Home / Transcripts / Western Digital Corporation (WDC) · August 5, 2026

Western Digital Corporation (WDC) Earnings Call Transcript & Summary

August 5, 2026

NASDAQ US Information Technology Technology Hardware, Storage and Peripherals earnings 48 min

What were the key takeaways from Western Digital Corporation's August 5, 2026 earnings call?

In the fourth quarter of fiscal 2026, Western Digital Corporation (WDC) reported revenue of $3.75 billion, a 44% increase year-over-year, and earnings per share (EPS) of $3.56, up 109% year-over-year. The company highlighted strong demand driven by AI and cloud services, with gross margins expanding to 54.4%. Management provided guidance for Q1 fiscal 2027, expecting revenue of $4.1 billion, reflecting a 45% year-over-year growth, and EPS of $4, indicating continued confidence in demand and pricing dynamics.

What topics did Western Digital Corporation cover?

What were Western Digital Corporation's August 5, 2026 results?

WDC's strong performance in Q4 fiscal 2026, driven by robust demand from AI and cloud services, positions the company favorably for fiscal 2027. The guidance for continued revenue growth and margin expansion suggests a positive outlook. However, analysts are cautious about the deceleration in exabyte growth and the implications of LTAs on pricing. Investors should monitor the ramp-up of new products and the evolving demand landscape.

Earnings Call Speaker Segments

Ambrish Srivastava executive
#1

Thanks, and good afternoon to all of you. With me on this call are Irving Tan, who serves as WD's Chief Executive Officer, and Kris Sennesael, WD's Chief Financial Officer. Ahead of getting started, we want to remind you that our conversation today includes forward-looking statements that rest on management's present assumptions and expectations, and those carry a range of risks and uncertainties. Among these forward-looking statements are what we expect for our product portfolio, our business plans and performance, ongoing market trends and our future financial results. We take on no obligation to update these statements. For additional detail about the risks and uncertainties that might cause actual results to differ materially from expectations, please see our most recent annual report on Form 10-K along with our other filings with the SEC; during our prepared remarks, unless we say otherwise, our comments will relate to non-GAAP results on a continuing operations basis. Reconciliations between the non-GAAP and comparable GAAP financial measures appear in the press release and other materials being posted in the Investor Relations section of our website at investor.wdc.com. Finally, I want to point out that when we say we, us, our or similar terms, we mean only Western Digital as a company and are not speaking for the industry. With that, I'll hand the call to Irving for his introductory remarks. Irving?

