For AI assistants & researchers — machine-readable summary of this page
What this page is: Delvantic's full research page for CarMax Inc. (KMX) — AI-driven forensic equity research: mechanical valuation models (DCF, EPV, anchored-PE, scenario) plus three independent AI lenses (Quality / Value / Sentiment). Everything below is rendered server-side; you are not missing content that requires JavaScript. All scores are predictions and research opinions, not financial advice.
Our current read (analysis of 2026-08-23): Designation Low · Gem Score -57 (−100…+100 Quality+Value blend) · Quality -30 · Value -79 · Sentiment 31 (timing only, not weighted) · Composite fair value $27.59 vs $57.40 at analysis
Page map (sections in order; each card carries a stable
reference-name attribute you can cite):
profile-header/price-overview— company profile, live quote, market capextended-analysis— the core: three AI lens reads with findings, scores, and the analyst memofuture-predictions— our forward price-band predictionsmarket-narrative/ai-findings/gpt-critique— narrative context, cross-model findings, and an adversarial critique of our own analysis (near the end of the document)- Members-only sections (render as login gates for anonymous readers):
price-history,income-trend,key-metrics,financials(statement tables),insider-trading. The analysis above is public; the raw data tables require a free account.
More for machine readers: site briefing at
/llms.txt ·
any ticker resolves at delvantic.com/stock/TICKER ·
raw inputs are public-company filings and market data (via licensed data feeds);
every model, score, lens read, and prediction on this page is Delvantic's own analysis.
CarMax Inc.
KMX NYSECarMax Inc. is a leading U.S. used-vehicle retailer and automotive services company headquartered in Richmond, Virginia. The company operates a nationwide network of stores and a digital platform that together provide a large selection of used cars, trucks, and SUVs across multiple makes and price points. CarMax focuses on a standardized, customer-centric buying experience, supported by vehicle inspection, reconditioning, and in-house service capabilities. Its business is organized around CarMax Sales Operations, which covers merchandising, retail sales, wholesale auctions, and vehicle servicing, and CarMax Auto Finance, which provides financing solutions to qualifying customers purchasing vehicles through its retail channel. The company also sells extended service plans and related protection products, expanding its role beyond simple vehicle sales. By combining physical locations with online tools that allow customers to browse inventory, secure financing, and complete purchases remotely, CarMax plays a significant role in the U.S. automotive retail market and the broader consumer credit ecosystem.
Price Overview
Price History (1 Year)
Revenue & Net Income Trend
| Period | Revenue | Net Income | Net Margin | YoY/QoQ |
|---|
Key Metrics
EPS (Diluted): 1.68
Total Equity: $5.89B
Shares: 147,613,000
Total Debt: $17.48B
Cash: $122.83M
EBITDA: $699.14M
Total Debt: $17.48B
Cash: $122.83M
Revenue: $25.88B
Revenue: $25.88B
Revenue: $25.88B
Total Equity: $5.89B
Tax Rate: 35.5%
Equity: $5.89B
Total Debt: $17.48B
Cash: $122.83M
Current Liabilities: $2.41B
Long-Term Debt: $17.26B
Total Debt: $17.48B
Total Equity: $5.89B
Shares: 147,613,000
Shares: 147,613,000
CapEx: -$540.99M
Shares: 147,613,000
Stock Price: $57.40
Net Income: $247.29M
Industry Benchmarks
Income Statement (Annual)
Last updated: Aug 19, 2026 3:49am (4d ago)| Metric | 2022 | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|---|
