A beat made of the ingredient WDC ran short of
Start with what is not in dispute. Revenue was $3,629M, up 48% year over year and 17% sequentially, $79M above the top of Seagate's own $3.45B guide. Non-GAAP EPS of $5.71 cleared the top of its $5.00 plus or minus $0.20 range by $0.51. Non-GAAP gross margin rose 570 basis points in one quarter to 52.7%. Dave Mosley opened the call with "Seagate delivered a very strong finish to an outstanding fiscal 2026," and the filing agrees with him line for line.
What makes this quarter worth a second look is the composition. On this site read Western Digital's June quarter as growth that changed hands: 44% revenue on 22% more exabytes, with price supplying 18.0%. Seagate's June quarter was the mirror image. Exabytes rose 34% to 218. Revenue per terabyte shipped rose 10.6%. Volume carried it. The market noticed the difference: the day WDC fell 13.03% on its print, STX rose 1.83% (unverified, market data). What the market is paying for in that gap is exabytes that arrive without the customer's permission.
My argument is that the September guide has already spent it. Read it in exabytes and it is a price quarter.
Read the ingredients, not the level
Read the ingredients, not the level. A hard-drive quarter is exabytes times revenue per terabyte, and Seagate gives both. Across eight quarters the second line ran $15.77, $15.42, $15.04, $15.05, $14.48, $14.87, $15.61 and $16.65. For most of that stretch capacity grew faster than revenue, which is what a density transition looks like when it is priced normally: every higher-capacity drive adds more terabytes than it adds dollars. The measure bottomed a year ago and has risen in each quarter since. Erik Woodring of Morgan Stanley put the June figure to management in one line: "You just reported 10% year-over-year price per exabyte growth in June." Romano's answer began: "You are correct."
Put the two ingredients in one bridge and the quarter sorts itself. Revenue rose $1,185M. At last year's $15.05 per terabyte, the extra exabytes account for $836.7M of it, 70.6%. The rise in revenue per terabyte accounts for $348.3M. In the March quarter, on the same arithmetic, volume supplied $839.3M of a $952M gain, 88.2%. June was still mostly a volume quarter. It was less of one than March.
The rest of this dive is for paid subscribers.
The headline numbers are above. The 8 sections below carry the mechanism, the peer read, the valuation work, and the dated tests that decide it.
- The June step
- HAMR needs no permission from the customer. It needs it from the factory.
- Price, said out loud
- Cash, and what it is being used for
- The strongest case against me
- What decides it
- The trade (analysis, not advice)
- Bottom line
Already a subscriber? Sign in.
The June step
Now the series that matters for September. Exabytes grew 4.7% sequentially in the December quarter and 4.9% in March. In June they grew 9.3%, with nearline going to 195 from 175.4. Gianluca Romano, the chief financial officer, described it as supply running a little hot: "our volume was a little bit higher in fiscal Q4," with "maybe a little bit of output available in fiscal Q1," and, in the same answer, "we will take a little bit more pricing benefit." The house rule for a single period is that it stays a phase until the series says otherwise, and the series here says June's step is the outlier, not the run rate.
Hold that against the guide. September revenue of $4.1B is "a 56% year-over-year improvement at the midpoint," against a quarter of 181.5 exabytes. If exabytes grow at the company's own "mid-20%" target, say 228, up 25.6%, revenue per terabyte has to reach $17.98, up 24.1%, and volume supplies 45.8% of the gain. Less than half, for the first time in this series. If exabytes repeat June's sequential pace and reach 237, up 30.6%, revenue per terabyte still has to rise 19.4% to $17.30. Woodring did the same sum on the call and got "pricing growth closer to maybe 20%." C.J. Muse of Cantor Fitzgerald called it "that strong 20 plus percent number embedded in the September guide." Either way the guide leans on price. The question is only whether volume keeps up enough to share the load.
HAMR needs no permission from the customer. It needs it from the factory.
The tkal WDC dive drew the line between the two roadmaps as density growth is elective; capacity growth is not: HAMR raises terabytes per drive, so a HAMR exabyte ships whether or not the customer changes anything. That remains true, and HAMR was "approximately 40% of our nearline exabyte shipment run rate" exiting the year. What the call added is where HAMR's permission comes from instead. Mosley: "we're not really increasing the box count." Moving from Mozaic 3 to 4 to 5 means "we have to actually put our factories on pause to go through the product transition a little bit." Romano: units were "absolutely flat" last year while disks and heads grew "between 15% and 20%." Asked whether areal density alone could meet 2029 demand, Mosley said: "I do not think that probably our areal density transitions are going to be sufficient." Note that I am not the one calling the transitions a bottleneck. The chief executive is.
The balance sheet shows the strain as inventory and construction. Work-in-process rose 29.8% to $1,088M while finished goods fell; construction in progress rose 26.4% to $421M; capital expenditure rose 114.7% to $569M, 2.1x depreciation of $268M against 1.06x a year earlier, and the 10-K says fiscal 2027 will be higher again. This is where Seagate and WDC part company, and where the house model needs care. Price does the work the capex hasn't done yet, in its asset form, described WDC, whose construction-in-process fell 14.10%. It does not describe Seagate. Seagate is building. It is building to hold exabyte growth at the mid-20s through three product transitions, not to exceed it, and every drive held in work-in-process during a transition is an exabyte that ships a quarter late.
