Nine Straight Quarters of Record Margins and a Floor Sold Out Through Mid-2026, but a Tax-Muddied Guide Tests a Stock Priced for Perfection — Initiating at Hold

Results vs. Consensus

FQ4 FY2025 Scorecard

Year-Over-Year Comparisons

Quarter-Over-Quarter Comparisons

Quality of Beat

Revenue: The $2.44B print landed $44M above the guide midpoint and modestly above every published consensus figure. The composition matters more than the magnitude: mass-capacity revenue crossed the $2B mark for the first time (+40% YoY), driven by 137 exabytes of nearline shipments into cloud and edge data centers, while the legacy and consumer lines (roughly 18% of revenue combined) merely held serve with modest sequential gains. Management said explicitly that demand exceeded what the company could supply and that some production volume was diverted to customer qualification units. In other words, the revenue line was supply-limited and quality-screened, not demand-limited: the cleanest kind of beat.

Margins: The 170bp sequential expansion to 37.9% non-GAAP gross margin came from three compounding sources: richer mix (nearline at 91% of mass-capacity exabytes, with the 24–28TB PMR platform setting a company record for quarterly sales and volume of any nearline product), the like-for-like pricing escalators embedded in renewed build-to-order contracts, and cost-per-terabyte declines as HAMR volume grows. Nothing in the reconciliation suggests accounting flattery: the delta between GAAP (37.4%) and non-GAAP (37.9%) gross margin is a routine $12M of share-based compensation, restructuring, and purchase-order items. Operating margin at 26.2% cleared the top of the "mid-20s" guide on operating expense discipline ($286M, in line).

EPS: The $2.59 is at the top of the guided range and operationally pure: the tax provision was a de-minimis $4M this quarter (the last before the global minimum tax regime begins), interest expense continued to decline with the debt paydown, and share count was stable. The optical wrinkle is GAAP EPS of $2.24 versus $2.39 a year ago, which screens as a decline; the prior-year figure was inflated by a $313M one-time gain on the SoC operations divestiture. Underlying pre-tax operating income rose 81% YoY. This distinction matters for screens and headlines but not for the investment case.

Segment Performance

Revenue by End Market — FQ4 FY2025

Key Volume & Mix KPIs

Mass Capacity / Nearline — The Entire Story

Mass capacity is now 82% of revenue and effectively all of the margin story. Nearline shipments of 137 exabytes grew 14% sequentially and 52% year-over-year, with cloud service providers absorbing the vast majority and enterprise OEM demand showing a modest sequential improvement that management expects to remain stable over coming quarters. The 24–28TB PMR platform set a company record for quarterly sales and volume of any nearline product in its history, while HAMR-based Mozaic 3+ units ramped in parallel. Implied mass-capacity revenue per terabyte was firm-to-rising sequentially, an outcome that would have been unthinkable in any prior HDD cycle where volume growth of this magnitude invariably bought price erosion.

Assessment: The combination of +52% YoY volume and stable-to-rising per-terabyte economics is the empirical proof of the build-to-order regime. Watch whether nearline exabytes can keep stepping up while qualification units consume production; management flagged that supply, not demand, is the governor through at least the December quarter.

Legacy & Other — Residual and Shrinking

Legacy HDD (client compute, consumer) at $270M grew 6% sequentially on seasonal factors but remains in structural decline, and the other-products line ($163M of SSDs and systems) was up 3%. Together they are roughly 18% of revenue, down from about a third two years ago. Management said seasonality in the overall model is "starting to really diminish" as these lines shrink relative to data center demand, with the March quarter now the only quarter with a visibly softer seasonal skew.

Assessment: The residual segments are a diminishing drag on mix and an accounting footnote to the thesis. The forthcoming re-segmentation (below) will formalize this: from the September quarter Seagate will report around a data center bucket (~75% of FY25 revenue) and an edge IoT bucket, which is a better map of what the company has become.

Key Topics & Management Commentary

Overall Management Tone: Confident and unhurried, with none of the defensiveness one might expect given the after-hours sell-off developing as the call progressed. Management repeatedly redirected quarter-timing questions toward the multi-quarter supply plan, declined to flatter the guide with hypotheticals, and twice told analysts their margin models were too conservative rather than managing expectations downward. The posture was that of a company that believes its Investor Day framework is being validated ahead of schedule and sees no need to re-argue it. Where management was least specific was HAMR's revenue contribution, which remains undisclosed.

