Renewal Rate Stops Falling, Member Growth Breaks Below 5%: Maintaining Outperform on a Narrower Margin

Results vs. Consensus

Costco reported fiscal second-quarter results for the 12 weeks ended February 15, 2026 after the close on March 5, with the call at 5:00 p.m. ET the same evening. The quarter beat on both lines by a margin consistent with the company's recent pattern: small, positive, and well inside the noise band of a business that does not guide.

Income statement, year over year

Every figure below is taken from the condensed consolidated statements of income filed as Exhibit 99.1 to the March 5 Form 8-K. Gross margin and SG&A are expressed as a percentage of net sales, which is the basis the company uses and the basis on which its own basis-point bridges are constructed.

Sequential comparison

Both FQ1 and FQ2 are 12-week quarters, so the sequential comparison is clean on period length. It is not clean on seasonality: FQ1 carries the holiday non-foods mix that lifts the reported gross margin rate, and FQ2 gives it back. The line to watch sequentially is operating margin, which improved despite the gross margin giving back 30bps.

Revenue

Total revenue of $69,597 million grew 9.2%, ahead of the roughly $69,286 million consensus by 0.4%. The composition is better than the 9.1% net-sales headline suggests once currency is removed, and worse once it is added back, which is the tension running through the entire top line this quarter. Foreign exchange added approximately 1.4% to comparable sales while gasoline price deflation subtracted approximately 0.7%, so reported total comps of +7.4% become +6.7% on an adjusted basis. Strip gasoline sales out entirely and adjust for currency and the figure is +7.4%, the same as the reported number by coincidence rather than construction.

The more useful read is the direction of travel. Adjusted comps accelerated to +6.7% from +6.4% in FQ1, and did so on traffic that held rather than on price. That distinction is the whole argument for the stock at this multiple.

Margins

Reported gross margin of 11.02% expanded 17bps, or 11bps excluding the gasoline effect. Management's own framing strips out the legal settlement as well and lands at roughly 5bps of underlying expansion. That is a small number, and it was earned in a quarter when the company was actively lowering prices on butter, coconut water, seaweed and olive oil, and absorbing tariff costs it chose not to fully pass through. Underlying margin expansion of 5bps while cutting shelf prices is a better outcome than 20bps of expansion earned by holding them.

SG&A deleveraged 13bps, or 8bps excluding gasoline. Six of those basis points are the general liability reserve. The operations component, which is the line that measures whether the technology investment is working, was 2bps worse reported but 2bps better excluding gasoline, and management stated that productivity gains "fully offset last year's wage investments and any impact of extended operating hours." That is the first quarter in our coverage where the productivity offset has been claimed as complete rather than partial.

EPS

Diluted EPS of $4.58 grew 13.9% against $4.02. Adjusting for the tax-rate benefit described above brings the comparable growth rate to roughly 12.4%, which is close to the operating-income growth rate of 12.5% and is the number we would carry into a model. Half-year EPS of $9.08 is up 12.7% against $8.06. The share count is doing almost nothing: buybacks of $419 million across 24 weeks reduced the diluted count by 0.1% year over year, consistent with management's stated policy of repurchasing only to offset equity-grant dilution.

Comparable Sales, Traffic and Ticket

The geographic comp table is where the currency distortion is most visible, and where the underlying picture is most useful. Traffic is unaffected by currency translation or gasoline pricing, so the traffic column is identical in both the reported and adjusted views. It is the cleanest signal in the release.

Adjusted figures exclude the impacts of changes in gasoline prices and foreign exchange; digitally-enabled adjusted excludes currency only.

United States

The U.S. is the only geography where the adjusted comp (+6.4%) exceeds the reported comp (+5.9%), because gasoline price deflation is a domestic drag that currency does not offset. Gasoline comps were negative mid-single digits on mid- to high-single-digit price deflation, partially offset by gallon growth, so the club is selling more fuel at lower prices. Adjusted U.S. ticket of +3.9% against traffic of +2.4% means basket growth is carrying the domestic comp.

U.S. traffic of +2.4% is the softest number in the release and sits meaningfully below the +3.1% worldwide figure. February compounds the question: U.S. traffic for the four weeks was +1.5% against +3.0% worldwide. Management attributed the February softness to Northeast weather that closed 55 warehouses for a full day, with a multi-day recovery, and declined to adjust the sales figures for it.

Assessment: we accept the weather explanation for February specifically, because a full-day closure of 55 clubs is a quantifiable event and the sales line did not show a corresponding gap. We do not extend it to the quarter. A U.S. traffic figure of +2.4% in a 12-week period is the metric we would want to see re-accelerate at FQ3, and it is the single number most likely to change our rating.

Canada

Canada delivered the largest gap between reported and adjusted comps in absolute terms outside of Other International: +10.1% reported against +7.6% adjusted, a 250bp currency and gasoline benefit. Traffic of +4.3% is nearly double the U.S. figure and is the second-strongest in the portfolio. Management noted that operating hours have been extended across all Canadian buildings specifically to absorb traffic growth, and that infill and business-center formats are the expansion vehicles. Canada added two warehouses in the quarter, both business centers, and the country count now stands at 114.