Tiang Yew Tan executive
#2

Thanks, Ambrish. Hi, everyone, and thanks for joining us this afternoon. I want to begin by reviewing our first full fiscal year running WD as a standalone HDD pure-play company. Revenue grew a remarkable 36% year over year, and we also expanded gross and operating margins considerably. Earnings per share reached twice what they were before, we generated $3.5 billion in free cash flow, and our balance sheet improved to a net positive cash position, all while continuing to return substantial capital to shareholders. Those robust financial outcomes resulted from satisfying our customers' rising storage needs through a focus on innovation and by pursuing operational excellence throughout the worldwide WD organization. We are now entering fiscal year 2027, supported by solid customer demand, greater visibility, and ongoing confidence in both the staying power of that demand and our capacity to fulfill it with our market-leading products and technology road map. Storage demand is being propelled by AI as well as by our core cloud services. A key dynamic is emerging within the AI infrastructure market. You can reuse compute cycles, and the workloads for compiling data, training, and inference will share and reuse those compute resources as time goes on. However, the data all these workloads generate — things like model inputs, outputs, locks, and retained contacts — just keeps accumulating. As AI usage grows, that drives an expanding need for storage infrastructure that can store and manage these enormous data sets in a cost-effective way. The inflection point we've talked about before, moving from AI training to inference to Agentic AI, has only gotten sharper. When you train models, you get substantial upfront data requirements, yet inference is continuously producing and keeping data. Right now, the biggest AI platforms are handling tens of billions of tokens every minute, along with billions of prompts each day. That produces a fast-growing collection of data that has to be kept, administered, and retrieved as time goes on. Newly released figures indicate token counts are rising by several full multiples compared with a year earlier, which highlights how quickly inference is expanding. At the same time, AI is shifting fast from simply giving answers to agentic AI that carries out the work — organizing tasks, reaching into data, and running nonstop through workflows made up of many steps. That change generates a workload that is far more data-hungry, and it is also becoming more lasting instead of short-lived. What this means for storage is considerable. Every action in an agent's workflow produces data, so the total amount created grows along with the amount that has to be kept as time passes. That's the reason Agentic AI still looks to us like a structural step function driver behind demand for capacity orientated storage. In addition to inference and Agentic AI, physical AI is starting to appear, along with autonomous vehicles, robotics and industrial automation systems and humanoids, and in those areas there isn't enough real world data to train the systems. So synthetic data sets have to be generated and stored, which becomes one more driver of storage demand. Investing in infrastructure matters, but that's just the starting point. The initial data foundation comes from training. Data gets generated continuously by inference. With agentic systems, both the amount of that data and how often it appears go up, and physical AI speeds the loop up even more. Combine those two trends and you end up with a stronger backdrop for demand in data storage — driven not just by the initial build-out of AI infrastructure, but also by the reality that data continues to be created and retained after that infrastructure is operational. As AI workloads move past the construction phase into steady operation, storage demand is no longer chiefly tied to one infrastructure spending cycle; instead it becomes chiefly tied to data accumulating — and that compounding effect is the secular growth engine for demand beneath our business. At present, about 80% of the data kept in hyperscaler data centers sits on hard disk drives. Odds are that stays true. It comes down to what HDDs are uniquely good at: supplying the scale, the economics and the power efficiency that long-retention, large-scale data storage calls for. Let me shift now to why these secular growth engines line up with what WD does best. WD's strong technology road map rests on industry-leading areal density per platter, with an emphasis on innovating to serve our customers' capacity needs at scale, plus our expanding capacity to satisfy their requirements by delivering cost-effective storage solutions across additional layers of the AI storage stack. Our 44 terabyte HAMR product is still on track to go out in the first half of calendar 2027. Customer response to the qualification effort remains very positive — the drives' capacity, performance and reliability are exceeding expectations. For the next-generation 40 terabyte ePMR drives, shipments started in our June quarter, and with two customers we are now transitioning into volume production. With the third major customer, we are currently ramping our out Ultra SMR technology. As fiscal 2027 comes to a close, we expect Ultra SMR to represent about 60% of our near line Exabyte shipments. Since the previous time we spoke, our pipeline for engaging and qualifying customers has continued to expand, and more hyperscaler and cloud clients are now working through qualification and planning for deployment. Beyond what capacity can do on its own, we're also driving innovation into additional layers of the AI storage stack; we're making progress on improving drive performance ahead of high bandwidth drives and have begun sampling with five customers, reaching up to 8x the throughput of drives available today while power drop doesn't rise correspondingly — exactly the kind of performance these workloads demand; in closing, the opportunity before us is real. And here at WD, we're well positioned to capture it. Data creation isn't slowing down. It's accelerating. When data volumes keep growing, the storage infrastructure needed to hold, administer, and safeguard that data grows too. For our company, that creates a lasting, long-term advantage, and we intend to make the most of it. We have a solid technology road map, deep ties with our customers, and the operational rigor required to convert this chance into ongoing earnings growth, free cash flow growth, and long-term shareholder value. I'll now hand things over to Kris, who will walk you through the financials and what we expect for Q1.