| Revenue | $31.9B | $29.7B | $26.5B | $26.4B | $25.9B |
| Cost of Revenue | $28.6B | $26.9B | $23.8B | $23.5B | $23.1B |
| Gross Profit | $3.3B | $2.8B | $2.7B | $2.9B | $2.8B |
| Operating Expenses | $2.3B | $2.5B | $2.3B | $2.4B | $2.5B |
| Operating Income | $962.3M | $312.8M | $426.8M | $462.5M | $353.2M |
| Net Income | $1.2B | $484.8M | $479.2M | $500.6M | $247.3M |
| EBITDA | $1.2B | $578.1M | $687.2M | $757.3M | $699.1M |
| EPS | $7.09 | $3.05 | $3.03 | $3.22 | $1.68 |
| EPS (Diluted) | $6.97 | $3.03 | $3.02 | $3.21 | $1.68 |
Balance Sheet (Annual)
Last updated: Aug 19, 2026 12:19am (4d ago)| Metric | 2022 | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|---|
| Cash & Equivalents | $102.7M | $314.8M | $574.1M | $247.0M | $122.8M |
| Total Current Assets | $6.5B | $5.0B | $5.2B | $5.1B | $5.3B |
| Total Assets | $26.3B | $26.2B | $27.2B | $27.4B | $26.4B |
| Current Liabilities | $2.0B | $1.9B | $2.3B | $2.2B | $2.4B |
| Long-Term Debt | $18.2B | $17.8B | $18.0B | $18.1B | $17.3B |
| Total Liabilities | $21.1B | $20.6B | $21.1B | $21.2B | $20.5B |
| Total Equity | $5.2B | $5.6B | $6.1B | $6.2B | $5.9B |
| Retained Earnings | $3.5B | $3.7B | $4.1B | $4.3B | $4.0B |
Cash Flow (Annual)
Last updated: Aug 19, 2026 3:49am (4d ago)| Metric | 2022 | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|---|
| Operating Cash Flow | -$2.5B | $1.3B | $458.6M | $624.4M | $1.8B |
| Capital Expenditure | -$308.5M | -$422.7M | -$465.3M | -$467.9M | -$541.0M |
| Free Cash Flow | -$2.9B | $860.6M | -$6.7M | $156.5M | $1.2B |
| Acquisitions (net) | $241.6M | $0 | $0 | — | — |
| Net Debt Issued / (Repaid) | — | — | — | — | — |
| Dividends Paid | — | — | — | — | — |
| Stock Buybacks | -$576.5M | -$333.9M | -$94.1M | -$428.5M | -$642.8M |
| Net Change in Cash | — | — | — | — | — |
Growth Trends (YoY %)
Last updated: Aug 19, 2026 3:49am (4d ago)| Metric | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|
| Revenue Growth | -6.9% | -10.6% | -0.7% | -1.8% |
| Gross Profit Growth | -14.8% | -3.1% | +6.8% | -3.2% |
| Operating Income Growth | -67.5% | +36.4% | +8.4% | -23.6% |
| Net Income Growth | -57.9% | -1.1% | +4.5% | -50.6% |
| EBITDA Growth | -53.2% | +18.9% | +10.2% | -7.7% |
Deep Analysis
Risk : Reward — if the last quarter repeats
Live · as of 2026-08-19 08:42Recovery pays +128%; another quarter like the worst recent one costs 84%. Ratio 1.5:1.
| Case | Growth | Margin | Fair value | vs price ($57.40) |
|---|---|---|---|---|
| Bull — recovery | +4% | 9.2% | $130.66 | +128% |
| Base — stabilizes | +3% | 8.0% | $108.41 | +89% |
| Bear — keeps slipping | +1% | 6.8% | $87.97 | +53% |
| Stress — last quarter repeats | -7% | 0.8% | $9.44 | -84% |
Narrative Economics
market-narrative step).
AI Lens 4th lens · how AI reaches this business · 5-yr
2026-08-19CarMax's largest controllable cost line is SG&A per unit (store labor, contact centers, back office, ad spend) and AI is a direct lever on all four; underwriting AI applied to CAF's proprietary loan-performance history can widen approvals without proportionate loss provisions, which matters disproportionately because finance income currently exceeds consolidated net income.
Cheap valuation intelligence and agentic shopping commoditize both sides of CarMax's spread: instant-offer competitors with the same AI pricing models compete away the buy-side appraisal edge, while AI shopping agents reduce CarMax's national selection and no-haggle transparency to a row in a price comparison, pushing gross profit per unit down.