1. Build-to-Order: The Floor Is Sold Out Through Mid-Calendar 2026

"The visibility afforded by our build-to-order strategy indicates our nearline exabyte production capacity is largely spoken for through the middle of next calendar year with visibility building into the second half." — Dave Mosley, CEO

This is the most consequential sentence in the prepared remarks. In April, management described visibility "into the first half of calendar 2026"; the commitment window has now extended to "largely spoken for through the middle" of 2026 with second-half visibility building. For a business that historically ran on 13-week purchase orders and suffered brutal inventory-driven whipsaws, a four-quarter forward order book is a different operating model, not an incremental improvement. Management also framed the contracts as mutually beneficial: customers get assured supply against data center buildouts that are expected to more than double installed gigawatt capacity by 2029.

Assessment: The BTO book is the load-bearing wall of the margin thesis. As long as capacity is oversubscribed, pricing holds, mix optimizes toward the highest-capacity drives, and the company controls its own factory loading. The test we will apply each quarter: does the visibility window keep extending roughly one quarter per quarter? It did this quarter.

2. HAMR / Mozaic 3+: Three CSPs Qualified, More Coming in September

"We are tracking to plan with shipments expanding to additional CSPs in the September quarter. Across the board, qualifications are progressing exceedingly well, and we continue to expect the key global CSP customers to be qualified by mid-calendar 2026." — Dave Mosley, CEO

The HAMR scorecard against April's commitments reads clean: the second major CSP qualification that was "nearing completion" has closed, and the CFO confirmed three major cloud customers are now qualified on Mozaic 3+ with additional qualifications "proceeding extremely well." Customer breadth was the honest bear question entering this print, since the ramp to date had been concentrated in one hyperscaler; management confirmed on the call that customers beyond the lead are "starting to ramp as well" with strong pull. The mid-calendar-2026 target for having the key global CSPs qualified remains the timeline that matters.

Assessment: Qualification breadth was the single biggest execution risk on HAMR, and it measurably narrowed this quarter. The remaining question is not whether HAMR works (it is shipping in volume, accretive to gross margin per the CFO) but how quickly the qualified base converts to volume. September's "expanding shipments" language gives us a near-term checkpoint.

3. Mozaic 4: The 4TB-per-Disk Qualification Has Started

"As planned, we recently started qualification with a global CSP on the 4 terabyte per disk platform and expect to begin volume ramp in the first half of calendar 2026. This time line is consistent with our target for exabyte shipment crossover on HAMR-based nearline drives in the second half of calendar '26." — Dave Mosley, CEO

The 4TB/disk platform (up to 44TB drives) is fiscal 2026's stated top priority and the product that management believes removes the old capacity ceiling entirely. Asked about the previously discussed ~160-exabyte practical limit on mass-capacity output, Mosley said continued Mozaic 3+ ramp allows growth and "4+ allows significant growth above that… it's not a wall so much anymore." The company is deliberately consuming operational efficiency to accelerate product transitions rather than adding gross drive-count capacity, which is the areal-density playbook: more exabytes per unit, not more units.

Assessment: The 4TB/disk qualification start, on schedule, is the first checkpoint in a chain (qualification → 1H26 volume ramp → 2H26 HAMR exabyte crossover) that underpins both the revenue CAGR and the incremental-margin math in the FY2028 model. This is the milestone track we will grade hardest in coming quarters, because a slip here cascades into everything.

4. Record Gross Margin and the Path to 40% "Fairly Soon"

"Our path to achieve the milestone or the first milestone that we discussed at our Investor Day just a few weeks ago is intact. Now we are going exactly in that direction. I think we can be there fairly soon." — Gianluca Romano, CFO

The exchange behind this quote was the call's sharpest moment: the opening analyst question calculated that the September guide implied only ~20bp of sequential gross margin expansion against an eight-quarter average above 200bp. The CFO's response was unusually direct, calling the estimate "significantly lower than what is implied in the guidance." Working the arithmetic ourselves: a $2.5B revenue midpoint with ~$290M of operating expenses and operating margin guided into the mid-to-high 20s implies gross margin around 38.5–39.5%, which would be a tenth consecutive record and meaningful progress toward the 40% milestone attached to the $2.6B quarterly revenue level in the Investor Day framework.

