AZO: The Margin Rebound Has a Refund Inside

The 60-second read

Five things to know

1. The margin rebound has temporary contributors. Tariff refunds and the comparative LIFO effect total 250 bp against a 182 bp gross-margin increase.
2. Commercial scale is outpacing productivity. Sales grew 8.6%, but weekly sales per program rose 2.7% as the program count expanded.
3. Cash after investment barely grew. FY2026 OCF less capital spending was $1.807B versus $1.790B, despite 7.4% sales growth.
4. Buyback discipline matters. About $2.026B was repurchased at an average $3,496, above the pre-release reference.
5. The price already allows for some recovery. Three valuation routes center near $2,800, with a $2,100–$3,900 scenario range.

Three scenarios, one probability-weighted number

Analyst assumptions, not company targets

Signposts: what would change our mind

Dates are analytical review deadlines. Compare like-for-like periods and reconcile the next filing before grading.

The price reset is already substantial

The reference price is the September 21 close of $2,803.25. It precedes this earnings release. The saved quote is timestamped at the prior session close, so its daily change cannot honestly be presented as the market's reaction to Q4. The research was prepared from the new release before the scheduled conference call; it does not assume that management answered any of the questions below. A later price would change the margin of safety, not the underlying statement arithmetic.

AutoZone is now priced much closer to ordinary retailer earnings multiples than its premium reputation might suggest. The stock is below the saved 52-week high of $4,332.68, but the distance from an old high is not a valuation method. A business can become cheaper while its reinvestment opportunity becomes less productive. The central question is whether the hub and store investment cycle will restore growth in cash available to owners, rather than simply expand the asset base.

Using the release's 16.173 million period-end shares gives $45.337 billion of equity value. Adding $8.752 billion of financial net debt produces $54.089 billion of enterprise value. These calculations deliberately use the primary share count rather than the provider's $45.764 billion market-cap field, which uses a different denominator. Diluted weighted-average shares are suitable for historical EPS; period-end shares are the transparent starting point for a spot equity valuation.

The company does not pay a common dividend, so the ownership return depends on earnings growth, repurchase discipline and valuation changes. A low quoted beta does not eliminate business risk or establish a downside floor. Neither the stock's prior high nor Wall Street's higher targets substitutes for cash generation.

Better headline earnings, less obvious operating progress

The quarter is sixteen weeks, and the prior-year comparison is also sixteen weeks. Revenue increased 5.6% to $6.595 billion, operating income rose 10.1% to $1.317 billion, and diluted EPS increased 15.1% to $56.05. Those are real GAAP results. The analytical problem is the mix of contributors to that growth, not whether the reported figures exist. The full fiscal year comprised fifty-two weeks, which must be distinguished from individual quarters when looking at sequential changes.

The release attributes 145 basis points of quarterly gross-margin benefit to tariff refunds and 105 basis points to the year-over-year noncash LIFO effect. Together those exceed the reported 182-basis-point gross-margin improvement. Subtraction leaves a negative 68-basis-point residual associated with the remaining factors, including the higher commercial mix. That does not constitute a complete company-adjusted income statement. It is a transparent way to see why headline margin expansion cannot all be extrapolated.

FMP's current earnings record presents EPS on a conflicting scale, and its derived quarterly statement has different weighted shares and EPS from the issuer table. We therefore do not publish an EPS beat percentage. Revenue consensus in the same saved record is $6.670 billion, compared with $6.595 billion reported, but the corrupted EPS field reduces confidence in treating that row as a clean all-purpose surprise table. The original release remains the controlling evidence.

Quarterly net income rose faster than operating income, and diluted EPS rose faster again as the share denominator fell. The decline from 17.181 million to 16.620 million weighted diluted shares aided per-share growth. That is legitimate financial engineering when shares are repurchased at attractive values; it is not a substitute for improving the economics of stores already open.

Commercial growth must earn its inventory

AutoZone reports a single operating and reportable segment. The FY2025 annual report says the company combined its previous operating segments to align with how management evaluates the business. Geographic store counts and commercial sales are useful operating disclosures, but they are not independent segment profit statements. We do not assign invented margins to Mexico, Brazil or the do-it-yourself business simply to fill a chart.