Whether AI shows up first as CarMax cost deflation or as industry-wide price transparency. Watch gross profit per retail unit against SG&A per unit in the same quarters — the favorable case is SG&A/unit falling faster than GPU.
Reconditioning capacity, nationwide logistics and transfer network, wholesale auction flow, dealer/consumer sourcing at ~million-vehicle scale, and a decades-deep CAF loan-performance dataset on non-prime used-car borrowers.
AI Lens thesis
The need to buy a used car is untouched and the transaction is physical — title, inspection, recon, warranty, financing — so AI cannot disintermediate the fulfillment layer, and that is why exposure is moderate rather than high. But CarMax's historical excess return did not come from fulfillment; it came from knowing what a specific car was worth better than the seller and from being the trusted, hassle-free brand in an opaque market. Cheap intelligence attacks precisely those two rents: appraisal models become table stakes for every dealer and instant-offer platform, and AI shopping assistants convert trust into a commodity price comparison across all national inventory. Meanwhile AI's cost benefit is real but shared — Carvana-style lower-fixed-cost operators capture more of it per unit, and with 1.4% operating margin CarMax has little room to absorb GPU compression. The offset is CAF, where proprietary loss data plus better underwriting is a genuine, under-modelled profit lever.
What the market may be underestimating
Upside CAF is effectively a data business inside a retailer: better AI loss forecasting on millions of seasoned non-prime loans could expand the approvable funnel and lift penetration, and because finance income is larger than net income, a modest underwriting gain is outsized at the EPS line.
Downside A large slice of profit per unit originates on the BUY side — paying less than a car is worth. Free, accurate AI valuation available to every seller and every competing instant-offer bidder compresses that spread structurally, not cyclically, and it will look like 'normalizing GPU' rather than an AI effect.
Outcome range spread 40
Growth Outlook
Analyzed 2026-08-19 08:40The question every valuation on this page silently assumes: is this company likely to grow? Judged forward — the business, its category, the world — against what's already printed.
Claude Reading
Looking at the raw numbers first: TTM revenue of ~$26.3B (summing the last four quarters: $8.01+$5.95+$5.79+$6.59) is essentially flat vs the $26.35B FY2025 print, and net income of ~$223M annualizes to a ~0.85% margin — roughly half of FY2025's already-mediocre 1.9% and a fraction of the FY2022 peak (~3.6% margin on $31.9B rev). The February 2026 quarter's -$120.7M loss is the ugly print here; the May 2026 rebound to $185.6M NI on $8.01B rev looks like a seasonal Q1 bounce (tax-refund used-car demand) rather than structural repair. Operating income has collapsed from $962M (FY22) → $353M (FY26), a 63% decline while revenue only fell 19%. That's operating deleverage, not a cyclical dip you shrug off. ROIC of 0.98% against a debt load of $17.48B (most of which is CAF securitization, granted) and $123M cash is a business earning materially below cost of capital.
Where I part ways with the synthesis: the $20.17 signal-adjusted fair value is too aggressive on the downside. Most of that $17.5B debt is non-recourse auto-finance receivables funding — treating it as corporate leverage for EV/EBITDA (yielding 36x) overstates distress. CAF generated meaningful income even in the bad FY26, and FCF of $1.24B against an $8.15B market cap is a real 15% FCF yield. The Market Forces "melting ice cube / short it" framing is overwrought — CarMax is the largest used-vehicle retailer in the US with ~4% share of a fragmented $800B+ market, and the fragmented mom-and-pop competition genuinely is worse-positioned in a high-rate environment. Carvana's near-death and recovery didn't kill KMX; the shared enemy is affordability, not disruption. So I dissent from the "existential" characterization while agreeing on direction.