Assessment: Management is guiding to continued margin expansion and telling the Street its models are too low, five days after the stock hit an all-time high. That is not the behavior of a team protecting a peak. The 40% gross margin milestone "fairly soon" is now a stated near-term expectation, and we treat it as a hard commitment for the scorecard.

5. Capital Returns: Buyback Resumes, Leverage Target Reached

"We don't have excess cash right now. I think we can maybe still increase a little bit our cash position, but the vast majority of our free cash flow, of course, will go back to our shareholders through the dividend and through the share buyback." — Gianluca Romano, CFO

The deleveraging-first sequencing communicated in April has completed on schedule: gross debt ended the year at $5.0B (down $684M in FY25, including retiring $505M in the June quarter via a smaller $400M note issuance and cash), net leverage reached 1.8x from 2.1x, and share repurchases resume this quarter against the $5B authorization announced at the May Investor Day. The stated first use of the buyback is offsetting dilution from the 2028 convertible notes, which begin diluting non-GAAP share count when the stock trades above roughly $108; the guided September share count of 221M (from 215M) reflects exactly that. FY25 capital returns were $600M of dividends, roughly 75% of the $818M free cash flow, consistent with the greater-than-75%-of-FCF framework.

Assessment: The balance sheet has moved from constraint to tailwind. With FCF guided to expand in the second half of calendar 2025 and capex staying inside the 4–6% of revenue band, the return engine (dividend plus anti-dilutive buyback, then discretionary repurchases) becomes a visible part of the total-return math for the first time since the FY23 downturn.

6. The Guide's Optics Problem: Taxes, Share Count, and 14 Weeks

The September quarter guide of $2.30 ± $0.20 in non-GAAP EPS screens as a $0.29 sequential decline from $2.59, and the after-hours tape treated it as a growth warning. The composition says otherwise. Three non-operational factors reset the EPS base this quarter: the Pillar Two global minimum tax takes the non-GAAP tax rate from near zero to approximately 16% (worth roughly $0.40 per share at guided profitability), the convertible-and-equity-driven share count rises from 215M to 221M (roughly $0.06), and the quarter contains 14 weeks, which lifts operating expenses to ~$290M while build-to-order supply, not the extra week, determines revenue. Adjusting only for tax and share count, the $2.30 midpoint is the equivalent of roughly $2.75–2.80 on the June quarter's basis: sequential growth, not decline.

Assessment: The EPS reset is real money (the tax is a genuine, permanent cash cost) but it is a one-time rebasing, not a demand signal. The correct sequential comparison is revenue (+2% guided, against a supply-constrained base), gross margin (guided up), and operating margin (guided into the mid-to-high 20s). All three point the same direction. The Street's initial read conflated a tax regime change with a growth inflection, which is precisely the kind of confusion that creates the next few months' debate.

7. AI Demand: Video Properties Today, Edge and Sovereign Data Tomorrow

"We're operating in a strong demand environment, driven in part by advancements in Gen AI and the march towards all these agentic models. These breakthroughs have solidified data as one of the world's most critical resources." — Dave Mosley, CEO

Management's AI framing is deliberately less breathless than the accelerator-supply-chain narrative and more specific about mechanisms: video-heavy cloud properties storing and serving diverse user-generated content at global scale; unstructured data pools retained for training and inference checkpoints rather than deleted; edge applications (factories, hospitals, safety) that now "snapshot" continuously because storage is cheap relative to the option value of the data; and sovereign-data requirements that will force localized storage buildouts, given that roughly half the world's data centers sit in just four countries. One concrete proof point: a top-tier CSP built a tiered storage solution for a major social media platform using hard drives both for mass storage and as a caching layer, and a leading enterprise IT vendor will introduce HAMR-based tiered storage systems for AI applications later this year.

Assessment: The AI story here is a second-derivative one (AI increases the value of retaining data, which increases exabyte demand) rather than a direct compute play, and management resisted the temptation to raise its mid-20s exabyte CAGR on the back of it. That restraint cuts both ways: it keeps expectations grounded, but it also means any upside from AI-driven retention is not yet in the model. The wildcard management itself flagged: viral, data-dependent applications at the edge.

