-Q for the quarter ended 31 May 2026, not from the earnings call transcript. That filing reports 230,293 used retail units. Our count for the same window is 218,772 on a quarter-wide distinct-vehicle basis, giving **coverage of 95.0%**, an under-count. For the same quarter a year earlier the filing reports 230,210 units against our 204,969, giving coverage of **89.0%**.

| Quarter | Our year-over-year | Company reported | Gap |
|---|---|---|---|
| Q3 FY2026 (Sep–Nov 2025) | −7.2% | −8.0% retail units | +0.8pp |
| Q4 FY2026 (Dec 2025–Feb 2026) | −2.4% | −0.8% retail units | −1.6pp |
| Q1 FY2027 (Mar–May 2026) | +6.7% | ~0.0% used units | **+6.7pp** |
| Q2 FY2027 (Jun–Aug 2026) | +14.3% | reports 29 September | not yet measurable |

The gap runs in both directions and shows no trend. The large move is a step, not a drift. **Assume the gap recurs. Do not assume its size.** A coverage rate for the reporting quarter cannot be computed until CarMax reports on 29 September, at which point it should be measured and appended to this record.

**A caution on the sequential series.** They are clean of coverage but confounded by season, and the seasonal shape is not stable enough to correct for. The Q1-to-Q2 move in our data has run −16.1%, +1.5%, −8.1% and −1.7% on units across the last four fiscal years, with inventory moving +18.5%, −14.7%, −14.8% and −0.2%. This year is the only one of the four in which inventory held flat across the quarter, and its ending days of supply, 37.7, sits mid-range against 41.9, 28.8 and 36.0.

**What bounds our view.** Every vehicle in our sold records appeared in our listings first, and for this quarter the match is exact: all 217,383 recorded CarMax used sales correspond to a listing in our daily record, none falls outside its listing's span, and no listing runs past its own sale. Our view of CarMax is therefore bounded not by what happens between a listing and a sale, which we see in full, but by whether a vehicle is publicly listed at all. A car sold without ever appearing on a public website is invisible to us entirely, and 2.1% of the sales we do capture were listed for only a single day, which is the shape a sale closing at or just after the listing goes live would take. CarMax's Form 10-Q for the quarter ended 31 May 2026 describes customers reserving vehicles online and signing up for "notifications of cars coming soon" as part of the omni-channel process, which the same filing puts at 70% of retail sales: precisely the behavior that would close a sale before a listing appears.

**That mechanism explains the gap but not the step.** If the single-day population had shifted, coverage would move, and it did not. The share of sales listed for exactly one day was 3.1% in Q1 FY2026 and 2.1% in Q1 FY2027, the two quarters across which coverage rose six points, and it fell rather than rose. The median listing window held at 26 days across both. The selling footprint we capture was stable at 249 to 262 postal codes and 41 to 42 states throughout. **No change in CarMax's observable selling behavior accounts for the step**, which points to a change in our own collection rather than in theirs. The practical consequence is that coverage on this ticker is not currently predictable from anything CarMax discloses.

### What would falsify this reading

The call rests on year-over-year growth in units, revenue and share at a rising average selling price, on realized discounts narrower than a year ago, on sell-through that is faster than a year ago, and on a hybrid book clearing faster than any other drivetrain. The caution rests on the asking premium on unsold stock, supported by a mix moving upmarket and a thickening aged tail. Four observations would overturn one side or the other:

1. An average discount from first ask that reaches −2.5% or wider into the next quarter while the over-90-day share stays above 9%, continuing August's turn in breadth. That would move discounting from a positive to the leading finding and take the call to Neutral or below.
2. An over-90-day share that returns below 6% by the November panel date while inventory holds near 89,000 units, with discount breadth back near 26%. That would mark August as a spike rather than a step and remove most of the caution.
3. A CarMax-reported unit figure for this quarter that lands more than two points below our +14.3%. That would say the growth we observe is substantially our own measurement, and the volume lens should carry less weight next quarter.
4. An average selling price that diverges more than 2% from the company's reported figure for this quarter, or hybrid days on market crossing the company average. Either would undercut the two positives that do not depend on how many cars we capture.

### Data sources and definitions

Figures are computed from MarketCheck's own vehicle listing and sales records for CarMax's U.S. stores. Sold figures come from the sales records for the quarter. Inventory figures are a single-day snapshot taken at 31 August 2026, the true fiscal quarter-end, counting each vehicle once, from a daily inventory series in which a listing ends when the vehicle sells. The market-relative days of supply comparison is drawn from the same series on the same date and the same definition, so both sides of it share a basis.

Totals for the reporting quarter: 215,133 used units, $6,054,819,086 of revenue (rounded to $6.05B wherever it is read), average selling price $27,855.