Domestic commercial sales reached $1.913 billion in the quarter, up 8.6%. That is stronger than domestic same-store sales growth of 1.6%, but the distinction matters. Commercial growth includes additional programs and their maturation. Stores with commercial programs increased from 6,098 to 6,443, while average weekly sales per program increased from $18.2 thousand to $18.7 thousand, or the company's rounded 2.7%. The network is expanding faster than the productivity of each program. Both are valuable, but they carry different capital requirements.

The quarter added 175 stores, including 97 in the United States, 68 in Mexico and ten in Brazil. Management also reported sixteen new U.S. Mega Hubs. More local availability can raise the probability of winning an urgent repair order, particularly when professional customers value reliable delivery. The cost comes first through inventory, people, facilities and distribution. A larger network will not automatically generate a proportionate increase in cash if demand remains modest.

International same-store sales grew 10.7% in reported dollars but only 1.3% in constant currency. Translation therefore makes the international headline much stronger than the local operating comparison. The relevant forward test is sustained local demand and store productivity, not an assumption that favorable currency will repeat. The domestic and international businesses can both remain competitive while the consolidated growth rate slows.

The refund and LIFO bridge changes the reading

The gross-margin bridge is the core of this episode. Tariff refunds are an economic benefit when the company is entitled to receive them, but their recurrence depends on facts that should not be assumed. LIFO changes affect the timing of inventory cost recognition and are explicitly described as noncash in the release. Neither should be treated as a new permanent pricing advantage merely because the quarter's gross margin improved.

Multiplying current sales by the disclosed rounded basis-point impacts gives approximately $95.6 million for tariff refunds and $69.2 million for the comparative LIFO benefit. Removing those comparative contributions from reported operating profit produces an illustrative $1.152 billion, about 3.7% below the prior-year $1.196 billion. This bridge is an analyst sensitivity, not a reported non-GAAP result and not a claim that all other costs stay fixed in the real business. The percentages are rounded; the illustration should not acquire false precision.

Operating expenses took 33.4% of sales versus 32.4% last year, with management attributing deleverage primarily to growth initiatives. That is the second important piece. Even if the investment is sensible, investors are paying today's expenses before the future sales are proven. Strong commercial sales alongside lower mix margins and rising operating costs can create a period in which sales growth does not translate into superior returns.

The annual comparison provides a useful restraint on the quarterly story. FY2026 sales rose 7.4%, but operating profit grew only 3.1% and net income about 3.0%. EPS rose 5.3%, helped by fewer shares. Annual gross margin declined despite the tariff-refund benefit, as the annual LIFO comparison was unfavorable. Quarter and year tell different timing stories, and neither should be discarded to make the thesis simpler.

Cash quality needs a working-capital lens

Operating cash flow of $3.303 billion exceeded annual net income of $2.573 billion, a healthy first check. After $1.496 billion of capital spending, however, free cash flow was $1.807 billion, only modestly above the prior year's $1.790 billion on the release's comparative basis. Free cash flow here means operating cash flow less capital spending, computed by Charged Alpha. It is not a separate GAAP line and does not include a judgment about which investments could safely be stopped.

Inventory grew 10.1%, faster than annual sales growth of 7.4%. That can be a deliberate service investment, especially as new stores and Mega Hubs stock a broader range of parts. It also raises the burden of proof: items must turn into sales at adequate margins before the investment earns its cost. The disclosed inventory-turn measure declined from 1.4 to 1.3 times. Inventory days calculated from year-end inventory are shown only as a directional check, not as an average-balance operating metric.

Accounts payable still exceeds inventory, which is an important structural advantage. Supplier financing supports a wide product assortment without requiring shareholders to fund every dollar on the shelf. Yet payable coverage declined from 114.2% to 111.1%, and negative net inventory per store narrowed from $131 thousand to $107 thousand. A negative number becoming less negative means more net cash is tied up relative to the prior relationship.