8. Pricing Discipline: Like-for-Like Escalators Continue

"We are not changing our pricing strategy that we have started more than two years ago. So our like-for-like pricing will continue to slightly increase every time we negotiate a new build to order." — Gianluca Romano, CFO

Asked directly whether the HAMR cost advantage would leak to customers through lower per-exabyte pricing as long-term agreements renew, management's answer was that the TCO benefit embedded in higher-capacity drives (a 40TB drive versus a 30TB drive amortized over six or seven years of data center operation) is itself the customer's incentive, so Seagate does not need to price-subsidize the transition. Like-for-like prices continue to step up with each contract renewal, while the mix shift toward higher capacities provides the customer's economics.

Assessment: This is the sharpest possible contrast with historical HDD economics, where technology transitions were priced away almost immediately. If like-for-like escalators hold through the Mozaic 4 transition, the incremental-margin framework (roughly 50% above $2.6B quarterly revenue) is credible. The pricing line is where any future crack in the thesis will show first.

9. The NAND Question: The Interface Is Not Moving

"When it comes to mass data storage, the interface between all the NAND that's being used and all the hard drive bits that are being used is not changing that much." — Dave Mosley, CEO

The perennial substitution question got a structural answer: in cloud architectures NAND and HDD are complements at a stable boundary (flash for performance tiers and caching, hard drives for the mass store), and the capital cost of replacing HDD exabytes with NAND remains prohibitive, a point management quantified at its Investor Day. Mosley acknowledged NAND's strength at the edge and noted that recent NAND capacity discipline has kept its cost curve from bending toward HDD territory.

Assessment: With hard drives still storing on the order of 90% of bits in large-scale data centers and HAMR bending Seagate's own cost-per-terabyte curve down, the substitution threat window is not closing on HDDs this cycle. We treat NAND encroachment as a tail risk to monitor at the annual level, not a quarterly one.

10. Re-Segmentation and the Capex Step-Up: Reading the FY26 Setup

Two housekeeping items carry signal. First, from the September quarter Seagate will re-segment reporting into data center (~75% of FY25 revenue) and edge IoT, aligning disclosure with the actual business and, helpfully for analysts, making the cloud demand engine directly visible each quarter. Second, capex steps up from a historically depressed 3% of revenue in FY25 ($265M) into the 4–6% target band for FY26; management was explicit that this is replacement gear plus staging for the 4TB and 5TB-per-platter transitions in FY27–28, not gross capacity addition. The 5TB/disk platform remains targeted for early calendar 2028, with 10TB/disk to be demonstrated in the lab in the same timeframe.

Assessment: The capex philosophy (spend on density transitions, not on unit capacity) is the structural difference between this cycle and every previous one, because it means supply discipline is embedded in the capital plan itself. The step-up is modest in absolute terms (roughly $400–600M against $10B-plus of revenue) and fully consistent with the FCF expansion narrative.

Guidance & Outlook

The guide's framing was the quarter's controversy, so it is worth being precise about what is and is not in it. Revenue of $2.50B ± $150M is up 2% sequentially and 15% year-over-year at the midpoint, against a June quarter that management says over-pulled some demand and with a slice of September production reserved for customer qualification units on Mozaic platforms. The CFO was explicit about the mechanism: "Demand is strong, is above supply. So our guidance is mainly based on what we think we are ready to supply during the quarter and that volume of exabytes, they will be fully sold." The extra 14th week adds operating expense but does not add supply, because build-to-order production plans were fixed six to nine months ago.

Implied margin trajectory: mid-to-high 20s operating margin on $2.5B with ~$290M of opex implies non-GAAP gross margin of roughly 38.5–39.5%, a tenth consecutive record at the midpoint. Management separately committed that the December quarter will carry "higher revenue and higher profitability" than September, extending the sequential-improvement pattern through calendar 2025.

Street at: pre-print consensus for September stood at $2.51B and $2.33 (LSEG); the revenue midpoint is a rounding error below, the EPS midpoint sits under it on the tax rebasing. One other provider carried $2.28, above which the guide's midpoint actually lands; the providers themselves disagreed about whether this was a miss. That disagreement, at a stock priced at its all-time high, is what produced the after-hours violence.

Guidance style: Seagate has beaten or hit the top of its EPS guide in each of the past five quarters, and the June quarter itself finished at the top of the range guided in April. The pattern says these guides are floors built off committed supply, not stretch targets. We treat the $2.30 as conservative and the margin commentary (which management twice told analysts was better than their models) as the real signal.

Analyst Q&A Highlights

The Implied September Gross Margin Math

The opening question did the arithmetic the whole tape was doing in real time: against an eight-quarter average of 200bp-plus sequential gross margin expansion, the September guide seemed to imply only ~20bp. Management's response was a flat rejection of the premise, twice, with a reaffirmation that the first Investor Day milestone (40% gross margin around $2.6B of quarterly revenue) is close.

