Academy Sports (ASO): the tariff cash boost behind the earnings jump

The 60-second read

Five things to know

1. Growth is real, but uneven. Total sales increased 3.0% while comps slipped 0.4%. The store base reached 327 and online sales grew 12.8%. Subtracting comp from total growth is not an exact new-store revenue bridge because the populations and denominators differ.
2. Profit did more of the EPS work than the share count. Net income grew 9.94%; diluted shares fell 6.11%. On an old-share-count-first bridge, higher profit adds about $0.18 per share and the lower denominator adds about $0.13. Neither effect should be mistaken for the other.
3. The tariff benefit is not a non-GAAP add-back. The $0.06 net EPS benefit remains in reported and adjusted performance. The actual adjusted-income bridge adds equity compensation and a debt-retirement loss, then deducts their tax effects.
4. Cash needs a wider lens. H1 operating cash includes tariff refunds, while $72.2M of tariff-claim remittances sits in financing. Reported adjusted FCF is $237.6M; subtracting that financing outflow gives a $165.3M cash illustration, not a measure of recurring FCF.
5. The rerating has used much of the valuation cushion. At our approximately $49.99 snapshot, the $50 central value offers almost no upside. The $33/$53/$68 scenarios have a $50 weighted value. A profitable retailer can still be an unattractive fresh purchase when uncertainty is high and the cushion is thin.

Three scenarios, one probability-weighted number

Twelve-month price scenarios; probabilities, earnings assumptions and multiples are Charged Alpha judgments, not guidance.

Signposts: what would change our mind

Dates for future earnings are estimates until the company announces them. These thresholds are our decision rules, not management promises. The next ASO packet should grade them explicitly. There is no previously published ASO research packet in our registry to backfill with an invented score.

The tape: price matters as much as the print

The important market question is how much of the improvement is already paid for. The previous close was $44.71; this edition uses a refreshed intraday valuation snapshot near fifty dollars. That change matters because the central appraisal remains around fifty: a price increase can consume the investment opportunity even when the underlying business has improved. We freeze the snapshot for internal consistency rather than allowing different sections to use different live prices.

The chart provides context, not a forecast. Its daily observations and the intraday quote use different clocks, so we do not describe the latest intraday move as a completed post-earnings session. Likewise, fiscal-year earnings and a quarter-end share count are explicitly labeled rather than marketed as current forward consensus. A reader updating the price can recompute the capitalization and multiples from the equations shown here without changing the analytical assumptions.

Our previous ASO episode covered the first positive comp in two years in Q1 FY2026. Q2 brings a different question: how much of the stronger reported earnings and cash can be carried forward? We do not recycle the old positive-comp framing after the comparable-sales line has turned negative. The historical presentation remains available and unchanged; this packet records a new period and a new assessment.

The print and what changed in guidance

Sales growth and earnings growth answer different questions. Revenue says customers collectively spent more with Academy; the negative comp says growth was not broad enough to lift the comparable population. Net income then captures merchandise economics, operating costs, financing, other items and taxes. Finally, EPS divides that profit among a smaller diluted share base. Collapsing those steps into one headline obscures the fact that both profit and the denominator improved this quarter.

In this explicit ordering, profit contributes roughly eighteen cents and the share count roughly thirteen cents. The unrounded EPS bridge grows about seventeen percent; the company’s seventeen-point-three-percent headline divides the two rounded EPS figures. That rounding difference is not an accounting error. Neither is the share-count benefit fake: existing investors really own a larger fraction of earnings. The caution is that repurchasing stock consumes cash and cannot be treated as a free, permanent growth engine.

The guidance table is more revealing than the phrase “EPS raised.” Sales and net-income ranges are unchanged, while the assumed diluted share count falls. Thus, the EPS upgrade should not be narrated as a new dollar-profit upgrade. Gross-margin guidance and adjusted cash-flow guidance do rise, but the gross-to-net gap and tariff cash treatment require their own explanations. Because guidance is forward-looking, we retain the ranges and do not turn the midpoint into a promised outcome.