Kris Sennesael executive
#3

Thanks, Irving, and good afternoon to all of you. For WD, fiscal 2026 proved to be a remarkable year, powered by wider demand, stronger relationships with customers, and careful operational discipline. Our revenue climbed 36%, reaching EUR 12.9 billion, as gross margins rose 970 basis points to 49.1%, and operating margins went up 1,290 basis points to 37.3%. Earnings per share more than doubled, hitting $10.22, and free cash flow came to $3.5 billion, which meant a solid 27% free cash flow margin. To shareholders, we gave back $3.1 billion, which shows both our belief in how durable this business is and our dedication to creating value over the long term. Now let me move on to our fourth quarter of fiscal '26. Revenue held at $3.75 billion, a 44% increase compared with the same period last year, supported by robust Exabyte growth and advantageous pricing trends. Earnings per share rose 109% versus a year ago, reaching $3.56. Revenue, gross margin, and EPS each landed at or above the upper end of our guidance range. Our shipments of Exabytes to customers rose 22% compared with last year. Nearline remained the primary engine of our Exabyte expansion, with healthy non-nearline Exabyte growth adding to it during the quarter. In our fiscal fourth quarter, we started shipping next-generation EPM drives offering capacities as high as 40 terabytes, and we anticipate a robust ramp across the coming few quarters. Cloud accounted for 89% of total revenue, or $3.3 billion, up 43% from a year earlier, as demand stayed strong for our high cast nearline products amid a favorable pricing backdrop. Client made up 6% of total revenue at $225 million, rising 61% compared with last year. Consumer accounted for 5% of revenue at EUR 187 million, an increase of 38% versus a year ago. Improved pricing helped both segments. Gross margin grew 1,310 basis points from last year to 54.4%, producing strong incremental gross margin year-over-year. The drivers were a shift in mix toward drives with greater capacity, advantageous pricing throughout our portfolio, and careful execution in our manufacturing operations. In the quarter, the blended average price increase per terabyte compared with last year went from high single digits in the prior quarter to high teens in this one, which mirrors how our predictable and sustainable pricing approach is paying off as we provide more value to customers. Operating expenses totaled $382 million, roughly 10% of revenue, a 170 basis point improvement sequentially, showing additional operating leverage in the model. Robust top line growth, widening gross margins, and leverage in the model pushed operating income to EUR 1.66 billion, up 126% from last year, which translated into a durable operating margin of 44.2%, an increase of 1,610 basis points year-over-year. Interest and other expenses came to $10 million, and our effective tax rate was $0.16. With a diluted share count of 388 million, earnings per share reached $3.56, up 109% compared with last year. Operating cash flow was $1.4 billion, and CapEx was $108 million; this produced $1.3 billion in free cash flow for the quarter and a strong 34% free cash flow margin. In the quarter, we finished monetizing the remaining 1.7 million shares of SanDisk by exchanging them for 4.8 million WD shares. On top of that, we bought back 2.3 million shares of common stock, spending $1 billion in total. The figures we're giving today for both the full year and the fourth quarter include $328 million used to settle the conversion premium on certain converts in cash instead of stock, which spared us from issuing about 773,000 new shares. Dividend payments came to $54 million, and we closed the quarter holding $1.1 billion in debt alongside $1.6 billion in cash, which leaves us at a net positive cash position of $500 million. Finally, the Board declared a cash dividend of $0.15 per share on our common stock, payable on September 17, 2026, to shareholders of record on September 8, 2026. Now I'll move on to the outlook for the first quarter of fiscal 2027. Demand remains strong, our longer-term visibility keeps improving, and pricing dynamics are favorable across our cloud, consumer and client end markets. We're projecting revenue of $4.1 billion, plus/minus $100 million. That midpoint works out to 45% growth year-over-year. Gross margin should land between 55% and 56%. Operating expenses are anticipated to come in between $390 million and $400 million. Interest and other expenses should be $15 million. We expect a 17% tax rate. Consequently, diluted earnings per share should be $4, plus/minus $0.15, on a non-GAAP diluted share count of 388 million shares. Looking forward, we have strong confidence in what's driving demand: AI, cloud and the spread of data-intensive workloads keep speeding up. When it comes to supply, our industry-leading technology and product road maps, paired with solid operational execution, allowed us to ship far more Exabytes to our customers; achieving that doesn't demand CapEx spending on additional unit capacity, though we are investing what's needed in our heads and media operations, plus automation, to boost productivity; in short, durable demand, disciplined supply and expanding margins put us in position to deliver sustained earnings growth, robust free cash flow and long-term shareholder value. With that, let's open the Q&A. Ambrish?

Ambrish Srivastava executive
#4

Thanks, Kris. Operator, please go ahead and unmute the lines for Q&A now. And so that we can get to as many analysts as we can, please limit yourself to a single question at a time. Once you've done that, we'll come back to you so you can pose one additional question.

CJ Muse analyst
#5

It's tough not to stack up your numbers against your biggest rival, since they're posting stronger quarter-over-quarter revenue gains and their gross margin target sits almost 200 bps above what you guided to in September; so I'm wondering how you read that? Does it come down to their head start on HAMR, or maybe contracts landing at different times, or possibly you're putting fewer bits out on the open market? I'd really like to hear what you think about that, and also how you view gross margin climbing from this point? Is it a matter of holding out until HAMR arrives in calendar '27? Or could other forces also let you drive it upward?