Q: "Over the last eight quarters, you've expanded gross margins by over 200 basis points sequentially on average. I believe at the midpoint of your guide… you're guiding to something like 20 basis points of sequential margin expansion. Can you maybe, one, just confirm kind of that math? And then two, just help us understand the puts and takes and maybe why we're not seeing that gross margin… expansion in September despite the confidence that we're hearing from you guys about getting to 40% gross margins in a few quarters."
— Erik Woodring, Morgan Stanley

A: "I would say your estimate is a bit low. Actually, I say is significantly lower than what is implied in the guidance… We are driving up gross margin and operating margin, I would say, significantly more than what you are modeling right now. So our path to achieve the milestone or the first milestone that we discussed at our Investor Day just a few weeks ago is intact. Now we are going exactly in that direction. I think we can be there fairly soon."
— Gianluca Romano, CFO

Assessment: A CFO telling the Street its margin models are too low, unprompted and with specificity, is the single most useful data point from the Q&A. If September prints gross margin near 39%, the 40% milestone likely falls in the December quarter, two quarters ahead of what consensus modeled entering this print.

Guiding Below Consensus While Claiming Demand Exceeds Supply

The apparent contradiction between a below-consensus revenue guide and sold-out capacity was put to management directly, along with a pointed observation about the receivables jump. The answer separated the two cleanly: revenue guidance is a supply statement under build-to-order, with a portion of production explicitly reserved for qualification units at the customers being onboarded to HAMR.

Q: "You had a strong quarter, clearly, but you guided revenue slightly below consensus for the September quarter and your DSO also jumped up. So I was wondering if you could clarify if there was anything that you would point out in linearity in the quarter? And I guess the question is, how well is your HAMR capacity ramp aligned with qualifications and demand?"
— Wamsi Mohan, Bank of America

A: "Demand is strong, is above supply. So our guidance is mainly based on what we think we are ready to supply during the quarter and that volume of exabytes, they will be fully sold. We also need to dedicate a little bit of our production to qualification… And when you look at our guidance, of course, you need to remember that starting this quarter, we will be subject to the global minimum tax… When you model all those things correctly, you will see actually a fairly good improvement in both gross margin and operating margin."
— Gianluca Romano, CFO

Assessment: "That volume of exabytes, they will be fully sold" is as close as a CFO gets to calling revenue guidance deterministic. The residual risk in the September number is production execution and qualification timing, not demand. The receivables question got a separate, adequate answer: the company simply stopped factoring receivables because cash flow no longer requires it, which is a quality upgrade dressed up as a DSO deterioration.

HAMR Adoption Beyond the Lead Customer

The most direct probe of the quarter's key de-risking event: whether customers beyond the first hyperscaler are actually taking HAMR volume, or whether the ramp remains a single-customer story with a qualification veneer.

Q: "A large part of the ramp thus far has been with a single customer. You're talking about multiple customers qualified. At this point, are customers two, three, et cetera, outside of customer one making up a significant portion of the ramp, like 10% or more, let's say? I just… I'm trying to get an understanding of the adoption outside of the first guy."
— Thomas O'Malley, Barclays

A: "Simple answer to your question is yes. The other customers are starting to ramp as well, and the pull is pretty strong also… everybody has pretty good visibility, and we have multiple customers pulling hard. And that's indicative of the exabyte demand… we're not really trying to add gross capacity of number of drives. We're trying to get through these transitions as fast as we can to be much more efficient with the exabytes."
— Dave Mosley, CEO

Assessment: An unhedged "yes" on multi-customer HAMR pull removes the concentration objection that dominated the pre-print bear case. The follow-up disclosure that the old ~160 EB output ceiling "is not a wall so much anymore" once Mozaic 4 ramps quantifies why the transition cadence, not tooling, is the capacity plan.

What "Booked Through Mid-2026" Actually Means

A skeptical question tested whether the extended visibility claim was genuinely new information or a definitional artifact of year-long build-to-order lead times, and pushed management to say whether December could decline. The answer produced the call's only piece of forward guidance beyond September.