Assessment: Canada is running at high volume per building and management is managing throughput rather than chasing new markets. The adjusted ticket of +3.1% against traffic of +4.3% is the healthiest mix in the portfolio, because it is a genuinely traffic-led comp rather than a price-led one. The capacity question raised in Q&A is real but it is a good problem.

Other International

Other International posted the largest headline comp (+13.0%) and the largest currency distortion: adjusted, it falls to +7.1%, and adjusted ticket collapses to +2.4% from a reported +8.0%. Nearly six points of the reported ticket growth is translation. Traffic of +4.6% is the strongest in the portfolio and is the number that matters. Strongest local-currency results came from China, Taiwan and Korea. The segment added no warehouses in the quarter and stands at 176 buildings.

February exaggerated the effect further: Other International comps of +17.9% reported fall to +10.9% adjusted, and roughly 4.0 points of the reported figure is the Lunar and Chinese New Year falling 19 days later this year, which shifted the holiday into the February reporting period. That calendar shift added about 0.5% to total company February sales.

Assessment: the international headline comps are not what they appear and should not be extrapolated. Adjusted, Other International is growing at +7.1%, only modestly faster than the U.S. at +6.4%, on far stronger traffic. The traffic differential is the real signal and it supports the long-run unit-growth case. Investors anchoring on +13.0% will be disappointed when currency turns.

February sales (four weeks ended March 1, 2026)

Currency added approximately 1.7% to total company February sales, with Canada up approximately 5% and Other International up approximately 8% from translation alone. Gasoline price deflation cost approximately 85bps, with the average worldwide selling price per gallon down 7.5%. Cannibalization from the company's own new openings cost approximately 60bps. Adjusted February comps of +7.0% are the strongest monthly print in the period we have covered, and they came with the Lunar New Year shift adding roughly half a point.

Membership

Membership fee income is 52.0% of operating income this quarter, up from 51.5% a year ago. It is the annuity that justifies the multiple, and its composition changed materially this quarter. Fee income grew $162 million. Of the 13.6% growth, roughly 4.5 points came from the September 2024 U.S. and Canada fee increase, approximately 1.4 points from currency, and 7.5 points from the underlying base. At FQ1 the fee increase accounted for a little under half of the growth; this quarter it is about one third.

"The September 2024 U.S. and Canada membership fee increase accounted for about 1/3 of our membership income growth. Excluding the membership fee increase and FX, membership income grew 7.5% year-over-year."
— Gary Millerchip, CFO

Assessment: this is a better composition than the headline. The deferred price increase is a wasting asset that stops contributing after it fully laps, and it is now contributing less while total growth held at roughly the same dollar amount. Organic fee growth accelerated to 7.5% from 7.3%. The engine underneath the price increase is running slightly faster than it was a quarter ago, which is precisely what we needed to see before the tailwind disappears in the second half.

Margin Bridges

The company discloses both bridges in basis points against the prior-year quarter, on both a reported and an ex-gasoline basis. Each foots to its printed total.

Gross margin: 11.02%, +17bps reported / +11bps ex-gas

The distinction between core (-3bps) and core-on-core (+22bps) is the one that matters and it is widening. Core margin measured against total sales is diluted by the mix shift toward pharmacy and e-commerce, both growing faster than core merchandise, plus higher 2% executive reward accruals as executive penetration rises and a lower contribution from the co-brand credit card program. Measured on core sales alone, margin expanded 22bps with all three categories positive: non-foods, food and sundries, and fresh.

Assessment: +22bps is the fourth consecutive quarter of broad-based core-on-core expansion, following +36bps, +29bps and +30bps. It is also the smallest of the four. We read the deceleration as mix rather than pricing pressure: the two largest dilutive forces named by the CFO, pharmacy growth and rising executive-reward accruals, are both consequences of things going right. But four quarters at +36, +29, +30, +22 is a decelerating series, and we would not underwrite continued expansion at the +30bps pace in the second half against a tougher compare.

SG&A: 9.19%, +13bps worse reported / +8bps worse ex-gas

"These productivity improvements fully offset last year's wage investments and any impact of extended operating hours."
— Gary Millerchip, CFO

Assessment: excluding gasoline, operations turned positive at +2bps. That is the payoff line for the multi-year technology spend, and it is the first quarter in our coverage where the offset was described as complete rather than partial. Of the 8bps of ex-gas deleverage, 6bps is the liability reserve, which is a prior-year claims true-up rather than a run-rate cost. Strip it and the underlying SG&A line is roughly flat. The healthcare cost pressure flagged as a new variable at FQ1 was not mentioned at all this quarter, which we treat as an open question rather than a resolution.

