Western Digital Corporation (WDC) Earnings Call Transcript & Summary

Good afternoon. Hi, I'm Tim Arcuri. I'm the semi and semi equipment analyst here at UBS, and we're very pleased to have Western Digital. And with WD, we have Irving Tan, who is the CEO; and we have Kris Sennesael, who is the CFO. So thank you to you both.

Pleasure.

Yes.

And I'll turn it over to Kris, who wants to read a statement.

Yes. And thanks, Tim, for having us at this wonderful conference. So before we begin, we will be making forward-looking statements in today's discussions based on management's current assumptions and expectations, including with respect to our product portfolio, business plans and performance, market trends and dynamics and future financial results. These forward-looking statements are subject to risks and uncertainties. So please refer to our most recent financial report on Form 10-K and other filings with the SEC for more information on the risks and uncertainties that could cause actual results to differ materially from expectations. We will also be making reference to non-GAAP financials and a reconciliation of our GAAP to non-GAAP results can be found on the Investor Relations section. So with that, turn it back to Tim.

Perfect. Great. So let's just start. Like your biggest competitor, you're enjoying a very favorable pricing environment along with very tight supply. Can you just talk about what's driving that? And how much of it is cyclical? And how much of it do you think is secular?

Yes. I think if you go back to the Investor Day that we had back in February, we laid out 2 cases, a base case where we saw Exabytes growing at a CAGR of 15% and that was primarily driven by what we saw would be just fundamentally cloud-based growth. And we also laid out an upside case where if what we saw the potential of AI would materialize, we would see a CAGR of 23% growth. What we've been seeing more recently is probably growth rates around in the mid-20s. And we think that trajectory will probably continue into the 12- to 24-month period. If you just take our last quarter results, Exabytes grew at 30% year-on-year as well. So ongoing strong growth in the cloud, the transition to more AI intermodal LLMs is driving a lot of growth demand as well. So we see that trajectory continuing, and that's translating into longer-term contracts that we are seeing from our customers. We have firm purchase orders from 5 of our largest customers for all of calendar year '26. One of those 5 customers have given us purchase orders for all of calendar year '27.

All of calendar year '27?

Correct.

Can you -- just in light of that, can you talk about your pricing strategy in this environment? Do you see -- if you keep increasing pricing, is there a point where you see some elasticity where customers are like pushing back on paying the higher prices?

Sure. Maybe I'll touch about philosophy towards pricing, and I'll talk about -- let Kris share a bit more about how we see that translated into the financials. Our focus around pricing is to ensure there's a fair value exchange with our customers to always ensure we're delivering better TCO value to them, and that's driven by 2 key areas: one, continuing to deliver higher capacity drives to them, so they get better TCO through better rack density, lower energy consumption, better real estate cost. And then we're also looking at innovations that drive better throughput in terms of those drives. So as we deliver a better TCO value, we're able to share in that TCO value that translates to a better pricing environment for us. So that's a fundamental philosophy, especially as we structurally have become a data center company. 90% of our business today is within the data center, concentrated on a few set of very large customers. And our approach is to really form long-term strategic partnerships with them to ensure we have fair value exchange, primarily through the TCO value that we deliver. And maybe, Kris, do you want to touch on the pricing?

Yes. And so historically, when the business was more cyclical, we have seen ASP erosion on a dollar per terabyte basis of approximately 7%. Although now when the business is more secular and we see a long-term secular tailwind, that has changed a lot. And so currently, I described the pricing environment as being stable, which means kind of flattish to slightly up low single digits on a year-over-year basis on a dollar per terabyte.

And you think that, that's going to continue, Kris.

Yes. Again, based on the long-term agreements that we have in place or are negotiating as well as the POs that we have in place covering '26 and '27, I expect the pricing environment to remain stable.

And can we just talk about the visibility and the way that your customers are actually placing orders. They're -- obviously, you're booked out through '26, it sounds like, and one through '27. There -- are they getting to the point or is there a risk that they get to the point where they force your hand to start to expand unit capacity because they say, I just don't want to wait 60 weeks for a drive?

Yes. I mean our focus is to really meet the demand that we have, which is Exabyte focus through improvements in the following areas: one, continuing to accelerate the road map that we have to continue to deliver higher capacity drives, both in terms of HAMR technology and our ePMR capabilities as well. Second, to up-level our customers to higher capacity drives, right? If you take the last quarter, the average capacity for all our nearline customers was 22 terabytes per drive. We're shipping the highest capacity is 32. So there's a lot more room to move up capacity points as well. And third is to transition more of our customers to our UltraSMR technology. And the benefit of that to them is UltraSMR gives them a 20% capacity uplift versus the standard CMR drive and 10% capacity uplift over a standard SMR drive. So for those reasons, we feel we can meet the exabyte demand requirements in the fastest way possible by transitioning them up to higher cap drives, accelerating the road map of high-capacity drives we are delivering and getting more of them to adopt UltraSMR. Last quarter, 50% of the nearline bits we shipped were already on UltraSMR. And we see that growing as more customers adopt our UltraSMR going forward.