Q: "You said the capacity is booked out to mid-'26 and visibility is extending into the back half. What does that mean? Because build-to-order… the lead time to build a drive is a year. So if you place an order now, you're not going to get the drive until this time next year anyway… could you tell us what you're going to ship in December? I'm just wondering if something changed for you to give that comment."
— Timothy Arcuri, UBS

A: "We said previously also in prior calls, we expect calendar '25 to sequentially increase revenue and profitability. So we continue in that direction. So December will be higher revenue and higher profitability."
— Gianluca Romano, CFO

Assessment: The December commitment is the answer to the guide controversy hiding in plain sight: a company worried about demand does not volunteer that the next un-guided quarter will grow revenue and profitability sequentially. Mosley's fuller answer also clarified that customers contract for exabytes rather than specific drive models, which is why qualification timing shifts mix but not the order book.

Who Keeps the HAMR Cost Savings

The pricing question that determines the entire margin thesis: as long-term agreements renew and HAMR's cost-per-terabyte advantage compounds, do the economics accrue to Seagate or get competed away to the hyperscalers?

Q: "Can you just touch on what pricing assumptions are you seeing embedded on an exabyte basis in these contracts? And really, as you think about the HAMR products starting to ramp up over the next 12 months, do you end up in a pretty good cost advantage… on a per exabyte basis on HAMR exabytes. Do you think these LTAs will enable you to keep those cost savings for Seagate?"
— Amit Daryanani, Evercore

A: "We don't have to incentivize the transition. I mean there's a significant TCO benefit of running these new products in your data center. So if you think about a 40 terabyte versus a 30 terabyte, for example, and you're going to run that for six or seven years, that's a huge TCO benefit. So there's an incentive baked in right there. We know what our costs are going to be, and we know what pricing we want to incentivize and we know what margin that we need to be able to go back and refeed our R&D and our supply chain and everything else."
— Dave Mosley, CEO

Assessment: The answer is that the customer's payoff comes from drive-level TCO, not from per-terabyte price concessions, which lets Seagate hold like-for-like pricing (the CFO confirmed escalators continue) while both sides win. This is the mechanism that converts areal-density leadership into gross margin, and it survived direct questioning intact.

The 50% Incremental Margin Framework

A modeling question tried to reconcile the September guide against the Investor Day's ~50% incremental gross margin framework that begins at $2.6B of quarterly revenue, probing for hidden near-term headwinds below that threshold.

Q: "It seems that your guidance implies incremental gross margins about like 50%, even though we are still below the $2.6 billion revenue baseline you outlined on the Analyst Day… why would your incremental gross margin not be better than the 50% number you guided at the Analyst Day?… are there some headwinds in the near term?"
— Hadi Orabi (for Krish Sankar), TD Cowen

A: "Every quarter will be a little bit different depending exactly from the mix that we change quarter-over-quarter. But I would say our first goal is to achieve the $2.6 billion in revenue and the 40% gross margin. And I think we are trending well in that direction. And after that, our goal is to continue to increase revenue in the low to mid-teens as we discussed at the Analyst Day and increase our profitability for an incremental 50% gross margin. So I would say nothing changed in the last eight weeks."
— Gianluca Romano, CFO

Assessment: "Nothing changed in the last eight weeks" is the summary of the entire call. The Investor Day framework ($2.6B revenue and 40% gross margin as milestone one, low-to-mid-teens growth with 50% incrementals beyond) survived its first quarterly checkpoint without amendment.

CapEx Step-Up and the Definition of Excess Cash

With the balance sheet at target leverage and the buyback resuming, the capital allocation question turned to why capex rises into the 4–6% band and how much cash the company intends to hold before returns accelerate.

Q: "That would seem to imply a fairly healthy uptick in the CapEx spend year-over-year for fiscal '26 versus fiscal '25? And I guess why would that be? And… as we see the generation of free cash flow, you've hit the sub-$5 billion gross debt level. How do we think about the right level of cash operationally you're willing to hold?… excess cash being built and capacity for share repurchase."
— Aaron Rakers, Wells Fargo

A: "Free cash flow is improving a lot. You have seen already in the June quarter, a major step-up. This will continue… We have reduced our debt, as you said, at the target level we were targeting since more than a year at this point. And Dave just announced that we are restarting share buyback… In terms of liquidity… we don't have excess cash right now. I think we can maybe still increase a little bit our cash position, but the vast majority of our free cash flow, of course, will go back to our shareholders through the dividend and through the share buyback."
— Gianluca Romano, CFO

Assessment: The capex answer (replacement gear plus staging for the 4TB and 5TB platter transitions, still inside the 4–6% band) confirms capital intensity stays structurally low; the cash answer commits the vast majority of a growing FCF stream to returns. The capital allocation setup entering FY26 is the cleanest in the company's modern history.