Tiang Yew Tan executive
#6

Appreciate that question, C.J. gross margin always moves around from one quarter to the next, since a wide set of drivers shapes it. Clearly, one is how pricing lines up against the long-term agreement contracts; existing LTAs will always expire and new LTAs will always begin under a different pricing regime at varying points in time; so that's one piece. The second piece is rolling out next-generation platforms. We're quite thrilled that in Q4 we began shipping up to 40 Terabyte ePMR drives. As of now, we remain very much on track to reach 50% of nearline bids on that form by third quarter of year '27. That, in turn, will strengthen our capacity to move more higher capacity drives out to the market and open up some extra pricing upside too, and clearly, we keep pursuing operational efficiencies the way we always have. So in brief, quarter-to-quarter swings will always be there, yet we're quite confident about what I just listed — the pricing structures we've established, the arrival of our higher cap drives later this year, and our steady operational efficiencies — and that all points to more robust Exabyte growth, sustained top line growth, and clearly continued margin expansion.

Ambrish Srivastava executive
#7

Do you have a quick follow-up, CJ, which wouldn't be a multipart question, please.

CJ Muse analyst
#8

Sure, absolutely. At Flash Memory Summit, Marvel ran this session discussing how HDDs are economically indispensable in the AI data center; so with that as the background, I'm curious how your visibility has improved? And in what ways do you see it getting even better once you begin signing contracts that renew.

Tiang Yew Tan executive
#9

Yes. Appreciate the question,. Well, I believe our visibility is strong and getting stronger, and to a significant extent it's customer driven too. When we last gave an update on LTAs, we mentioned holding one LTA with a large customer extending all the way to calendar year '29; however, we are very much in the thick of talks with customers to set up LTAs for calendar year '29, '30, and '31 as well. So visibility stays very strong. customer-driven demand for LTAs stretching all the way out to '31 stays very strong; so we keep working on them. I think we have a very good line of sight when it comes to the exabyte demand. What we're working through with them is more about the pricing commercial construct of what those LTAs would look like going forward.

Amit Daryanani analyst
#10

So, perhaps the initial point is that exabyte shipments rose roughly 21%, as I recall, compared with a year earlier, reaching 21%. That falls short of the 30% exabyte increase you people have recorded over the past several quarters. Could you simply address whether this might be a pause tied to timing or supply? Or does it instead signal a more normalized growth rate? And what's your thinking on exabyte growth across fiscal '27?

Tiang Yew Tan executive
#11

Yeah. What I'd say is that we've provided guidance around an exabyte figure. Demand, in our view, is expanding beyond 25% — 25% plus — moving forward; so our road map is unquestionably built to sustain that. Once again, as I noted, comparable to margins, we'll see quarter-to-quarter swings when it comes to exabyte growth rates, based on which products we ship. And once we get into the year's second half, when we truly scale up 40-terabyte shipments — as I mentioned, that will exceed 50% of nearline exabyte ship — we expect the exabyte growth rate to pick up. Then, clearly, HAMR arrives in the year's second half and in calendar '27 too, at a 44-terabyte level.

Kris Sennesael executive
#12

Right, and I'd also mention that during Q4, our exabyte shipments rose 22% compared with the same period a year earlier, and if you consider the entire fiscal year '26, exabytes climbed 25% versus the prior year.

Ambrish Srivastava executive
#13

Do you have a follow-up, Amit? .

Amit Daryanani analyst
#14

Turning to pricing, the price per unit rose quite sharply — up in the high teens compared with a year ago — and Kris, I believe you addressed this. Could you explain how you expect pricing to develop going forward? And perhaps more precisely, the new LTAs you're signing — do they make this sort of high-teens pricing possible? Or might there be additional upside ahead?

Tiang Yew Tan executive
#15

Well, I'd say we gained from pricing gains across the board, in our nearline offerings and clearly in the client consumer segment too. When it comes to nearline, particularly with our big customers, we've consistently emphasized predictable pricing. That approach is what we keep following. It gives them visibility so they can make architectural commitments for the long term with the proper economics supporting them. So — yet even so, as we bring in drives with greater capacity, we deliver improved TCO value to them, which creates a chance for us to raise pricing, though you'll find that happens on a highly predictable cadence moving forward. When it comes to areas outside nearline, especially client and consumer, we found a bigger chance to raise pricing, mostly because of the pricing frameworks for competing products that are flash based; that led to a somewhat stronger pricing climate. Kris, would you like to — anything else you'd like to add?

Kris Sennesael executive
#16

No, I think you summarized it well. .