Price, said out loud
Management does not call this a change. Romano: "it's not that we are changing our strategy, but for sure the gap between supply and demand is now a little bit bigger than few quarters ago." Mosley described how the extra output gets sold: when Seagate has "more exabytes to give," then "usually they'll pay more than that contract price." Take them at their word. Demand is contracted through calendar 2027, and "the vast majority of our nearline exabytes are now allocated into calendar 2028." When the volume is already allocated, the only thing a tighter supply-demand gap can change is the price. Mix is the softest margin to capitalize, and the 10-K's own attribution of the year's 11-point gross margin gain is "pricing actions" and "product mix shift to higher capacity products," in that order.
The concentration raises the stakes on that price. One customer was about 14% of fiscal 2026 revenue, up from 10%; three customers were 18%, 16% and 10% of receivables at year end. A buyer that size, watching its supplier move from a 37.4% to a 52.3% GAAP gross margin in a year, will price its next long-term agreement with that in mind. The call did not say how much of calendar 2028 is priced rather than merely allocated, and nobody asked it that way.
Cash, and what it is being used for
The cash is real. Free cash flow was $1,118M in the quarter, a 31% margin, and $3,105M for the year against $818M. Net debt fell to $1,899M, 0.4x trailing adjusted EBITDA; Seagate redeemed $1B of senior notes in July, settled the last of its 2028 exchangeable notes on with about $150.97M in cash and 1,647,862 shares, and Romano expects debt of about $2.4B at the end of fiscal Q1. Fiscal 2026 capital return was dividends of $634M and repurchases of $176M. None of this is fragile. It is also not evidence about exabytes, which is the only line I am arguing about.
The strongest case against me
Here is the other side at full strength, and it is not weak. HAMR is on a published ramp: 40% of nearline exabytes now, 50% of HAMR exabytes on Mozaic 4 exiting calendar 2026, 70% of nearline on HAMR by , and management says it is "on track." Seagate has beaten its own mid-20s target for a year: fiscal 2026 exabytes grew from 595 to 789, nearline from 497 to 695. Citigroup's Asiya Merchant framed the question as whether "this about 30% exabyte growth rate" can hold, not whether 25% can. Romano himself said there may be "a little bit of output available in fiscal Q1." A company that just produced a 9.3% sequential step, with Mozaic 4 only in its second quarter of shipments, may simply do it again. On that reading June was the new pace, the guide is conservative on volume as it was on revenue last quarter, and I am reading a supply ceiling into a factory that is in the middle of lifting it.
I take it seriously and still land the other way, for two reasons. The first is that management chose to repeat "mid-20%" as the target on a call where it had every incentive to raise it, and paired it with a description of paused factories and strained component fabs. The second is the work-in-process line: a 29.8% build in partly finished product is what longer transition routes look like on a balance sheet, and it argues for exabytes arriving later, not sooner. If the steelman is right, the number shows it within a quarter, which is why the test is set where it is.
What decides it
One number, on one date: HDD exabytes shipped for the quarter ending , reported with fiscal first-quarter results. Below 228, about 25.6% above the 181.5 of a year earlier and a return to the 4.7% to 4.9% sequential pace of December and March, and the $4.1B quarter is being met mostly on price. Seagate would then be running the model tkal found at Western Digital, with more capex behind it. At 237 or more, up 30.6%, June's step was HAMR capacity arriving on schedule, volume still carries more than half the growth, and I'm wrong. Between the two is a partial result, and I'd call it that. The secondary check is revenue per terabyte: at the guide it lands between $17.30 and $17.98 across that exabyte range, against $16.65 in June. WDC reports two days later, on , so the two compositions will sit side by side within the week.
The trade (analysis, not advice)
At $913.45 the stock is 22.23% above its pre-print close and trades at 58.6x fiscal 2026 non-GAAP EPS of $15.58 and 31.3x the $7.30 guide annualised (unverified, market data). The guide is almost certainly achievable. What the multiple is paying for, judging by how STX traded on WDC's day, is the belief that Seagate's growth has a volume floor WDC's lacks. If first-quarter exabytes print below 228, that belief takes its first direct hit even on a revenue beat, because the beat will have come from the ingredient both companies share. If they print at 237 or above, the HAMR premium is earned for another quarter and the gap to WDC widens. The asymmetry sits in the composition, not the headline, which is why the headline on will say very little.
Bottom line
Seagate's June quarter was excellent and mostly volume: 218 exabytes, up 34%, supplying 70.6% of a $1,185M revenue gain, with revenue per terabyte up 10.6%. The September guide is a different quarter. It needs revenue per terabyte up 19.4% or more at any exabyte count up to 237, and management's own target puts volume at less than half the gain. My view is that June's 9.3% sequential step was the outlier and the transitions set the pace. Wrong if first-quarter exabytes reach 237. Right if they come in below 228.