Key Topics & Management Commentary

Overall Management Tone: management was steady and unusually specific on operating detail, and explicitly non-committal on the two questions the Street most wanted sized, retail media revenue and a special dividend. The posture on tariffs moved from the "operationalized" framing of the past two calls back toward acknowledged fluidity, with the CEO opening the call on the topic rather than waiting for it in Q&A. Where the quarter was weakest, on paid-member growth falling below 5%, the response reframed the long-run growth rate rather than defending the trend, which is candid but also lowers the bar the company will be measured against.

1. Tariffs: from operationalized back to fluid

For three consecutive calls the tariff narrative had been compressing toward routine. That reversed here. The CEO opened the prepared remarks on tariffs before anything else, and described a policy environment that had changed shape again since the last call.

"The future impact of tariffs remains extremely fluid as the recently eliminated IEEPA tariffs have now been replaced with new global tariffs for at least the next 150 days."
— Ron Vachris, CEO

The mitigation playbook is unchanged and now well rehearsed: shifting country of production where it makes sense, consolidating buying globally, leaning into Kirkland Signature where the company controls the supply chain, and sourcing more items domestically. The company also began cutting prices on categories where tariffs came down, naming certain textiles, bedding and cookware.

Assessment: the financial evidence says tariffs remain contained. The LIFO charge was $12 million, or 4bps, and non-foods inflation stayed low single digit. A 150-day replacement regime is short enough that it will be re-litigated inside the FY26 window, so this reverts from a resolved item to an active one. It does not yet threaten the thesis, but it is no longer background.

2. The IEEPA refund question, and who gets the money

The more consequential tariff disclosure concerned refunds. With the IEEPA tariffs eliminated, there is a question of what happens to duties already paid and, in many cases, already passed through to members in the form of higher shelf prices. Management pre-committed the answer.

"As we've done in the past, when legal challenges have recovered charges passed on in some form to our members, our commitment will be to find the best way to return this value to our members through lower prices and better values."
— Ron Vachris, CEO

Management was explicit that the process, the amounts, and the timing are all unknown, and noted that the layering of different tariffs at changing rates through the year makes it genuinely difficult to attribute a cost to an individual item sold.

Assessment: this is a pre-emptive answer to a question the sell-side would otherwise have asked at FQ3 or FQ4, and it should be taken at face value. Any IEEPA refund is a member-value pass-through, not an earnings event. Modeling a one-time gain here would be a mistake. The second-order effect is more interesting: a refund deployed into price is a competitive weapon funded by the government, and it arrives in a period when the company is already claiming it can expand margin while cutting prices.

3. Paid-member growth breaks below 5%

Paid households grew 4.8% to 82.1 million, the first sub-5% reading in the arc we have covered. Management gave three reasons, and they are not the same reason.

First, fewer openings in genuinely new markets. New-country openings such as Japan or China produce a spike in sign-ups that inflates the growth rate, and there have not been many recently. Second, the company is cycling strong sign-ups from a year ago. Third, and most important for the model, the CFO reset the structural expectation.

"If you look at our long-term growth rate, it really is in more of that 5% growth range in terms of new members."
— Gary Millerchip, CFO

Assessment: the first two reasons are cyclical and reversible; the third is a reframing. Note the interaction with the unit-growth plan: management's own explanation says membership growth accelerates when openings enter new markets, and the stated 30-plus-per-year plan is roughly half U.S. infill, which by that same logic drives frequency rather than sign-ups. A 5% membership growth rate paired with an infill-weighted opening schedule is internally consistent, but it means fee-income growth from FY27 depends on upgrades and price rather than on member count. Executive upgrades at +9.5% are carrying that weight today.

4. Renewal rate: the worldwide number finally stops falling

The renewal rate has been the live bear point since FQ4 FY25. The arc of sequential change now reads -30bps, -40bps, -10bps and, this quarter, -10bps in U.S./Canada and zero worldwide. The 89.7% worldwide rate is the first non-decline in the series.

"The slight decline in the U.S. and Canada renewal rate was due to the factors we have discussed in prior quarters and reflects new online members growing as a percentage of our total base and renewing at a slightly lower rate than warehouse sign-ups."
— Gary Millerchip, CFO

The mechanism has been consistently disclosed: digital sign-ups renew at a lower rate than warehouse sign-ups, and as they grow as a share of the base they pull the blended average down. Management said the decline would have been larger without the targeted digital-communication and retention programs, and continues to guide to a few more quarters of slight decline before the mix effect matures.

Assessment: this is the most important data point in the release for our rating. We upgraded at FQ4 FY25 partly on the expectation that the mitigation playbook would work, and it has now produced two consecutive quarters of moderation and a flat worldwide print. The caveat is that the U.S./Canada rate is 10bps from the 92% level we named as a downgrade trigger, so the margin of safety on this specific test is one quarter's slip.

5. Core-on-core expansion decelerates but keeps its breadth

Core-on-core margin expanded 22bps, with non-foods, food and sundries, and fresh all higher year over year. Management named commodity deflation, supply-chain efficiency and rising Kirkland Signature penetration as drivers, and was direct about why the reported core line looks different.