Aaron Rakers analyst
#17

So, picking up where we left off with those earlier questions, it appears to me there's a product transition happening with the 40 terabyte ePMR drive. When I look at how your costs are structured, it seems the plus per terabyte came in more or less flat, perhaps slightly lower on a sequential basis; so what I'd like to know is, as this product transition plays out, what should our expectations be for cost per terabyte, or cost of goods sold per terabyte, as we go through '27?

Kris Sennesael executive
#18

Yeah. Erin, in Q4 of fiscal '26 — the quarter we just reported — cost per terabyte declined roughly 8% versus the same period a year earlier. As we've said previously, over the long term cost per terabyte should decline somewhere around 10% year-over-year; and the primary factor, which Irving already went through, is that shifting toward higher capacity drives with better aerial density delivers a meaningful drop in cost per terabyte; and as we deliver in the near, mid and long-term, deliver against our technology road map, deliver against our product road map, bring out the next-generation ePMR and bring our HAMR products to market, we'll keep seeing cost per terabyte decline over time. follow-up.

Ambrish Srivastava executive
#19

Aaron, anything else you'd like to ask?

Aaron Rakers analyst
#20

Yes, I have one. Kris, perhaps returning to — you've accomplished quite a bit regarding the balance sheet, correct? There's still $1.1 billion remaining on that debt, roughly. And you're holding $1.6 billion in cash. I'm just wondering, what's your thinking on the capital structure going forward from here? Is it basically 100% free cash flow return? Would you consider adding more debt to the balance sheet? I'm just curious whether there are any fresh thoughts on that.

Kris Sennesael executive
#21

Yeah. Our strategy hasn't changed, nor has our dedication to giving free cash flow back to shareholders, just as we've consistently done across the last 4 or 5 quarters via a mix of dividend payments and share repurchases. So both of those 2 programs have our complete commitment.

Wamsi Mohan analyst
#22

I wanted to ask, Irving, whether you might unpack a bit more that remark about mix you gave when responding to an earlier question regarding the 22% exabyte increase, since the expectation before was that those rooms would ship additional exabytes during the quarter, and the slowdown looks quite pronounced. So I hoped you might elaborate somewhat on what you said about mix. Simultaneously, you clearly anticipate an uptick based on what you've said, both at a macro level and regarding your product cadence. So is there anything you can share about what specifically is shifting within that mix that perhaps produced the weaker growth this quarter? I realize 22% is quite robust, yet it's certainly a slowdown, and then with further acceleration, is there some near-term ceiling on that prior to when you might begin ramping HARM?

Tiang Yew Tan executive
#23

Yes; thanks for asking, Wamsi. Much of it, as you're well aware, involves our large customers. And so various customers adopt various kinds of technology; so should a specific customer in a specific quarter take somewhat more CMR products, clearly that means we send fewer bits out into the market relative to the quantity of units we have, correct? So in another quarter when we have customers that lean more ultra-smart heavy and with an identical quantity of units, we can supply far more exabytes out into the market. So variations from quarter to quarter will always exist, and their purchasing isn't done in a very linear fashion. It's somewhat lumpy. So you'll observe those quarter-to-quarter fluctuations within that; yet over a medium- to long-term trajectory, we remain quite confident regarding continued exabyte growth, the capacity to achieve the percentage growth trajectory we provided, both concerning the 40 Terabyte ePMR that we're ramping very aggressively. Then clearly, we have the 44-Terabyte HAMR launching in the first half of calendar '27, which we're ramping, and clearly, we've stated in our road map, we'll have 50-Terabyte products launching toward the second half of calendar '27.

Wamsi Mohan analyst
#24

And if I may also ask a quick follow-up. Looking at your sequential gross margins versus your guidance — sorry to keep harping on this — but compared with last year, you're still pointing to fairly robust incremental gains. I understand that. But is there something particular in there, once again, in terms of mix and the pricing trend, that shifts quarter over quarter? Or is that pricing trend something that ought to stick around, and it's really more a matter of mix that's perhaps limiting any additional upside to gross margins?

Kris Sennesael executive
#25

So yes, once again, on a quarter-to-quarter basis, there are going to be valuations. That's the reason I don't get too hung up on my sequential incremental gross margins. My focus is more on my incremental gross margins compared with a year ago; and to give that some numerical context, correct, across fiscal '25, on a full year basis, our incremental gross margin came in at 60%, and during fiscal '26 — the fiscal year we just closed, incremental gross margins reached 75% versus a year ago, finishing Q4 at an incremental gross margin versus the prior year of 84%, 85%. And baked into the guide for Q1 is a year-over-year incremental gross margin of 80%, 81%. So I'd say we're performing very well when it comes to generating incremental gross margins, and pushing gross margins still higher. I hold strong conviction that we'll keep enhancing gross margins across many quarters ahead given everything we've laid out. Demand is extremely strong; we're transitioning to higher capacity drives that deliver greater value to our customers, and that lets us raise our price per terabyte while simultaneously lowering our cost per terabyte. And I'd call that an excellent formula for additional gross margin gains.