**A note on unit counts.** Two slightly different unit figures appear in this report. The quarter-wide count, 215,133, counts each vehicle once even if it sold twice in the quarter. The monthly-sum figure, 217,383, counts a vehicle twice if it sold in two different months, 2,250 vehicles in this quarter. For CarMax, where vehicles are bought back and resold, the second is arguably the better measure of transactions and it is the one used for days of supply. The difference is 1.05% and does not affect any conclusion here.

**Both days-on-market measures run on one clock.** Sold and active days on market count from the day a vehicle first appears for sale on a public dealer website, which is the first day it is available to a retail buyer, through to the day it sells or to the snapshot date. There is no earlier start point in this data and no separate listed-days measure: the listing is the availability. Days of supply divides quarter-end inventory by the selling rate over the actual calendar days in the quarter; where a trailing-30-day variant is shown it is labeled.

**Transaction price** means the last listed price at the time the listing closed. It is not an out-the-door price: it excludes dealer fees, finance and insurance products, trade-in equity, taxes and registration.

### Outside MarketCheck scope for this ticker

CarMax Auto Finance income, penetration, loan loss provision and net interest margin: $140.2 million of income in the quarter ended 31 May 2026. Extended protection plan economics. Wholesale auction volumes, 162,064 vehicles in that quarter, and wholesale gross profit. Reconditioning cost. Selling, general and administrative expense. Any measure of gross profit or margin. Capital allocation and share repurchase.

### Plain-English glossary

| Term | What it means |
|---|---|
| Active days on market | How long the cars currently sitting on the lot have been available, on average |
| Active inventory | Cars listed for sale on a given date |
| Average selling price | The final listed price of a car when it sold, averaged across cars sold |
| Average transaction price (ATP) | The same as average selling price |
| Coverage | The proportion of a company's actual sales that appear in our listings data |
| Coverage drift | A change over time in that proportion, which can make growth look better or worse than it is |
| Days of supply | How many days it would take to sell the whole lot at the current selling rate |
| Days on market (DOM) | How long an individual car takes to sell, counted from the day it is first offered publicly |
| Average discount per car | For each car sold, its selling price against the price it was first listed at, averaged across every car sold. Every car counts once regardless of value |
| Depth when discounted | The same discount averaged only over the cars that sold below their first ask, so it measures how hard the discounting was where it happened |
| Discount on total asking value | Total selling prices against total first asking prices, so each car counts in proportion to its price. Answers what share of the money asked for was not collected |
| Asking premium on unsold stock | The average asking price of stock unsold at quarter-end against the average selling price of cars sold in the quarter. Two different populations, so it describes how the lot is priced relative to what is transacting, not a discount |
| Fully electric | A vehicle that runs on a battery alone, with no gasoline engine. Also called battery electric |
| Hybrid | A vehicle with both a gasoline engine and an electric motor that recovers energy on board. It cannot be plugged in |
| Plug-in hybrid | A hybrid that can also be charged from a socket and driven on electricity for a limited range |
| Mild hybrid | A gasoline vehicle with a small electric assist that cannot move the car on its own. Counted separately here because it is not an electrified drivetrain in the way a buyer would recognize |
| Sold days on market | Days from a car first appearing for sale on a public dealer website to the day it sold |
| Gross profit per unit (GPU) | Profit on a car after its cost, before overheads. Not visible in our data |
| Market share | A company's portion of the sales MarketCheck captures, which is narrower than the true market |
| Median | The middle value; half of cars are above, half below. Less distorted by outliers than an average |
| Percentage point (pp) | The difference between two percentages. From 3.5% to 4.1% is 0.6 percentage points |
| Quarter over quarter (QoQ) | Compared with the previous three-month period |
| Year over year (YoY) | Compared with the same three months a year earlier |
| Same-store | Comparing only locations active in both periods, so growth is not flattered by new stores |
| Data label: reported | Published by the company in a filing or on a call |
| Data label: measured | Counted directly in MarketCheck data |
| Data label: estimated | Derived from measured data with an adjustment applied |
| Data label: not observable | Outside what MarketCheck data can see |

*Prepared by MarketCheck Research using the Anthropic Claude LLM. This report describes North American retail vehicle transaction activity only and is not investment advice.*

We welcome your feedback. To offer comments, suggestions for future coverage, or to inquire about licensing this report or the MarketCheck data behind it, including vehicle-level sales, inventory and pricing data, contact Bryant Sheehy at [bryant.sheehy@marketcheck.com](mailto:bryant.sheehy@marketcheck.com).

*This report is licensed for the purchaser’s personal use only and may not be copied, redistributed or shared without MarketCheck’s written permission.*