"On the core-on-core, specifically, side of it, as you heard us say, we were up 22 basis points. I wouldn't say there's one particular sort of driver of that."
— Gary Millerchip, CFO

The CFO added a specific mechanism worth carrying forward: when prices fall, Costco leads the market down, but because inventory turns so quickly the company captures a transitory financial benefit on the way through. That means a portion of the +22bps is a function of the deflationary environment in produce, eggs, butter and cheese rather than a structural gain.

Assessment: we would discount the deflation-timing component. The structural pieces, Kirkland penetration and supply-chain efficiency, are durable; the inventory-turn benefit on falling prices is not, and it reverses if commodity prices turn. This is the strongest argument for modeling core-on-core at +15 to +25bps rather than extending the +30bps run rate.

6. Gross margin expansion while cutting shelf prices

The CFO's own framing of the quarter's margin performance stripped out both the gasoline effect and the legal settlement, and landed on a much smaller number than the headline.

"we had a gain for a nonrecurring legal settlement in there for 6 basis points. So overall, we look at it as being up by 5 basis points in the quarter and being able to achieve that growth when we were also lowering prices for members and managing the impact of tariffs."
— Gary Millerchip, CFO

The price cuts were specific and disclosed at the item level: Kirkland Signature butter fell from $13.89 to $8.49 between the end of FQ1 and the end of FQ2, 12-count organic coconut water from $12.79 to $10.99, organic seaweed from $10.99 to $9.99, and two-liter Italian extra virgin olive oil from $29.99 to $24.99. Roughly 30 new Kirkland items launched in the quarter.

Assessment: the honest read is that underlying margin expansion was about 5bps, not 17bps. That is a small number in isolation and a good number in context, because the price file moved down materially at the same time. The company's stated position is to be "the first to lower prices and the last to raise them," and this quarter it demonstrably was. We would rather own the version of this business that converts margin into price and takes share than the version that reports 25bps of expansion by holding the file.

7. Ancillary businesses become the margin engine

Ancillary and other businesses contributed +19bps of the +17bps total gross margin change, meaning the rest of the P&L was net negative. The drivers were higher gasoline profitability and strong pharmacy growth, with pharmacy and food court both posting double-digit comps and optical and hearing in the high single digits.

Pharmacy in particular has become a real business rather than a traffic driver. Management cited AI tools deployed to improve in-stock positions, digital enhancements to speed pharmacy checkout, and growth ahead of total company sales. That last point is precisely why reported core margin looks weak: pharmacy carries a different margin structure and is growing faster than the core.

Assessment: the ancillary contribution is doing more work in the margin bridge than at any point in our coverage, and it is partly gasoline profitability, which is a function of falling crude prices and is not durable. Management flagged that instability in the Middle East could push fuel costs and shipping schedules the other way. If gasoline profitability normalizes, the +19bps ancillary contribution compresses and the reported gross margin loses its largest positive line.

8. Digital accelerates and personalization gets sized for the first time

Digitally-enabled comps accelerated to +22.6% (+21.7% excluding currency) from +20.5% in FQ1. E-commerce site traffic rose 32% and average order value 15%. Top categories were gold and jewelry, toys, small electrics, housewares and pharmacy, and the same-day delivery channel through Instacart, Uber Eats and DoorDash grew faster than digital overall.

The new disclosure is a number attached to the personalization work.

"In Q2, our personalized product recommendation carousels drove over $470 million of e-commerce sales, and our newly modernized product display pages are driving incremental sales on our dot-com site as well as increased traffic to our same-day sites."
— Gary Millerchip, CFO

Assessment: this is the first time the personalization investment has been given a dollar figure, and $470 million in a 12-week quarter is meaningful even before asking how much of it is incremental rather than redirected. The acceleration from +20.5% to +22.6% on a metric that only began reporting under this definition at FQ1 is the cleanest evidence in the release that the digital roadmap is compounding. We would like the incrementality question answered before capitalizing it into the thesis.

9. Unit growth: 924 today, 942 by year end, and a back-half that has to deliver

The company opened three warehouses in the quarter (one U.S. net new, one U.S. relocation, and two Canadian business centers, netting to three) and stands at 924 worldwide: 634 in the U.S. and Puerto Rico, 114 in Canada, and 176 across the rest of the world. The full-year plan was reaffirmed.

"We currently expect to have 28 net new openings in fiscal year '26 and are targeting 30-plus new openings per year in the coming years."
— Ron Vachris, CEO

The supplemental deck makes the cadence explicit: seven openings in FQ1, three in FQ2, and 18 estimated across the remainder of the year to reach 942. Management also put a more specific frame on the long-run target than at FQ1, describing a five- to ten-year horizon with "just over half" in the U.S. and "just under half" across Mexico, Canada, Asia, Europe, Australia and New Zealand. Creative real-estate structures, including parking decks and residential development above warehouses, are the mechanism for entering dense urban markets where 25-acre sites do not exist.