Sreekrishnan Sankarnarayanan analyst
#26

First off, Irving — it's clear your pricing is getting better at a solid clip. You brought up how customers are looking for LTAs reaching into 2029, '30, and '31. So I want to ask: given that 90% of your bids land in cloud, and most of those go toward LTAs that stretch further out, the pace of price hikes you're seeing — moving from high single digits up to high teens last quarter and likely 20% in the September quarter — would that ease off since most of it sits under LTA? Or do you figure there's still room for that pace of price increases to keep climbing?

Tiang Yew Tan executive
#27

Yes. Appreciate the question. Like I said, when it comes to the cloud business, as you noted, a fairly large portion of it falls under LTA. Of course, not all LTAs begin at the same moment; so across the fiscal year we'll still see price adjustments from LTAs, depending on when those contracts activate with the new pricing, and that's clearly shaped by some of the new platforms we've brought to market. And let me restate our overall pricing approach there — ensuring it stays predictable, while giving our customer plenty of visibility. The area where we've found somewhat greater ability to push faster price increases has been non-nearline. If we jump ahead to what we're examining in our discussions with customers regarding '29, '30, '31, we've — we're sorting through the particular commercial constructs of the LTAs, yet we have solid line of sight and visibility into the volume requirements of our customers, which has somewhat reinforced our conviction around that 25% plus exabyte growth. What we're sorting through is the pricing regime for how we'd bring that about in those years ahead; in short, I believe there remains plenty of opportunity for us to deliver more value to our customers, deliver better TCO, and capture that benefit through pricing, just as we've always said we would.

Ambrish Srivastava executive
#28

Do you have a follow-up, Krish? .

Sreekrishnan Sankarnarayanan analyst
#29

One more quick question for Kris. You said something about cost per terabyte declines being maybe in the high single digits at present, and that it could reach roughly 10% reductions. Does that assume hammer qualification has already happened? Or do you figure that's achievable within the coming few quarters?

Kris Sennesael executive
#30

I've mentioned this before and I'll say it again today: over the mid to long term, cost per terabyte comes down by around 10% year-over-year. Looking at the past couple of quarters, that's what we've been delivering, and I have strong confidence once more that over the longer horizon we can deliver that as well. Sure, there will always be some quarter-to-quarter fluctuations, but over the mid to longer term — and all of this stems from our technology road map and our product transitions toward greater aerial density and higher cost drives, which will push cost per terabyte downward.

Mike Cadiz analyst
#31

Mike Cadiz here, covering Asia merchant at Citi. So, to start with, considering the LTA discussions you've had, and naturally the forward visibility those give you — could you speak to what kinds of workloads your customers intend to support, whether agentic or physical, and so on, and whether those workloads shift over time, from LTAs that mostly cover the nearer term, 2027, to LTAs and talks reaching '28, '29, and even '30? I just wanted to understand how workloads evolve from where you sit.

Tiang Yew Tan executive
#32

Certainly. I'd say most of the workloads will remain the same. There are a few growth drivers we're seeing, and customers are definitely giving us visibility. First — and this is something people tend to underestimate — core cloud services keep expanding. That's especially true for video-heavy applications; take the collaboration tools so many of us rely on daily. A great many of these video meetings we join are placing rising demands, which pushes a lot more storage requirements onto core cloud services. And clearly, within AI, over the past 24, 36 months, model training and development have been the main drivers. What we're observing now, and expect ahead, is growth coming from inference, Agentic AI; and in fact, we're beginning to see the early stages of growth powered by physical AI too. We're working with numerous large enterprise customers active in the physical AI arena, and we have solid visibility into their growth trajectories as well. So growth is certainly arriving from those vectors I just mentioned. Beyond that, clearly, with innovations we've introduced, such as the high bandwidth drives now being sampled with five customers, that unlocks the ability to deliver hard drive capabilities too, with the superior economics it brings in higher bandwidth foot loads. So we view that as another future growth driver we're quite excited about, one customers are engaging with very closely, Aon.