Stores, online and the limits of the split

Academy is analyzed here as one reportable operating business. We do not invent merchandise-segment profit bars when the release does not provide category profit. The useful operating disaggregation is instead store expansion, comparable sales, e-commerce and inventory productivity. These describe different populations and must not be added as if they were separate revenue segments. In particular, e-commerce growth is not incremental to total growth: online sales are already included in it.

Twenty-one more stores at quarter end is a larger percentage increase than revenue growth, but it would be misleading to divide revenue by ending stores and call that mature-store productivity. Locations opened at different times, have different sizes, and have not all traded through a full quarter. The same caution applies to the simple gap between total growth and comps. The direction supports an expansion contribution, but the release does not provide the exact weighted bridge needed to assign every percentage point.

For the next quarter, the operating question is whether growth becomes less dependent on adding square footage. Positive comps alongside the planned openings would strengthen the case. If comps remain negative, investors need evidence that new locations earn adequate returns after construction, staffing, working capital and lease commitments. Sales alone are not a store-return calculation. We would rather leave an unavailable category margin blank than create a visually attractive but unsupported decomposition.

Profit bridge: follow the money below operating income

The headline gross margin reaches 40.4%, versus 36.0% a year earlier. Operating income rises strongly, but $58.0M of other expense and a debt-retirement loss sit below it. The cash-flow statement also identifies a $61.8M loss on tariff-refund monetization. These disclosures make it unsafe to extrapolate operating income in isolation or to describe the quarter as an uncomplicated merchandise-margin expansion. They also do not justify moving the entire tariff effect into the non-GAAP reconciliation.

Management separately discloses a six-cent net tariff-refund EPS impact, including reinvestments. Subtracting six cents from adjusted EPS gives an analyst normalization of $2.25, and subtracting it from GAAP EPS gives $2.11. Those calculations help size the net benefit but are not company-reported alternative EPS measures. They do not remove every gross tariff transaction, nor can the six-cent net effect be substituted for the $61.8M cash-flow add-back. Net economic benefit and gross accounting movements are different quantities.

Stock compensation is another economic distinction. It is noncash in the current period and the company adds it back, but employees receive value and shareholders bear dilution or the cost of offsetting repurchases. We therefore retain GAAP earnings as a valuation anchor, show the adjustment transparently and do not award an automatic premium to the adjusted number. Debt-retirement expense is less likely to recur each quarter, yet the refinancing that produced it has an ongoing financing cost.

Earnings quality and the cash classification test

Cash conversion looks strong at first glance because operating cash exceeds net income. That comparison is insufficient this period: the cash-flow statement adds back the monetization loss and places tariff-claim remittances in financing. We show both the company definition and the broader cash perspective. We cannot certify recurring cash flow by simply removing one convenient item, because refund timing, receivables, working capital and the monetization arrangement may interact. The next filing must clarify the remaining exposures.

The annual record also matters. Fiscal 2025 simple free cash flow was $222.1M, below the roughly $328M levels in the two preceding fiscal years. That decline prevents us from capitalizing the unusually strong first-half cash result as if it were a permanent new annual base. New stores require investment before they mature. Healthy unit growth can therefore coincide with weaker cash conversion, and investors should test whether the future earnings justify that capital rather than reward openings mechanically.

Inventory rises 4.4% in total, slightly faster than sales, while dollars per store fall 2.3%. Both statements can be true as the footprint grows. The per-store disclosure is reassuring about aggregate stocking discipline but does not rule out category markdowns or out-of-stocks. Inventory days use ending balances and tariff-affected COGS, so their apparent movement is not a clean operational verdict. We use the metric as a prompt for questions and compare it with comp demand, availability and subsequent markdown commentary.

Balance sheet and capital allocation

The balance sheet has manageable borrowing debt, but describing the company as nearly debt-free would omit substantial contractual rent obligations. Operating lease liabilities are already recognized on the balance sheet; they are not hidden off-balance-sheet debt. Our enterprise-value route excludes them and uses earnings after rent, avoiding a mismatch between lease-adjusted debt and unadjusted operating earnings. A fully lease-adjusted valuation would need corresponding rent and earnings adjustments, not merely a larger debt subtraction.