The contrarian bull case worth taking seriously: if you normalize to a mid-cycle 2.5% net margin on $27B revenue, that's $675M NI, or ~$4.30 EPS, putting the stock at ~13x normalized — not screaming cheap but defensible for a market leader with real scale. The narrative layer's "65-70% narrative premium" claim is only true if you refuse to normalize earnings. That said, the bull case requires (a) rate cuts flowing through to CAF spreads within 4-6 quarters, (b) used-car ASPs stabilizing rather than continuing to mean-revert down from 2022 highs, and (c) no credit-loss spike in the subprime book as consumer stress builds. Item (c) is where I'd push hardest — CAF's loan loss provisions have been climbing and a recession would blow a hole in both the retail and finance segments simultaneously. The synthesis undercounts this dual exposure.
Net: the stock is overvalued at $57 on current earnings power but not the disaster the composite $20 fair value implies. My fair value is $40-$45 — roughly 10-11x normalized $4 EPS with a haircut for cyclical risk and the credit tail. That's still ~25-30% downside from here, so directionally I agree with the synthesis but with materially less conviction and a much narrower gap. The 34x trailing P/E is the wrong lens; the right lens is normalized earnings power against a mid-cycle multiple, and even that lens says wait. I'd want to see either (i) a print showing CAF net interest margin re-expanding, or (ii) a re-rating to the low $40s before adding. Shorting here is a bad risk/reward — the 15% FCF yield and cheap P/B of 1.44 mean the downside is bounded and squeeze risk exists on any rate-cut catalyst. The models are directionally right, quantitatively too harsh, and the "platform monopoly" narrative framing is a red herring — KMX is priced as a mediocre retailer having a bad year, not as a SaaS company.
GPT Reading
What stands out to me is that CarMax is being valued on a recovery that is only partially visible in the actual income statement. The business has stabilized, but it has not yet re-earned a premium multiple. Annual revenue has drifted from $31.9B in FY2022 to $25.9B in FY2026, and annual net income has collapsed from $1.15B to $247M over the same span. That is not a temporary paper cut; it is a four-year reset in earnings power. Even if I give credit for the strong latest quarter — $8.01B of revenue and $185.6M of net income versus $7.55B and $210.4M in the comparable quarter a year ago — the company is still running at low-single-digit margins in a business where tiny changes in gross profit per unit and credit losses drive enormous swings in EPS. Paying 34x trailing earnings and 36x EV/EBITDA for a used-car dealer/auto financier with 1.4% operating margin and 1.0% net margin is simply too aggressive.
The key contradiction in the raw data is between earnings and cash flow. On the surface, $1.78B of operating cash flow and $1.24B of free cash flow against an $8.15B market cap looks attractive. But for an inventory-heavy auto retailer, that level of cash generation can be heavily flattered by working-capital release rather than true franchise economics. The annual P&L says gross profit was $2.81B, operating income only $353M, and net income $247M on $25.88B of revenue. Those are the economics of a low-return retailer, not a compounding platform. ROE is 4.2% and ROIC is under 1%, which are the most revealing numbers in the pack: the company is employing a lot of capital for very little residual return right now. If the bull case is “trough earnings,” fine — but the stock price already reflects a substantial normalization. At $57.40, investors are implicitly looking past the current $250M earnings base toward something much closer to the $500M-$1B band. I do not see enough evidence in the annual trend to underwrite that with confidence.
I also think the balance sheet deserves more skepticism than a simple current ratio might suggest. Total debt is $17.48B against just $122.8M of cash and $5.89B of equity. Some of that debt supports financing receivables, so I would not frame it as existential in the sensational way some models do, but it absolutely magnifies cyclicality and limits strategic flexibility. In a business whose annual operating income is $353M, leverage matters even if the asset-liability matching is functional. The problem is not that CarMax is about to disappear; it is that the equity can look optically cheap on sales or book while still being expensive on normalized distributable earnings because so much of the capital structure sits ahead of shareholders. A 1.44x book multiple is not low if the return on that book is only 4%.