Ambrish Srivastava executive
#33

Can we go to the next question, please.

Erik Woodring analyst
#34

During the prior quarter's call, you said you were seeing some interest in nearline drives for your systems coming from sovereign and neocloud data centers. Could you perhaps expand on that a little, given that 3 months have passed since you made that remark? Specifically, what exactly are they purchasing? To what degree does price matter to them? What size are they? How large might they eventually become, and so on? Then I have one quick follow-up.

Tiang Yew Tan executive
#35

Of course, Erik. That's an excellent question. My view is that we treated it as an opportunity, and that opportunity is now becoming real; once more, I'd point to 3 areas; we clearly observe rising demand from neo clouds, and in fact even frontier AI labs are approaching us for supply; and as I mentioned in response to the earlier question, a physical AI company. So we've witnessed quite robust demand from an autonomous vehicle company seeking additional storage as they scale up their autonomous vehicle capabilities, we're seeing greater sovereign demand from new clouds, from AI frontier labs, and we especially notice the economic advantages of hard drives along with the growing storage requirements their business models are generating too. . Regarding pricing, I'd say they find themselves in the very tax supply climate we currently have, which gives us a chance for greater pricing leverage as well.

Ambrish Srivastava executive
#36

And Erik, you said you had a follow-up?

Erik Woodring analyst
#37

Yeah. Just one thing real fast. Irving, I wanted to get some clarity on what you said about pricing. Do I have this right — some of your contracted pricing includes step-ups in price over time? And since demand has gotten stronger and you're now providing extra value to your customers through those higher-capacity drives, should we think of those price step-ups as speeding up? Or am I reading too much into it?

Tiang Yew Tan executive
#38

Yeah, Erik, you might be reading a little too much into it. What I'd say is that the LTAs clearly carry a base price tied to a base volume. And when we've been able to give our customers upside, that gets handled through various pricing structures. So that's basically what we've always described as the general way the LTAs are set up commercially.

Ben Reitzes analyst
#39

I'd like to ask, considering the product ramp and the better demand picture, should we be thinking about the first quarter of fiscal as the bottom for both sequential revenue growth and incremental margins, with things speeding up as the year goes on, with regard to FY '27.

Kris Sennesael executive
#40

Yes, Ben, as you're aware, our guidance covers just one quarter at a time. And once more, I believe this quarter we're guiding toward some sequential growth along with year-over-year growth and some solid incremental — robust incremental gross margins compared with a year ago. Again, we have — with the strong demand environment we're seeing, with our excellent execution on technology and the product road map, with the fact that we're at the start of a ramp of the launch of our new generation of ePMR reaching 40 terabytes and then followed by our HAMR launch, we hold high conviction that we can keep shipping many more exabytes at improved pricing and bring the cost down over time, which will produce continued strong revenue growth along with additional gross margin improvement. That's all I'll say.

Ambrish Srivastava executive
#41

Ben, anything else you'd like to ask?

Ben Reitzes analyst
#42

Yeah. I was just hoping you might shed a bit more light on physical AI. What I think could help investors is that there's an enormous amount of training video, plus synthetic video too. And you're bringing it up more prominently now than you did, as I recall, in earlier quarters. So I was just curious, did anything shift? And how do you expect your tier to be affected by what will likely be a surge in both synthetic and video data used to train these systems?

Tiang Yew Tan executive
#43

Appreciate that question, Ben. Yes, we've called it out somewhat more this quarter, mainly because our visibility into this space is getting much better too. I gave an example involving an autonomous vehicle company where, I'd say, the rise in exabyte demand coming from them for this calendar year — or rather, calendar year '27 — has gone up by a multiple, correct? So we are certainly witnessing that. And when you extrapolate across how many players exist in this area, you carry that over to human, or you carry that over to industrial automation systems too. And as I mentioned during the prepared remarks, in a lot of these situations, they're producing fresh data from the tools and vehicles they gather data from, yet that remains insufficient for training their models too. So they're additionally taking the data already on hand, feeding it into AI tools to create synthetic data for further model training and to power reinforcement learning, and that calls for far more storage data. as well. So the cycle we observed in how model training, development, and inference came about will carry over into what's taking place in physical , but it's turning very real at this point and we view that as a growth driver for '27 and past that too.

Karl Ackerman analyst
#44

If I could, I have a couple of questions. To start with, what's the right way to look at the gross margin difference right now between nearline and non-nearline? The reason I'm asking is that since there's more room to push prices up in the non-nearline HDD market, my assumption would be that the gap narrows. And then, over the coming 12 months, what portion of these LTAs get repriced for nearline?