There is a presentation difference between the current release and the original FY2025 audited cash-flow statement: both the operating cash and capital-spending comparators are $38.064 million higher in the new release. The free-cash-flow difference is zero. We preserve both source bases and avoid claiming that the reclassification itself produced additional cash. The full FY2026 filing should explain the complete cash-flow presentation before finer working-capital claims are made.

Supplier funding is an advantage with obligations

Financial debt was $9.078 billion against cash of $326 million at year-end. The resulting $8.752 billion net-debt balance is material, but it should be interpreted alongside the recurring parts business and the company's stated investment-grade capital objective. Cash alone is not the full liquidity picture; equally, a durable business model does not make debt free. Net interest expense consumed $473 million of the year's earnings before tax.

The company reports adjusted debt to EBITDAR of 2.5 times. Its definition adds finance leases and six times rent to debt, then compares that balance with earnings before interest, tax, depreciation, amortization, rent and share-based expense. That ratio is useful for following management's own capital framework, but it is not the same as conventional debt to EBITDA. Our enterprise-value routes retain rent as an operating cost and exclude operating lease liabilities from financial net debt; this avoids charging rent once in earnings and again through the full lease balance.

The FY2025 annual report disclosed approximately $5.699 billion of confirmed supplier-financing obligations, including amounts classified outside current accounts payable. Participation is optional for suppliers, and the annual report says the company's original payment obligations are not changed by a supplier's choice. Nonetheless, the ability of suppliers to finance receivables and the willingness of financial institutions to support them remain relevant to the working-capital model.

Repurchases cost about $2.026 billion during FY2026, compared with our $1.807 billion free-cash-flow calculation. The average price paid was $3,496, well above the pre-release reference. That does not prove every repurchase was wrong; hindsight prices are not the same as contemporaneous value. It does mean that shrinking the denominator should be judged alongside price paid, funding and foregone flexibility. Negative book equity mainly reflects the accumulated capital-return structure and is not a stand-alone insolvency test.

Three valuation routes point near the current price

Our call is HOLD, conviction three out of five, with High uncertainty and a base fair value of $2,800. The scenario range is $2,100 to $3,900, and probability-weighted value is $2,810. These are analyst judgments, not company guidance or a promise about a twelve-month trading price. A twenty-percent margin of safety against base value implies a conditional $2,240 entry reference. The call can change above that price if operating evidence improves enough to increase value.

The earnings route assumes $3.900 billion of forward operating income, $470 million of net interest, a twenty-one-percent tax rate and 16.173 million shares. That produces $167.54 of earnings per share. At eighteen times earnings, the value is $3,015.80. This assumes an improvement in operating income and does not credit future share repurchases without paying for them. The multiple is a judgment reflecting the company's durable service network and the current investment burden.

The enterprise-value route applies thirteen times the same $3.900 billion operating income, subtracts $8.752 billion of financial net debt and divides by the same shares. The result is $2,593.69. Rent remains within operating expenses, and operating leases are excluded from financial net debt for consistency. This route is deliberately more conservative than a premium earnings multiple, exposing how debt changes the value attributable to shareholders.

The DCF starts with assumed normalized unlevered free cash flow of $2.700 billion, grows it six percent annually for five years, discounts at nine percent and applies three-percent terminal growth. The result is $2,719.74 per share after financial net debt. The starting cash assumption is above current post-investment owner cash plus after-tax interest: it requires some investment normalization and operating recovery. It is not an observed run rate. Present value of terminal cash exceeds the explicit-period value, so small changes in discount and terminal growth assumptions matter greatly.

The three routes average $2,776.41, rounded to the published $2,800 base. Their agreement is not independent proof because all depend on the same business recovering acceptable returns. At the current enterprise value, thirteen times EBIT would require about $4.161 billion of operating income, above our $3.900 billion assumption. Alternatively, at eighteen times earnings the price implies about $155.74 EPS. The market is therefore demanding progress, but not an impossible bull case.

Wall Street has more recovery priced in

The saved provider consensus target is $3,664, with a low of $3,200 and high of $4,023. These are a provider snapshot, not targets all refreshed after the September 22 release. The most recent clearly identified records are pre-print reductions: Evercore ISI to $3,500 on September 21, Oppenheimer to $3,500 on September 18 and Wells Fargo to $3,500 on September 10. The direction of these recent changes matters more than presenting an old average as a new endorsement.