The best case against my view is that FY2026 may already be the trough year and the latest quarter hints that the turn has begun. Revenue in the May quarter rose to $8.01B from $7.55B, and net income of $185.6M nearly matched the prior-year $210.4M despite a much uglier FY2026 full-year print. If that quarter is the new run rate, trailing earnings are stale and the 34x P/E collapses quickly. There is also a legitimate argument that the free cash flow matters more than the depressed EPS, because inventory normalization and disciplined capex can support equity value before reported margins fully recover. And compared with FY2023-FY2026 annual revenues clustered around $26B, one can argue the top line has largely bottomed and the next move is margin repair, not demand rescue. I weigh those points less heavily because this company has now shown several years of structurally lower profitability, not just one bad patch, and because the difference between 1% and 3% net margin in this model is the whole valuation. The market is already paying as though that repair is likely.
What would change my mind is evidence that the business can sustainably earn at least $600M-$700M annually again without heroic assumptions. Concretely, I would want to see two or three consecutive quarters with revenue holding above roughly $7B, net margins back above 2.5%, and annualized operating income trending toward $700M+, not just one quarter helped by easy comparisons or working-capital tailwinds. I would also want proof that cash flow remains strong even without inventory release. If CarMax can show that FY2027 earnings power is rebuilding toward the FY2024-FY2025 range of roughly $480M-$500M and then beyond, the stock starts to look more defensible around the current price. Until then, this is a cyclical retailer/financier valued on a recovery that is not yet earned in the numbers.
Grok Reading
CarMax at $57.40 is a mature used-car retailer whose income statement has been in multi-year retreat while the multiple still prices a clean recovery. Annual revenue has slid from $31.90B in FY2022 to $25.88B in FY2026, a cumulative drawdown of roughly 19 percent, with net income collapsing from $1.15B to $247.3M—an earnings CAGR near −28 percent. Operating margin sits at 1.4 percent and net margin under 1 percent; ROE is 4.2 percent and ROIC under 1 percent. Those are trough-cycle returns, not platform-compounder economics. The latest spring quarter ($8.01B revenue, $185.6M net income, 2.3 percent margin) looks better than the prior winter loss of $120.7M, but it is seasonal and still leaves trailing profitability far below the $480–500M net income of FY2024–25 and nowhere near the billion-dollar era. Against that backdrop a 34× P/E, 36× EV/EBITDA, and a price that embeds a large premium to any normalized cash-flow anchor look aggressive.
What the numbers do show on the constructive side is cash generation and a still-intact top-line base. FY2026 free cash flow of $1.24B on an $8.15B equity value is a mid-teens FCF yield before leaning on any multiple expansion, and operating cash flow of $1.78B with only $541M of capex confirms the business can throw off cash even when GAAP earnings are compressed. Gross margin held near 10.8 percent annually, and the most recent quarter’s revenue acceleration is real. Debt of $17.48B against $5.89B of equity and a thin $123M cash balance looks alarming until one remembers a large share is auto-finance related; even so, leverage remains a hard constraint on equity duration if credit losses or funding spreads widen. The story the raw data tells is therefore not collapse but stagnation with a financing-heavy balance sheet: volume and unit economics have not reclaimed mid-cycle levels, yet cash conversion has stayed adequate.
I part company with the most apocalyptic model language—“melting ice cube,” “existential risk,” 66 percent further EPS decline—as overstated relative to the still-national scale, brand, and $25B+ revenue run-rate. I do not part company with the valuation conclusion. Composite fair-value work around the low-to-mid $20s versus $57.40 is directionally right: roughly two-thirds of the current price is narrative (national platform, financing flywheel, share gains as independents exit) rather than capitalized current or near-term cash flows. A 34× multiple on depressed earnings only works if net income re-expands toward $800M–$1B and stays there; the last five years of declining revenue and margin have not demonstrated that path. Sector lag and a consumer-cyclical rate sensitivity reinforce that the burden of proof sits with the bulls.