The financing table corrects the idea that debt simply sat unchanged. Academy refinanced: it issued notes, repaid notes and term borrowings, and paid fees. Similar ending debt balances do not mean no financing activity occurred. The interest burden and debt terms remain relevant even when net leverage is modest. We therefore distinguish the balance at a date from the gross flows over a period, and carrying-value net debt from face-value principal.

H1 repurchases are $182.1M including excise tax in the release headline, versus $180.5M on the cash-flow statement excluding that tax. For a comparable quarterly subtraction, the latter minus Q1 cash-flow repurchases of $98.4M gives $82.1M in Q2. Mixing tax-inclusive and tax-exclusive figures produces a spurious quarterly number. Buybacks may create value below intrinsic value, but their existence does not prove the purchase price was attractive. The capital-allocation test is return on cash after funding stores, leases and other obligations.

Three routes and the market-implied hurdle

At $49.99, the price is approximately 8.42 times our $5.94 normalized trailing EPS illustration. That EPS uses reported quarterly figures of $1.05, $1.98, $0.80 and $2.17, then subtracts the disclosed six-cent net tariff benefit. Summing quarterly EPS is an analytical approximation because weighted share counts vary; it is not the company’s annual EPS. The low multiple says the market already doubts the durability of earnings, but a low multiple alone does not establish underpricing.

These routes are related, not independent votes. Earnings, operating profit and free cash flow come from the same business; agreement does not eliminate model risk. We select a rounded $50 central appraisal and a $46–55 working range because the approaches bracket that area. The scenario distribution remains much wider. With high uncertainty, our chosen twenty-percent margin of safety suggests a price around forty dollars before considering a new Buy, provided the operating case has not weakened.

The market-implied test is simple: a fifty-dollar price at nine times earnings requires about $5.56 of sustainable EPS. That hurdle is below the updated GAAP guidance midpoint, but reported guidance may contain effects that do not persist indefinitely. A bull must explain why recurring earnings deserve a higher multiple; a bear must explain why the current profit base erodes. Our Hold recognizes a plausible base case while refusing to treat the guidance headline as a guaranteed recurring cash stream.

Our appraisal versus dated Street targets

Analyst targets are dated opinions. The saved FMP history contains ten selected firm observations with a mean target of $57.50, all before this earnings release. The newest observation in that saved set is BMO’s September 8 initiation at $42. We retain their dates and do not relabel them post-print research. Differences between firms do not establish a time trend; a sequence from a bullish firm to a cautious firm is not the same as one analyst repeatedly cutting a target.

Our appraisal is below that dated average, but a lower target does not automatically mean a better forecast. The useful comparison is the assumption gap: are analysts capitalizing the updated guidance, assigning a higher long-run multiple, or expecting healthier comps? Without their complete models, we cannot answer on their behalf. We label the target set as context and make our own arithmetic available so the disagreement can be evaluated.

A fresh post-earnings target is useful only if its reasoning incorporates the negative comp, lower share assumption and tariff cash treatment. We do not infer an analyst’s Buy or Hold recommendation from whether a price target rose or fell. Rating distributions, firm targets, management guidance and our valuation are separate datasets. Where an authenticated provider snapshot does not contain a current-period consensus comparison, we omit a beat-or-miss claim rather than substitute an old estimate.

Management scorecard and five questions

Management deserves credit for providing the guidance comparison and explicitly identifying the net tariff EPS benefit. That makes it possible to distinguish per-share guidance from total profit guidance. But disclosure quality is not identical to earnings quality: a transparent unusual item is still unusual. Our scorecard separates what management explained from whether the underlying economics deserve a higher valuation.

Capital allocation also needs a price-sensitive assessment. A lower share count can reflect intelligent repurchases, yet cash spent above subsequent trading prices is not automatically a success. Conversely, a share price below a past repurchase price does not prove the decision was irrational at the time. We compare the cost with an explicit value range and the opportunity cost of store investment, debt service and liquidity. We avoid grading management solely on EPS growth that its own capital decisions helped produce.