And I think you mentioned that there could be more to your ePMR road map even beyond 36T. What is so hard about scaling beyond 36T? And how do you think about stretching that road map to 40T or beyond and marrying that with -- you want to now pull HAMR in a little bit. So how do you think about that?

Yes. So we have shared that we are pulling forward our HAMR qualification to start in the first half of calendar '26 with one customer, and then we look to extend that with two more customers in the second half of the year. So through the course of calendar year '26, we'll be qualifying 3 customers on HAMR. The current road map for that is to introduce a 36-terabyte CMR and 44-terabyte UltraSMR HAMR drive. We also have indicated we've pulled forward the qualification of our next iteration of ePMR to Q1 of calendar year '26. We will be sending an invitation to all of you for an Innovation Day that we'll be having in New York on February 3. The invitation will go out latter part of this week where we'll be highlighting all the latest innovations we are going to bring to the market and how that's going to translate into the new road map for both HAMR and PMR and also the performance improvements that we look to deliver for our drives going forward. So stay tuned for that.

So do you think you'll have a seamless transition, your 44T Ultra SMR in mid-'26 will sort of lead right into HAMR?

Yes. I mean our goal is really to focus on not only having the highest capacity drives, but to be able to deliver that at scale. There's been a lot of talk about we have a 40-terabyte drive or 44-terabyte drive. But ultimately, what our customers want are exabytes, right? If you look at those contracts that we've talked about just recently, those contracts are exabyte based. They're technology agnostic. They don't have any mention of units. Customers want -- they're signing up for a certain amount of exabytes. And so the ability to deliver exabytes at scale is what we're really focusing on. And if you look at our current latest generation of ePMR drive, last quarter, we shipped 2.2 million units, which totaled a total of 70 exabytes. This quarter, that same platform will ship well north of 3 million units. So that's a really great example of what our customers appreciate from us; fast qualification of our ePMR products, the ability to ramp up very quickly and deliver exabytes at scale. And we anticipate that will be the case with our next iteration of ePMR. And as we transition to HAMR, we want to make sure that customers have a seamless transition to HAMR that will give us 50, 60 up to 100 terabytes in the future.

And what went into the decision to pull in qualification for HAMR? Was it more that you saw your competitor had gone through the trials and tribulations and you saw that the customers were sort of adopting HAMR. And so you thought, "Oh, I better pull that in?"

No, it's really driven by the progress that we saw that we had been making. And in the prior earnings calls, I've highlighted that we were pleased and ahead of the internal milestones that we had set for ourselves. We were definitely comfortable with the aerial density improvements that we were seeing in our HAMR development as a result of us being ahead of those milestones and the confidence that we had, we felt it was the right time to pull forward that qualification. As I mentioned, we are comfortable with the aerial density improvements we can deliver through HAMR. The focus on our qualification is really twofold. It's ensuring we can get the right reliability for HAMR because our customers are telling us what they're concerned about is what's going to happen 1 year, 2 years, 3 years down the road? Will there be systemic failures in the HAMR fleet? So we want to make sure we avoid that for them. And secondly, making sure we can get the yields for HAMR in production up to the same yield levels that we've been able to deliver on our ePMR portfolio, which we've shared based on the platforms has been in the high 80s to low 90s percent range.

So your average -- and you mentioned this before, but your average capacity per drive is still well below what these higher densities are. Why is that number so far below? And what applications are still on these older smaller drive, less dense drives? And what effort are you making to migrate those to higher capacities?

Yes. Part of it is because there's still a material number of customers on CMR, right? So as they move to UltraSMR, they get a capacity uplift. In order for them to transition from CMR to UltraSMR, there's a degree of software work that they need to make on the application side of the house. There's always been a bit of contention to get their application colleagues to free up resources to do some of the rewrites. But in the environment that we are seeing where supply is being challenged because of the demand that we're seeing driven by cloud and AI, there's been a very concerted effort by many of our customers to get their application teams to rewrite software to take advantage of UltraSMR. So we've seen very fast adoption and qualification of it currently. And then also some of the -- as they refresh their fleets, they're moving very quickly up to capacity points as well. We're also doing work in our Platforms business to eliminate some of the heavy lift that they potentially might have to do on the conversion from CMR to UltraSMR by using our platforms to be able to convert UltraSMR to CMR recording technology for them as well.

And can you just sort of double-click on the HAMR road map for us? I know you're pulling in the qualification. But sort of once you do qualify HAMR, where does HAMR go from there?

Yes. So we're looking to qualify HAMR through the course of calendar year '26. The current road map is we will ramp that in the first half of calendar year '27. And in the current road map, the capacity points will be 36 terabytes on CMR and 44 terabytes on UltraSMR.

And customers don't care whether it's on HAMR or it's on...