The strongest case against this read is the FCF. At $1.24B, simple capitalization at a mid-cycle 10–12× FCF multiple supports equity value in a $12–15B range—above the deep-value DCF prints and close enough to today’s $8.15B that a moderate recovery in retail margins plus stable CAF spreads could justify holding rather than shorting. Opponents would also note the May quarter’s $186M profit as evidence the trough is behind, argue that P/E on trough earnings systematically overstates expensiveness, and point to the platform-monopoly narrative’s durability among GARP allocators who will keep a bid under the stock as long as used-vehicle prices stabilize. I weigh those points as real but incomplete: FCF in a finance-linked auto retailer is partly working-capital and securitization timing, not pure owner earnings, and one strong seasonal quarter does not reverse a five-year earnings CAGR of −28 percent or a sub-1 percent ROIC. The narrative premium remains fragile to credit tightening or a renewed slide in used-car ASP.
I would flip toward neutral or constructive if two consecutive quarters delivered net margins back above 2.5 percent with revenue stabilizing above $7B, or if FY2027 net income guidance and CAF credit metrics pointed credibly at $600M-plus of sustainable earnings without a spike in loss rates. A material debt refinance that cut interest burden or a sustained FCF print above $1.5B with deleveraging would also force a re-rating of the equity value floor. Absent those, the stock at $57 remains a recovery story priced as if the recovery were already earned.
Big-3 Panel — where each AI stands
Advanced Analysis Forensic deep-dive · separate lenses
CarMax is a mature, scaled auto retailer with modest but real free cash generation (FCF $1.24B in FY2026 after a stretch of negative-to-flat FCF in 2022-2024) and a disciplined share count that has fallen from 165.2M to 147.6M diluted shares, a -2.8% CAGR with buybacks running 378% of SBC. Earnings quality looks clean on the surface: OCF/NI 1.97x, accruals only 1% of assets, and a Beneish M of -2.7 indicating no manipulation flags. However, the operating trajectory is the real story: revenue has contracted from $31.9B (2022) to $25.9B (2026), operating margin has compressed from 3.0% to 1.4%, and net income has fallen from $1.15B to $247M - a roughly 78% earnings decline over four years, only partially explained by industry-wide used-vehicle normalization.
Verify before trusting this (6)
- Composition of the $17B+ debt: how much is non-recourse CAF warehouse/securitization vs. corporate debt with recourse to the retail business
- CAF loan-loss provisioning trend and delinquency rates - a key earnings swing factor not visible in the summary financials
- Gross profit per unit (retail and wholesale) trend versus the 2022 peak to gauge whether margin recovery is structural or cyclical
- Inventory days and any writedown risk given used-vehicle price deflation
- Capex composition (maintenance vs. new store growth) to judge sustainability of the FY2026 FCF print
- Any covenants on securitization/warehouse facilities that would constrain capital returns
The e2e composite pegs deserved value at $27.59 and the signal-adjusted FV at $20.17 against a $57.40 price - implying roughly 50-65% downside if you take those methods at face value. Even the most generous input, the DCF at $38.98, still sits 32% below the current quote, and the EPV floor of $18.41 and anchored P/E of $14.00 corroborate that the through-cycle earnings power does not justify today's cap of $8.1B. Sanity check: the methods cluster in the high-teens to high-thirties, so this is not a single runaway model - three independent frames all say rich. On fundamentals the market is capitalizing CarMax like a compounding platform while operating margin is 1.4%, net income is a quarter of 2022, and revenue has declined three years running. To defend $57 you need a sharp used-price stabilization, CAF finance-book expansion without credit losses, and margin recovery back toward 2021-22 levels - a stack of 'has to go right' outcomes. The quality lens is Mixed, not Great, so there is no premium-business argument to bail out the multiple. Modest earnings-quality haircut reinforces, not offsets, the gap.