What They're NOT Saying

1. HAMR's revenue contribution remains undisclosed: Management declined, again, to quantify Mozaic's share of revenue or exabytes. The refusal is defensible while qualification mix shifts quarter to quarter, but it means the market cannot independently verify the ramp's slope, and the eventual first disclosure will itself be a market event.
2. No fiscal-year 2026 guidance: The company guided one quarter and pointed to Investor Day CAGRs for everything beyond, with only the volunteered December commitment ("higher revenue and higher profitability") bridging the gap. A company with four quarters of contracted visibility could guide further; choosing not to preserves negotiating room with customers but leaves the Street to model the tax-rebased year on its own.
3. The whisper on gross margin was answered with adjectives, not numbers: The CFO twice said analyst margin math was too low without stating the implied September gross margin. Our arithmetic says 38.5–39.5%; if management believed a 40%-handle was possible in September, saying so would have cost nothing. The reticence suggests the 40% milestone lands in the December quarter, not September.
4. Enterprise OEM demand is "stable," not recovering: Nearline sales into enterprise OEM improved modestly and are expected to be "stable over the next few quarters." There is no server-attach recovery call embedded in the model, which is conservative, but also means the cloud is carrying the entire growth load.
5. Nothing on competitive dynamics: No analyst asked about the pure-play HDD competitor's roadmap and management volunteered nothing. With the rival promoting energy-assisted PMR extensions while Seagate ships HAMR, the silence is comfortable incumbency; the risk is that the competitive gap is being taken for granted at precisely the moment the competitor's own margins are inflecting.
6. Tariffs got one sentence: "Minimal tariff-related impacts" based on the current outlook, with mitigation strategies "ready to deploy." Given the July trade-policy environment and Seagate's Asian manufacturing footprint, the brevity is a choice. Any change in the tariff treatment of storage hardware would hit the model with no cushion in guidance.

Market Reaction

• Pre-print setup: STX closed July 29 at $152.68, an all-time closing high territory print (52-week closing range $66.54–$152.76), up 76.9% year-to-date against the S&P 500's +8.3%, up 47.7% over the trailing twelve months, and up 7.9% over the trailing 30 days. The entire year-to-date gain accrued after the April earnings report.
• After-hours move (July 29): Down more than 8% in extended trading within an hour of the release, with headlines uniformly attributing the drop to the September guide (midpoints below one major consensus provider) rather than the June quarter itself.
• Reaction session (July 30): Opened at $141.00 (down 7.6%), printed the session low of $138.30 (down 9.4%) in the morning, then recovered through the day to close at $147.42, down 3.4%, on 12.0M shares, 3.1x the 30-day average volume. The S&P 500 was flat (down 0.1%).
• Peer tape: No contagion. The closest HDD comparable gapped down about 1.6% in sympathy at the open and finished up 1.2% ahead of its own results that evening; the major memory name rose 2.5%. The market treated Seagate's move as company-specific positioning, not a storage demand signal.

The whisper reset: the mechanics of the sell-off are a case study in expectations math. Consensus sat above the guide midpoint on both lines for September, the stock had run 77% in seven months, and four consecutive beats had trained the market to expect guides that clear consensus. A guide that was arguably in line (one provider's consensus was below the EPS midpoint) triggered a 9% drawdown at the low because at an all-time high, in-line is the new miss. The tax rebasing amplified the optics: a $0.29 sequential EPS decline headline reads as deceleration to anyone who has not modeled the 16% tax rate.

The defended pullback: the recovery from down 9.4% to down 3.4% over a single session, on triple average volume, is itself information. The morning brought a wave of pre-open sell-side support, including an upgrade explicitly framed around the sell-off and multiple raised targets characterizing the pullback as a buying opportunity, with new Street-high objectives well above the pre-print price. Long-only demand met the gap. When a 9% morning drawdown cannot hold through a single session despite genuinely disappointing guide optics, positioning is heavy but conviction in the underlying story is heavier.

Street Perspective

Debate: Is the Flat Sequential Guide a Demand Signal or a Supply Artifact?