1. How much tariff-related cash remains collectible, committed or subject to remittance, and where will each flow appear?
2. What portion of the updated gross-margin outlook is durable merchandise economics versus tariff timing and reinvestment?
3. What are mature-store and recent-cohort returns after occupancy, capex and working capital?
4. Which categories explain the negative comp, and did lower units per store affect availability?
5. At what valuation and cash-coverage thresholds does management choose repurchases over incremental expansion?

Risks and the strongest opposing cases

The strongest bull argument is that Academy remains profitable, is opening stores and has manageable borrowing debt. Stabilizing comps could turn today’s cautious multiple into an opportunity without requiring spectacular growth. The strongest bear argument is that the gross-margin and cash headlines flatter the durable improvement while expansion creates more fixed commitments. In that case, repurchases support EPS temporarily while the earning power deserves a lower multiple.

The loss distribution is asymmetric enough to matter. Our bear value is meaningfully below the current quote, while the central weighted value is approximately equal to it. That does not predict a fall; it says the price does not pay us much for accepting uncertainty under these assumptions. Hold is a valuation judgment, not a statement that the company is broken. If price falls while the source-backed operating case holds, the same company can become more attractive without any change in the latest EPS.

We do not assign false precision to tariff policy, consumer income or the timing of analyst revisions. These can change independently of management execution. The packet therefore pairs scenarios with observable thresholds and separates expected check dates from confirmed events. Readers should revisit the assumptions after material new disclosures rather than treat this quarter’s target as a permanent rating.

Catalysts and dated checks

The next useful catalyst is evidence, not merely a calendar entry. If the quarterly filing reconciles remaining tariff cash commitments and future collections, it can improve confidence in recurring cash without changing current EPS. If comps turn positive with disciplined inventory and store economics, the operating case improves. A target increase alone does neither. We will judge new information against the signposts rather than move the goalposts to defend the original call.

Source tables, methods and glossary

Current-quarter source: Q2 FY2026 Exhibit 99.1, the financial and reconciliation tables for the thirteen and twenty-six weeks ended August 1, 2026. Historical data: SEC companyfacts with accession, dates, taxonomy and exact duration retained for each observation. The accompanying historical filing snapshots support quarterly EPS, share counts and structural disclosures. Market history and dated target records are sourced to the saved Financial Modeling Prep responses; provider keys are excluded from public source links.

Method: dollar inputs are converted consistently from thousands to dollars and then millions for display. Growth is current divided by prior minus one. Margins divide the appropriate profit by sales. Borrowing net debt is carrying debt plus current maturities minus cash. Share-count attribution holds the old denominator first; reversing the order reallocates the interaction term but leaves final EPS unchanged. Ending-balance working-capital days are simple screening measures and are not average-balance turnover calculations.

Valuation multiples, required yields, scenario probabilities and the margin-of-safety threshold are our assumptions. They are not filed company facts. Scenario prices are rounded only after computing the underlying arithmetic, and weighted value uses the displayed rounded scenario prices. The central fair value is a judgment within the model band, not the mechanical average of three correlated methods. Prices can move after the frozen snapshot; that does not silently update the verdict in this archival quarterly packet.

Glossary: comparable sales measure the company’s eligible operating population, including defined online sales. GAAP is the accounting basis used in the filed financial statements. Adjusted earnings exclude the company’s specified items and may still contain unusual benefits. EPS divides common-stock earnings by a weighted share count. Enterprise value combines equity and the specified net-debt convention. FCF here always names its formula, because OCF less capex differs from Academy’s adjusted FCF definition.

The result is a stronger profit quarter with a more complicated cash story. Profit growth contributed more to the EPS bridge than the smaller share count, while tariff accounting and financing flows limit what can be extrapolated. At the updated price, the valuation cushion is thin. Those observations support a Hold near fifty dollars. Durable cash and a price that compensates for uncertainty would strengthen the investment case.

Historical normalization limits: FY2023 earnings retained a $15.9M credit-card litigation gain, and FY2024 retained a $15M settlement; the original FY2024 release also identifies a $7.1M sale-leaseback gain. Adjusted earnings are therefore not a universal recurring-profit measure. Some direct quarterly tables differ from annual-minus-nine-month calculations by $1,000–$2,000; the source dataset preserves both contexts. Such differences do not justify estimating EPS or changing signs to force agreement.