Assessment: 18 of 28 openings landing in the back half is a heavy concentration, and the same schedule that delayed two Spanish buildings out of FY26 at FQ1 is the reason to watch it. Capex tells the same story: $1.29 billion in the quarter and $2.82 billion across 24 weeks against a roughly $6.5 billion full-year plan, implying about $3.7 billion in the second half, up roughly 31% from the first. The plan is credible and the pipeline commentary is more specific than a quarter ago; the risk is schedule, not intent.

10. Capital allocation: the cash keeps building and the special dividend keeps not happening

Cash and short-term investments reached $18.24 billion at quarter end, against $5.69 billion of long-term debt. Operating cash flow of $7.68 billion across 24 weeks grew 27.9%, and free cash flow after $2.82 billion of capital expenditure was $4.87 billion, up 35.0%. Asked directly whether the balance sheet points to another special dividend, management neither committed nor dismissed it, and introduced a constraint we had not heard framed this way before.

"What I'd say on special dividend is while our cash balances are back to the levels that they were pre the last special dividend, I think, it's important to remember that to achieve a similar yield to last time when our stock was at $660, the cash would need to be greater."
— Gary Millerchip, CFO

The stated priority order was unchanged: invest in the business first, grow the regular dividend, repurchase enough stock to offset equity-grant dilution, and treat a special dividend as the most efficient vehicle for genuine excess because it preserves flexibility. Dividends paid across the 24 weeks were $1,154 million against $515 million a year ago, which works out to $2.60 per share versus $1.16, a payment-count difference across the comparison window rather than a doubling of the run rate. Buybacks were $419 million against $412 million.

Assessment: the yield-math point is a real constraint and it is the first time management has explained the special dividend as a function of the share price rather than the cash balance. At a materially higher stock price the same dollar payout is a much smaller gesture, so the bar rises as the stock rises. We would not model a special dividend into FY26. The more useful observation is that free cash flow grew 35% on 24 weeks while capital expenditure grew 17%, so the cash build is accelerating even as the investment plan steps up.

11. AI: positioning for agentic commerce without spending on it

Management's AI commentary split cleanly in two. Internally, the framing was deliberately unglamorous: pharmacy in-stock tools, employee pre-scan technology, automated pay stations piloting at roughly eight seconds per transaction, and the personalization work already discussed. Externally, the framing was new.

"We're working closely with the leading AI companies to ensure our values will be visible to existing and potential future Costco members as they engage with these tools."
— Ron Vachris, CEO

Asked directly how the company wants to show up when a consumer queries a large language model, the answer was that Costco wants to surface everywhere, on the theory that a price-and-quality-ranked comparison is a format that favors it.

Assessment: this is the right instinct and the argument is genuinely sound. A shopping interface that ranks on price and quality rather than on advertising spend is structurally favorable to the lowest-markup operator in retail. But the membership wall cuts the other way, because an agent comparing prices across open web retailers may not surface a price that requires a paid annual membership to access. Nothing here is thesis-relevant yet. We note it as a genuine two-sided risk that neither management nor the Street has framed.

Outlook & Forward Markers

Costco does not provide formal revenue or earnings guidance. The call produced the following actionable forward markers.

Implied second-half ramp. Half-year EPS is $9.08. FY25 delivered $18.21 with $8.06 in the first half, so the second half carried $10.15. Growing that at the first half's 12.7% pace gives roughly $11.44 and an FY26 total near $20.52. Growing it at 10% gives $20.25. The second half also loses most of the deferred fee tailwind and absorbs the March 2026 wage increment, so the low end of that band is the prudent anchor.

Guidance style. Unchanged. The company gives capital expenditure, opening counts and monthly sales, and nothing else. That discipline is a feature, but it means the FY26 EPS number is entirely an analyst construction and the Street's range should be treated as wider than it is quoted.

Analyst Q&A Highlights

Weather distortion and whether the consumer changed

The opening question tried to separate genuine demand from calendar and weather noise across January and February, with a competitor having already flagged a February weather headwind. Management gave a specific, quantified answer on the traffic line and declined to adjust the sales line, then broadened to a read on the consumer that emphasized consistency over any inflection.

Q: "There's been a lot of noise in January and February on the weather, whether there was a net benefit to January."
— Christopher Horvers, JPMorgan

A: "The thing that we think may have caused that to look a little bit lighter was because of the weather we had in the Northeast, in particular, we have 55 warehouses that were closed for a full day and then took a couple of days for the local communities to get back up to sort of speed."
— Gary Millerchip, CFO

Assessment: the answer was better than a deflection because it named the mechanism and the count. Management explicitly declined to adjust the sales figures for it, which is the conservative choice and consistent with prior practice. The broader consumer read, that results have been "very consistent in that 6% to 7%" once calendar shifts and short-term blips are stripped, is the same framing used at FQ1, so there is no tone shift to read here.

Sizing the productivity savings and where they get reinvested

A recurring theme across recent calls has been whether the technology investment produces a margin benefit or a reinvestment budget. The question asked management to size the savings from pre-scan, pay-ahead and automated checkout, and to say whether returns on the reinvestment are diminishing. Management declined the sizing and answered the diminishing-returns half directly.