Bull view: Revenue guidance under build-to-order is a statement about factory output, not demand: capacity is sold out through mid-calendar 2026, qualification units are consuming a slice of September production, and management volunteered that December grows revenue and profitability sequentially. The +2% guide on a quarter that over-pulled is supply arithmetic, and every exabyte produced "will be fully sold."

Bear view: Every storage cycle in history has ended with management insisting demand was intact while sequential growth flattened. The June quarter's DSO jump, the extra week producing only 2% growth, and cautious enterprise commentary are early-cycle-top tells; sold-out today says nothing about the price and volume at which fiscal 2027 contracts get written.

Our take: The bull mechanics are verifiable and the bear pattern-match is not, this quarter: receivables rose because factoring stopped (a cash flow quality improvement), the extra week adds opex but not supply by construction, and the December commitment is the opposite of cycle-top behavior. We side with the bulls on the September question specifically, while conceding the bears' deeper point that build-to-order has never been tested by a demand downturn.

Debate: Is 38% Gross Margin the New Baseline or the Cycle Peak?

Bull view: Nine consecutive record quarters, like-for-like pricing escalators in contract renewals, HAMR cost-per-terabyte declines still ahead of the volume ramp, and a stated 40% milestone "fairly soon" describe a structural repricing, not a cycle. The mix shift to nearline under multi-year contracts has permanently changed the demand elasticity Seagate faces.

Bear view: Two years ago this company printed sub-20% gross margins and burned cash; the current margin is the product of a historic supply squeeze meeting an AI demand pulse. When capacity catches up (or a hyperscaler digestion cycle arrives), the contracts roll off and the margin structure reverts toward the long-run HDD mean. Paying an all-time-high price for peak margins is the classic cyclical trap.

Our take: The structural case has the better evidence today: the margin expansion has been continuous across eight quarters of very different demand conditions, the pricing mechanism is contractual rather than spot, and the capex philosophy (density transitions, not unit capacity) keeps industry supply growth bounded. But "structurally better" and "never cyclical again" are different claims; we underwrite the first, not the second. The margin floor in the next downturn, whenever it comes, is the single biggest unknown in the valuation.

Debate: After a 77% Run, Does the Pullback Create the Entry?

Bull view: At roughly $147, the stock trades near 14–15x the forward earnings run-rate implied by the tax-rebased guide, for a company guiding low-to-mid-teens revenue growth with 50% incremental margins, an inflecting FCF stream, and a resuming buyback. The multiple has not caught up to the margin structure; the pullback is the entry the morning's upgrades called it.

Bear view: The stock has nearly doubled since April on a story that is now fully told: the Investor Day model is public, the HAMR roadmap is public, and consensus already extrapolates the margin march. With EPS optically declining next quarter on the tax reset and no fiscal-year guide, the catalysts for the next leg are thin while the air under a 77% YTD gain is thick.

Our take: Both sides are right on their own horizon, which is what a Hold means. The forward multiple is genuinely undemanding if the FY28 framework holds, but the market just demonstrated (a 9% morning drawdown on an in-line guide) how little tolerance the price has for anything short of acceleration. We prefer to pay up after the September print evidences the 40% margin trajectory and the Mozaic 4 qualification, accepting the risk that the evidence arrives at a higher price.

Model Update Needed

Valuation impact: On a tax-rebased FY26 non-GAAP EPS path of roughly $9.75–10.50 (four quarters at $2.30 rising toward $2.75 as revenue steps toward $2.7B and margins toward 40%+), the July 30 close of $147.42 implies 14–15x forward earnings. Our initiating fair-value range of $145–160 (14–16x the midpoint, justified against the 50%-incremental-margin framework but haircut for cycle history) brackets the current price, implying roughly +4% to the range midpoint: consistent with a Hold. Evidence that the 40% gross margin milestone lands by December and that Mozaic 4 qualification is converting would justify migrating toward 17–18x, or roughly $175–185.

Thesis Scorecard Post-Earnings

Overall: The operating thesis (HAMR moat, margin transformation, exabyte supercycle) was confirmed on every measurable checkpoint this quarter, while the valuation bear point was also confirmed by the tape's behavior. That tension is the rating.

Action: Initiate at Hold with a $145–160 fair-value range. Add on evidence, not on price: the September print showing ~39% gross margin and December guidance consistent with the "higher revenue and higher profitability" commitment would move us to Outperform even at a higher entry. A HAMR qualification slip or any crack in like-for-like pricing would move us the other way.