The rating snapshot contains one Strong Buy, thirty-two Buy and twelve Hold classifications, with no Sell classifications. We show those categories as delivered rather than assume each corresponds to a separate current analyst. Coverage databases can mix update dates and rating histories. A favorable distribution describes professional opinion; it is not an earnings input or a reason to raise a valuation mechanically.

The annual estimate table shows about $21.868 billion of FY2027 revenue and $175.09 EPS, with twenty EPS analysts. That EPS is above our $167.54 operating model, and the release itself does not provide a quantified FY2027 EPS target. Management expects sales in all three countries to accelerate, but translating that into profit requires assumptions about commercial mix, new-store expenses, LIFO, refunds, interest and repurchases. We keep the Street projection, management language and our model on separate lines.

Some individual target-news rows contain a mismatch between the firm field and the headline. Those rows are preserved in the dataset but excluded from the named comparison. The current provider EPS surprise is also excluded because of the scale conflict. This conservative treatment may leave a less dramatic comparison, but it avoids building a valuation argument on a false impression of precision. A meaningful post-call update would require newly dated analyst evidence, not a refresh of the same stale average.

Grade investment on returns, not square footage

Management's account of the quarter describes a difficult first eight weeks followed by stronger sales in the last eight. That direction is encouraging, but the release does not quantify the second-half improvement or provide a full monthly trajectory. We therefore treat it as a management observation awaiting corroboration, not an independently measured acceleration rate. The next earnings report should show whether the improvement carried into the new fiscal year.

The strategic argument is coherent: broaden inventory, increase local availability and improve delivery speed so professional repair shops can complete jobs quickly. The risk is timing and scale. If demand grows slowly, the expanded network carries inventory and labor before its productivity matures. Store openings and Mega Hubs are leading activity measures. Returns on invested capital, program sales per week and cash conversion are the outcome measures that determine whether shareholders benefit.

Company-adjusted ROIC declined from 41.3% to 35.8%. That remains a high absolute return, but the direction deserves attention because the invested-capital base has expanded. The company uses a lease-adjusted calculation based on trailing-five-quarter averages, so it is not interchangeable with a generic screen's ROIC. We retain the company's definition and interpret the change alongside the investment cycle rather than compare it casually with an unrelated retailer's ratio.

The capital-allocation score is mixed. Maintaining an investment-grade objective and strong supplier relationships supports the model. Repurchasing more than current free cash flow while reinvestment is elevated reduces flexibility. The most useful management questions concern incremental hub returns, recurring margin after tariff effects, the working-capital cost of expansion and the evidence behind the expectation for acceleration in each country. A larger authorization would not answer any of those questions.

The risks reinforce one another

The first risk is weak underlying demand. An aging vehicle fleet can support replacement-parts demand over time, but it does not prevent short periods of customer deferral, difficult weather comparisons or discretionary pressure. When comparable sales are modest, the fixed cost of a larger distribution and store network can weigh on earnings. This is a high-likelihood operating sensitivity rather than a prediction of a recession.

The second risk is extrapolating favorable accounting and tariff effects. A refund can improve a quarter without creating a recurring source of margin. LIFO movements can reverse and can obscure the relationship between current purchase costs and reported cost of goods sold. The third risk is investment productivity: new hubs and stores may take longer to mature, leaving inventory elevated and cash conversion weaker than expected. These risks interact because slower demand delays the payback from assets already placed in service.

The fourth risk is financing and working capital. Accounts payable funds more than inventory, but a deterioration in supplier terms or access to supplier financing would increase cash requirements. Leases and debt add fixed commitments even when a simple net-debt ratio appears stable. Currency is a separate source of volatility: reported international sales can change without an equivalent change in local demand, as this quarter's reported and constant-currency comparisons demonstrate.