Q: "Did you size the potential savings from things like prepaying your card or line breaking from your associates?"
— Michael Lasser, UBS

A: "As far as investing in the business, seeing the same values in that, no, we feel that we still get the same return from our members as we continue to invest in the business out there."
— Ron Vachris, CEO

Assessment: no number was given and none was likely to be. The useful content sits in the SG&A bridge rather than the answer: operations turned +2bps favorable excluding gasoline, and management stated the productivity gains fully offset the prior-year wage investment and extended hours. That is the sizing, expressed in the only place it can be verified.

Core-on-core trajectory into a tougher second-half compare

The most substantive exchange of the call. The question flagged that the year-ago comparison gets harder in the back half and asked how to think about the core-on-core line from here. Management redirected to the total gross margin, stripped out both the gasoline effect and the legal settlement, and landed on a materially smaller underlying number than the headline.

Q: "I was hoping that you could expand on core-on-core margins in the quarter and then maybe how we should be thinking about the back half? The compare seems a little bit tougher there, but just thoughts on how we should be thinking about that would be great."
— Edward Kelly, Wells Fargo

A: "I share all that context because I think from our perspective, when you think about looking forward, the rate is going to fluctuate and the different elements are going to fluctuate quarter-to-quarter, and we tend to not get too fixated on one individual element of the margin."
— Gary Millerchip, CFO

Assessment: the answer explicitly declined to guide the back-half trajectory, which is consistent with a company that does not guide. What it did do, unprompted, was volunteer that part of the core-on-core gain comes from quick inventory turns during a deflationary stretch. That is management telling the model to discount a portion of the +22bps. We take the hint and lower our core-on-core assumption accordingly.

Paid-member growth falling below 5%

The question named the number directly and asked what was driving it. The response is the closest thing to a guidance revision the call produced, because it reset the structural expectation rather than treating the quarter as an anomaly.

Q: "So just going back to membership growth, so it's sub-5% this quarter. So if you could maybe walk through some of the dynamics at play."
— Rupesh Parikh, Oppenheimer

A: "If you look at our long-term growth rate, it really is in more of that 5% growth range in terms of new members."
— Gary Millerchip, CFO

Assessment: this is the exchange with the largest model consequence. Two of the three reasons given are cyclical, but the third recasts the historical 6%-plus prints as flattered by new-market openings rather than as a baseline. That is probably accurate and it is certainly candid, and it means fee-income growth beyond the price-increase lap has to come from upgrades and pricing rather than member count. Executive members at +9.5% are doing that work today, which is why the fee line held while the member line did not.

Whether the renewal rate has found its floor

A direct question on whether this is the bottom, and on the mechanics of the auto-renewal push visible in warehouses. Management would not call a bottom but confirmed that the trajectory has improved relative to what the mix math alone would have produced.

Q: "The U.S. down about 10 basis points, worldwide flat. So is this the real bottom here?"
— David Bellinger, Mizuho

A: "And if we'd have just played out the impact we would have expected without any of that activity, it would have been a higher decline just with the math of the number of digital members that we're feeding into the overall renewal rate calculation."
— Gary Millerchip, CFO

Assessment: the claim is unfalsifiable as stated, because the counterfactual is unobservable. What is observable is that the sequential decline went from -40bps to -10bps to -10bps, and that the worldwide rate is flat. Two quarters of moderation and one flat print is a trend, and it is the trend we underwrote at the FQ4 FY25 upgrade. Management still guides to a few more quarters of slight decline, so the flat worldwide print should not be read as the turn.

Decomposing inflation out of the ticket

The question tried to pin a basis-point figure on inflation inside the U.S. ticket. Management gave a directional answer, declined the arithmetic, and made a distinction that is more useful than the number would have been.

Q: "I wanted to follow up on inflation. You mentioned how, I believe, it was a little bit less this quarter than the prior quarter. And I'm just curious how much less."
— Gregory Melich, Evercore ISI

A: "But directionally to your point, the inflation as in the actual price increases would only have been a fraction of the total, and the mix changes and the increasing units would have been a meaningful part of the growth as well."
— Gary Millerchip, CFO

Assessment: the substantive content is that price is a minority of ticket growth and that units and mix carry the rest. If that holds, adjusted ticket of +3.5% is mostly real basket expansion rather than pass-through, which is a materially better quality of comp than the same number driven by price. Inflation trending toward low single digits from low-to-mid also removes a support from reported comps in the back half, so the traffic line has more to do.

Excess cash and the special dividend

The balance sheet invited the question and it came late in the call. The answer restated the capital-allocation hierarchy and then added a constraint that has not previously been articulated on these calls.

Q: "I know it's only been about 2 years or so, but the last time you had this much cash on the balance sheet, you did pay out a special dividend. So is that something we could see in the next few quarters?"
— Scot Ciccarelli, Truist Securities

A: "And so we'll continue to review the question of special dividend with our Board, but there are no plans that we could share at this time around a plan for special dividend."
— Gary Millerchip, CFO

Assessment: the refusal is routine but the reasoning is new and it matters for anyone modeling a capital return. Framing the decision against the yield the payout would deliver at the current share price, rather than against the absolute cash balance, means the threshold rises with the stock. Since the shares have roughly reached $1,000 against $660 at the last special, the same dollar payout is worth roughly a third less in yield terms. The practical conclusion is that the cash keeps compounding on the balance sheet.