Verify before trusting this (5)
- CAF loan-loss provisions and 60+ day delinquency trend in latest 10-Q
- Retail units sold and gross profit per unit trajectory
- Used-vehicle wholesale price index (Manheim) direction
- Management guidance on SG&A leverage and store-growth capex
- Any one-time items inflating recent GAAP EPS
The macro tape is mildly risk-on with a contained VIX at 15.8, and KMX's 1.16 beta means it gets a modest lift from that backdrop rather than a hammering. More importantly, the active narrative around KMX has quietly shifted from 'cyclical used-car retailer' to 'platform-monopoly' with strong intensity and medium cult - the market is willing to pay a SaaS-like multiple on the story that CarMax is the only national systematized used-car platform. That is a real, durable-feeling sentiment tailwind independent of what the fundamentals justify. The most telling live signal is the news flow: Carvana just posted a blockbuster quarter and is DOWN 7%, while the article explicitly flags that 'a traditional rival is up over 50% this year' - that rival is KMX. That is the market voting with its feet for the platform-incumbent story and against the pure-digital challenger. Rotation within the used-auto cohort is running in KMX's favor right now. The counter-pressure is that the story is running well ahead of DCF (price 57.40 vs DCF 20.17), which makes the narrative fragile to any crack - a credit tightening headline, a used-car price roll, or a soft comp print could snap the multiple fast. Analyst tone is not flagged as extreme either way, so there is no visible upgrade cycle amplifying the move; the tailwind is narrative- and rotation-driven, not broker-driven, which limits its magnitude.
Verify before trusting this (4)
- Whether the Carvana-versus-KMX rotation persists after the next KMX earnings print or reverses on a soft comp
- Any subprime auto credit or delinquency headline - the fastest way to crack the platform narrative
- Used-vehicle price index direction; a renewed decline would revive the 'just a cyclical' framing
- Sell-side target revisions - a visible upgrade cycle would extend the tailwind; downgrades would snap it
The need to buy a used car is untouched and the transaction is physical — title, inspection, recon, warranty, financing — so AI cannot disintermediate the fulfillment layer, and that is why exposure is moderate rather than high. But CarMax's historical excess return did not come from fulfillment; it came from knowing what a specific car was worth better than the seller and from being the trusted, hassle-free brand in an opaque market. Cheap intelligence attacks precisely those two rents: appraisal models become table stakes for every dealer and instant-offer platform, and AI shopping assistants convert trust into a commodity price comparison across all national inventory. Meanwhile AI's cost benefit is real but shared — Carvana-style lower-fixed-cost operators capture more of it per unit, and with 1.4% operating margin CarMax has little room to absorb GPU compression. The offset is CAF, where proprietary loss data plus better underwriting is a genuine, under-modelled profit lever.
Verify before trusting this (8)
- Buy-side spread per acquired vehicle
- Self-sourced acquisition mix and cost
- Third-party appraisal tool adoption
- SG&A per retail unit trend
- Headcount per store per unit sold
- GPU stability while SG&A falls
- US used SAAR and units sold
- Household vehicle ownership rates
Used-vehicle retail is normalizing out of the 2021-23 price bubble: vehicle values have deflated toward trend, which restores affordability and unit demand but compresses the per-unit gross dollars that made the bubble years look profitable. The structural change around KMX is distributional, not existential — cars still get bought and sold, but the transaction is migrating to whoever can source, recondition and deliver at the lowest cost per unit, and that contest is being won on logistics density and cost, not brand trust. Financing is the second front: with the 10-year at 4.72% and a neutral-but-tightening credit backdrop, captive finance economics decide the earnings line more than retail units do. Tariff-driven new-vehicle price increases are a genuine tailwind to used demand — they push buyers down-market into KMX's inventory — but they also raise trade-in acquisition costs. Net: the world is getting better for used-car volumes and no better for used-car margins, which favors the low-cost operator over the scaled incumbent.
When we made this prediction on Aug 19, 2026, KMX was $59.51. We expect it to be $47.50 by Feb 2027, and we consider it great value under $32.00. This is an early model (v0.6.0) — the direction is more reliable than the exact price. Made Aug 19, 2026.
Blue is our prediction, starting the day we made it. Grey is a slower route to the same place — the same destination, taking longer. Black is the actual price, so you can see how we are doing.