The bull case is substantial. The service network may continue taking commercial share, investments may mature into higher utilization, and a better sales environment may spread operating costs across more volume. Under our bull assumptions, $195 of earnings at twenty times supports $3,900. The bear case assumes $150 at fourteen times, or $2,100. Those are explicit sensitivities with judgmental probabilities, not statistical confidence intervals or worst-case bounds.

Use the next evidence to update the call

The immediate information event is the September 22 conference call, scheduled in the release for ten in the morning Eastern time. This packet does not claim to have heard it. The call could clarify tariff refunds, LIFO assumptions, the pace of recovery in the final eight weeks, and the near-term burden of new infrastructure. A useful answer would quantify these items or define a measurable range; a repetition of optimism would not justify changing the model.

The FY2026 Form 10-K is also pending at the research cutoff. It should supply the complete cash-flow statement, more detailed balance-sheet accounts, lease information, supplier-finance balances and other disclosures omitted from the release. Its exact publication date is not asserted. We will reconcile the release to that filing rather than assume that a provider's zero-filled statement is a valid substitute. Any material new disclosure would require an update to the affected valuation inputs.

The next quarterly release is the practical checkpoint for the strategy. We use December 31 as an analytical review deadline, not a claimed company-announced earnings date. Green thresholds include domestic same-store sales above three percent, commercial weekly sales per program above $19 thousand, and better evidence that inventory growth is converging toward sales growth. Red thresholds include domestic same-store growth below one percent, additional SG&A deleverage or a deterioration in payable coverage below 108%.

Investors should also distinguish price catalysts from business catalysts. A sharp price decline can create a margin of safety without changing the business; a reassuring call can improve sentiment without changing cash value. Our base call changes when either the discount is wide enough for the uncertainty or new evidence materially raises expected cash. That framework is intentionally more demanding than calling every target reduction bearish and every store opening bullish.

Sources, comparability and calculation notes

Current-quarter and fiscal-year figures come from AutoZone's September 22, 2026 Form 8-K, Exhibit 99.1, for the sixteen-week quarter and fifty-two-week year ended August 29. The release tables place the current period before the prior period. Dollar statement entries are in thousands in the original document; this packet converts them to millions, except per-share values. Shares are expressed in millions. Original HTML, saved provider responses, retrieval metadata and hashes are retained with the episode research.

The historical nine-quarter table uses the saved FMP statement series, with the current quarter overwritten by the primary release where the provider conflicts. Older quarterly EPS and revenue series can carry fiscal-calendar comparability issues; Q4 contains more weeks than the first three quarters, and FY2024 was a longer fiscal year. Sequential quarter-to-quarter growth is therefore not used as evidence of operating momentum. Current and year-earlier Q4 comparisons have matching sixteen-week durations.

The audited FY2025 annual report supplies the segment definition, auditor, employee count, supplier-financing structure and historical cash-flow detail. It is a prior-period source, not proof that every structural number is unchanged in FY2026. Historical source data includes balance-sheet and below-the-line statement fields, but missing current receivables, complete cash-flow components and longer operating-KPI histories remain explicitly unavailable. We do not fill those gaps with zeros or interpolation.

Derived margins divide profit by revenue; free cash flow subtracts capital spending from operating cash flow; financial net debt subtracts cash from reported debt. The ending-inventory days calculation uses 364 days because FY2026 is fifty-two weeks. It is not average inventory days. Valuation assumptions, scenario weights and safety thresholds are ours. A dollar per-share chart uses plain dollar units, while operating amounts use millions. No split-scale EPS inference or unverified post-print price move appears in the analysis.

This is the first finalized packet in this episode workflow for the current quarter, so there is no prior packet signpost set to grade. The recovered Q3 video is retained as historical content and its beat-and-crash headline is not reused. Future packets should grade the dated signposts here and compare the old fair value with the subsequent evidence, including cases where the thesis proves wrong. The aim is an auditable decision process, not an unchanging recommendation.

The original issuer release describes financial debt, adjusted debt and EBITDAR separately. They are not interchangeable. No separate geographic profit series is disclosed, and this packet therefore uses commercial and store metrics without manufacturing segment economics. The scorecard flags absent data as unavailable, not as a pass.