What They're NOT Saying

1. No explanation for the 97bp decline in the effective tax rate. The rate was stated, 25.2% against 26.2%, and never revisited. No analyst asked. It contributed roughly $0.06 to EPS, twice the size of the consensus beat, and without knowing whether it is discrete or structural the FY26 modeling assumption is a guess.
2. Healthcare costs disappeared from the discussion entirely. At FQ1 management flagged the first quarter in which healthcare grew faster than sales and called it a new variable with mitigation in progress. It was not mentioned once this quarter, and the SG&A bridge attributes the "Other" line to general liability reserves instead. Either it resolved, or it is now inside the Central line and unlabeled.
3. The March 2026 wage increment went unmentioned. The FQ1 call flagged a top-of-scale increase arriving in March 2026, inside FQ3. The FQ2 call, held on March 5, did not reference it, quantify it, or frame the offset. It is the single largest known cost step in the back half.
4. Retail media revenue remains unsized, now for the fourth consecutive quarter. Two separate analysts asked. Management confirmed roughly 1,000 participating suppliers and double-digit growth, then declined to give a base, explicitly reasoning that the value will show up in top-line growth rather than in the margin profile because it will be reinvested.
5. Affirm was not mentioned at all. The big-ticket buy-now-pay-later partnership has now gone four quarters without being sized, and this quarter it was not raised by management or by any analyst.
6. No sizing of the IEEPA refund exposure. Management pre-committed the use of any refund but gave no indication of the magnitude of duties paid, the share passed through, or the recovery timeline. Given the commitment is to return it to members, the disclosure gap matters less for EPS than for competitive positioning.
7. The incrementality of the $470 million personalization figure is unaddressed. Sales driven by recommendation carousels are not necessarily sales that would otherwise have been lost. Without an incrementality frame the number is a gross attribution, not a contribution.
8. No U.S. traffic trend commentary beyond February. U.S. traffic of +2.4% for the quarter was disclosed in the deck and never discussed. The weather explanation was offered for February alone. The quarterly figure, which is the one that matters, was left without a frame.
9. No update on the two delayed Spanish openings. Those buildings moved from FY26 to FY27 at the FQ1 call and were the reason the opening plan fell from 35 to 28. With 18 of 28 openings still to land in the back half, their status is a reasonable thing to have asked about, and nobody did.
10. No comment on the next U.S. membership fee increase. The September 2024 increase followed a seven-year gap. With the deferred tailwind lapping and member growth resetting to 5%, fee pricing is the most obvious lever available, and it went entirely unaddressed for the second consecutive quarter.

Market Reaction

• Pre-print setup: the stock closed at $982.57 on March 5, up 13.9% year to date against the S&P 500's -0.2%, up 0.5% over the trailing 30 days, and down 6.2% over the trailing twelve months. The 52-week closing range entering the print was $850.00 to $1,056.85, so the shares were roughly 7% below their 52-week closing high and 16% above the low.
• Reaction session (March 6): the shares opened at $967.01, 1.6% below the prior close, traded as low as $960.46 (-2.3%) and as high as $1,002.50 (+2.0%), and closed at $998.10, up 1.6% or $15.53. The close sat within $4.40 of the session high.
• Volume: 2.7 million shares against a 30-day average of 2.2 million, or roughly 1.2x. Elevated but not extreme.
• Relative: the S&P 500 fell 1.3% on the same session, so the shares outperformed the index by roughly 290bps. Among peers, KR rose 3.6%, WMT and TGT each rose 0.4%, and BJ fell 1.7%.

The shape of the session is more informative than the close. The stock gapped down with a weak tape, extended to a 2.3% loss intraday, and then recovered the entire move and more, finishing near its high on a day the index fell. That is a pattern consistent with an initial read that focused on the two soft headline numbers, paid-member growth of 4.8% and another 10bp slip in the U.S./Canada renewal rate, followed by a session in which the detail was digested: the worldwide renewal rate was flat, worldwide traffic held at +3.1%, adjusted comps accelerated, and the underlying fee-income growth rate improved as the price-increase contribution shrank.

The peer tape supports that reading rather than a sector-wide move. BJ, the closest club comparison, fell 1.7% on the same session while Costco rose, so this was not a warehouse-club bid. Outperforming a 1.3% index decline by nearly three points, on a beat of only $0.03, is the market paying for quarter composition rather than for the beat itself.

Street Perspective

Debate: is the membership annuity decelerating or just changing shape?

Bull view: fee income grew 13.6% while the September 2024 price increase fell from roughly half of the growth to roughly one third, meaning organic growth accelerated to 7.5% and did the heavier lifting. Executive members grew 9.5% and now account for 49.2% of the paid base and 75.8% of sales. The annuity is broadening, not fading, and the upgrade flywheel is a renewable driver in a way that a one-time price increase is not.