Kris Sennesael executive
#45

Regarding gross margin for nearline versus non-nearline, which takes in consumer, client, and all of that, at present everything sits in roughly the same range, around the 5 ZIP code. So these days there isn't much of a distinction left there.

Karl Ackerman analyst
#46

Irving, it's clear that demand from hyperscalers is strong. At this point, how big is the on-prem enterprise segment, and can you also get long-term agreements signed with those customers?

Tiang Yew Tan executive
#47

Yes. I appreciate that question, Karl. Once again, I'd say most of the new line bitter shipments are headed to hyperscale clients. More and more, we're actually seeing greater demand coming from enterprise OEMs in storage, particularly as they move toward more hybrid-based storage options, whereas not long ago they were considering systems that could be entirely 100% flash arrays. A swing back to hybrid systems is generating additional demand for us. So we're also incorporating forward-looking requirements into the projections we're developing regarding supply, which is giving us improved visibility too. And clearly, our engineering groups are collaborating with them so they can take on our Ultra products. plus next-generation ePMR HAMR products, which will provide them high capacity drives and, as a result, more exabytes more quickly as well.

Ambrish Srivastava executive
#48

And operator, we'll go to the last caller, please. .

Ananda Baruah analyst
#49

Kris, Irving here — I was just thinking, with everything happening over in China right now, what would be a helpful framework for looking at the opportunity there in hyperscale and the AI labs?

Tiang Yew Tan executive
#50

Sure. Thanks for raising that, Ananda. To me, Asia is clearly a really thrilling opening for our business. We've obviously been picking up plenty of signals and seeing a great deal about how AI frontier lab models are spreading over there, and that they're open source. Our take is that this will genuinely benefit the wider sector, since it should widen availability of various models, which will help AI spread much further with better economics. So based on whatever the use case happens to be, you're able to apply differing economic models to generate growth and value across the various applications. Yet fundamentally, regardless of whether we're talking about the conventional frontier models familiar from the U.S. that carry a premium, or those newer open source models that are more economical, the basic need they'll have is for data to underpin training of these models; and these models will produce far more data, which means storage demand lands with us. That's why we see it as quite favorable, because even if compute perhaps delivers better efficiency when it comes to compute and memory resources, the call for storage will simply keep compounding and expanding; so for demand ahead, our view is quite positive.

Ambrish Srivastava executive
#51

One more quick question on hand?

Ananda Baruah analyst
#52

Yeah, a fast one; picking up right where that left off, how the relationship works between ramping up aerial density by a meaningful amount, weighed against the fact that we're only at the beginning of inference, which is truly the killer application for hard drives; so in what way do you expect those two factors to shape the supply-demand imbalance over the next few years?

Tiang Yew Tan executive
#53

So, our view is that inferencing growth will be a positive for demand when it comes to higher capacity drives, and of course influencing will actually create still more data. That data will need to be kept, both for reinforcement learning purposes and because retaining the context of generated data will cost less than running it again through compute and memory resources. Thus, in our view, inference in the genetic growth will in fact push even greater storage demand, and the optimal route to that storage, with better TCO and greater speed, is delivering high capacity drives, which is certainly something we are striving to provide our customers.

Kris Sennesael executive
#54

Thank you. Now, at WD, when we think about the coming fiscal year and what lies past it, we do that with a purpose that's very clear to us. The people who buy from us depend on us to provide storage infrastructure at a scale and a speed that keeps up with how quickly AI is being adopted and how fast it's moving, and we embrace that challenge wholeheartedly. We have faith that our technology is out in front, that the way we execute is disciplined, and that everyone across the WD team stays concentrated on getting results. So I truly want to use this moment to express gratitude to all of WD's team for a year that was outstanding, to the investors and the analysts who are with us today, and thank you for the time you've given and for the confidence you keep placing in Western Digital.

Operator operator
#55

That brings us to the end of this call. We appreciate your joining us for today's presentation. You're welcome to disconnect at this time.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Western Digital Corporation transcript - plus 132,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.

Get an API key View API docs →

For developers and AI pipelines

Programmatic access to Western Digital Corporation earnings transcripts and 132,000+ others is available through the EarningsAPI REST API and the hosted MCP server. Quarterly plans from $162 - full transcripts, speaker segments, full-text search, and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.