Bear view: paid-member count grew 4.8%, the first sub-5% print in the arc, and management responded by reframing the long-run rate downward rather than treating it as a lap effect. Once the price increase fully laps in the second half, fee income growth converges toward the sum of member growth and upgrade mix, which is high single digits at best. The 13.6% headline is the last quarter that looks like this.

Our take: the bear has the arithmetic and the bull has the direction. Fee income will decelerate to high single digits in the back half; that is not in dispute and it is already in our numbers. What matters is whether the deceleration is to 7% or to 5%, and the answer turns on upgrade penetration, which at 49.2% of the base still has room, and on whether the company takes fee price again. Management has now gone two quarters without addressing the next increase, which we read as a live option rather than a closed one.

Debate: does the margin story survive the loss of its two temporary supports?

Bull view: core-on-core expanded 22bps with all three categories positive, the fourth consecutive quarter of broad-based expansion, and it was achieved while cutting shelf prices on named Kirkland items. Operations SG&A turned favorable excluding gasoline for the first time, meaning the technology spend now fully covers the wage step-ups. Both are structural.

Bear view: two of the largest positive contributors are transitory. Ancillary businesses added 19bps on gasoline profitability that is a function of falling crude, and management itself flagged Middle East instability as a risk to fuel costs. Part of the core-on-core gain is an inventory-turn benefit that exists only while commodity prices are falling. Strip both and underlying expansion is the roughly 5bps management named, which does not support a 49x multiple.

Our take: the bear identifies the right two items and overstates the conclusion. Underlying expansion of 5bps earned while lowering the price file is a different and better outcome than 5bps earned while holding it, because the price investment buys traffic and share. But we agree the reported gross margin line is flattered and we would not extend the +22bps core-on-core into the back half against a harder compare. Our model carries +15 to +25bps.

Debate: is 49x forward defensible when the growth algorithm is decelerating?

Bull view: the business compounds EPS in the low-to-mid teens with a contractual fee annuity, a net cash balance sheet, free cash flow growing 35%, and a unit-growth runway management has now framed over five to ten years at 30-plus openings annually. Quality of that order rarely gets cheaper, and the stock is 6.2% lower than a year ago while earnings grew, so the multiple has already compressed.

Bear view: at $998.10 the shares trade at 51.9x trailing and roughly 49x forward earnings on a business whose comp is 6.7% adjusted, whose member growth just reset to 5%, and whose EPS growth this quarter was 12.4% once the tax benefit is removed. That is a mid-teens multiple business priced at a high-40s multiple, and the second half loses the fee tailwind while absorbing a wage increment.

Our take: the bear is right about the multiple and wrong about the alternative. We do not defend 49x as cheap and we are not underwriting multiple expansion. The Outperform rests on the earnings stream compounding through a period in which the two things that could break it, renewal rates and traffic, both stabilized this quarter. If either resumes deteriorating at FQ3 the valuation offers no cushion at all, which is exactly why we are describing the margin as narrower rather than reaffirming it.

Model Update Needed

Valuation impact: the FY26 EPS revision is roughly a 3% cut at the midpoint. Held against an unchanged forward multiple that is worth approximately $30 per share of fair value, which is inside a single session's range for this stock and does not on its own change the rating. The offsetting consideration is qualitative: the renewal-rate stabilization and the flat worldwide traffic reading both reduce the probability of the downside scenarios that would compress the multiple, and those matter more at 49x than a 3% earnings revision does.

Thesis Scorecard Post-Earnings

Overall: the standing thesis is unchanged in substance and better balanced in evidence than it was a quarter ago. The two bear points we carried as the live risks to the FQ4 FY25 upgrade, the renewal-rate slip and traffic moderation, both improved and both moved from emerging to contained. Against that, paid-member growth moved the other way and management reset its own long-run framing, the tariff environment reverted from resolved to fluid, and our FY26 earnings estimate comes down 3%. The bull case is intact; the margin of safety is smaller.

Action: Maintaining Outperform. At the FQ1 report we named four conditions that would return us to Hold: sustained worldwide traffic below +3%, the deferred fee landing materially below management's framing, a tariff-driven non-foods inflation acceleration large enough to swamp LIFO, and a U.S./Canada renewal rate below 92%. None tripped. Worldwide traffic held at +3.1%, fee income grew 13.6% with better composition, LIFO cost 4bps, and the renewal rate is 92.1%. We hold the rating on that basis and note plainly that two of those four tests are now within 10bps of failing. The second half carries the wage increment, the lap of the fee tailwind, 18 of 28 planned openings, and a tariff regime with a 150-day clock on it. We are underwriting a 12-month total return above the S&P 500 from $998.10, with FQ3 framed as a decision point rather than a formality. Holders should treat a U.S./Canada renewal rate below 92% or worldwide traffic below +3% as sufficient on its own to move us to Hold.