No. In fact, all those POs and contracts that we have '26, '27, and we're in discussions with customers talking about '27-'28 time frame contracts, they are exabyte-based, they are technology agnostic. What they want is highest velocity of volume of drives at the highest capacity that's the most reliable. And these contracts are not only volume, but they come with fixed pricing.

Is that right?

There's a degree of fixed pricing associated with it as well. Yes.

Degree of fixed pricing. So it's not -- you will -- that will take...

There's a degree of base price -- base volume at fixed price and there's incremental upside in pricing.

Got it. Okay. And can you talk about your willingness to expand unit capacity? In other words, are your largest customers offering to co-invest in your CapEx to get you to expand unit capacity? Or are they happy with you just moving density higher?

We've been very clear that we are not expanding unit capacity. None of our customers have had the discussion with us about investing in unit capacity. And the company is in a very different position. We have a very healthy balance sheet. We're generating a lot of cash flow. We can make the appropriate investment decisions that we need to. We feel that with the areal density improvements that we are going to make with the road map, with the ability to upscale customers to higher drive capacity points and through UltraSMR adoption, we're going to be able to meet their demand requirements without adding any unit capacity.

And can you just talk about the interplay between HDD and SSD and whether these extended lead times and is this pushing any demand in nearline over to SSD?

We don't see that. I mean the good news is right now with the ongoing growth of the cloud, with AI driving a lot the increased data generation and demand for storage, we are seeing an environment where in a way, it's lifting all boats, right? You're seeing demand for SSDs go up. You're seeing demand for HDDs go up. Even [ good old tape ] is seeing demand increase as the value of data increases and more data is getting stored. What we also see is data center architectures are pretty much set with the chassis that they have for a period of time, forward-looking 2, 3 years out. So you don't see any rapid changes between the ratio of HDDs and SSDs. So the ratio is going to be -- roughly about 75% to 80% of data is going to continue to be stored on HDDs, 10% to 15% on SSDs and then the remainder on flash. And that ratios may adjust a little bit quarter-to-quarter, but by and large, they will stay the same. HDDs continue to have very strong TCO benefit, 6x delta in terms of acquisition cost, 3.6x delta in terms of total cost of ownership. We probably think that gap has widened as SSD prices have gone up as well. And with the announcements and some of the innovations we'll share in February, we'll also share some of the performance improvements we are making on throughput that will address some of the gaps that we have in terms of throughput performance that HDDs sort of lag SSDs.

Can you speak just from a strategic perspective about what the puts and takes are now that the company is back to being independent and the split happened? What are some of the puts and takes? I mean, obviously, there was a thesis to put them together in the first place. So if you disaggregate them, obviously, there are some negatives because you wouldn't have put them together in the first place, not that you did, but the company wouldn't have put them in the first place. So can you speak a little bit to how you run the company differently now versus how you ran it...

Yes. I would say, at a high level, maybe the one dissynergy that we have from splitting the company up, we would -- we did have to incur more G&A costs, right, because you could amortize that over a larger business. But I think that has been far outweighed by having 2 independent businesses that are singularly focused on the success of the HDD business and the SSD business, right? Because if you look at it, technology-wise, they're really quite different. One is electromechanical based, one is semiconductor based. It has no commonality in technology, in the manufacturing assets, in the technology, intellectual property, even in the supply base. And so having that singular focus of each company on what is the right processes to run the business by the right KPIs, the right financial model to run each business by has really given the ultimate -- it's proven that the spin has made sense. And if you look at the market cap of the 2 companies combined today, they are in the mid-$80-plus billion. So it's proven to be the right decision ultimately.

Yes. The markets help, too, but...

Yes, [ across ].

For sure. So can we just talk about there's a third competitor that people don't talk about very often, but they do have a MAMR road map and some of the hyperscalers are talking about trying to keep them alive. So can you just talk about the relative share between you and your main competitor and the third competitor? Do you see that third competitor like withering away? Or do you see the customer trying to keep them alive and maybe they become more...

Well, I treat all our peers with a lot of respect and a healthy dose of competition. I would say I think we trust in our technology. We have a 32-terabyte drive and PMR. We have HAMR qualification starting. We are introducing our next generation of PMR, which currently is slated for 36 terabyte, but probably will be slightly north of that when we bring it out in the first half of calendar year '26. I think the third player is a good competitor. We look forward to competing with them in the marketplace, and I'll leave it at that.

Great. And then on the last quarter's call, you discussed your systems integration and test lab in Rochester. How important is joint development in this lab? And what kind of partners do you work with there?

Yes. I would say it plays a very critical role in a couple of things. One, in the Rochester lab right now, we have 2 environments. Each environment is a complete replica of 2 of our largest customers' production data center environments. Everything from the racks that are in the lab to the temperature that we operate the lab in, right, to the speed at which we run the hard drives. In fact, I can't even enter those labs because they are effectively my customers' environment. So when I visited the lab recently, I had to get approval from the 2 customers to be able to enter the lab to see what they're doing. The only people that can enter it are our engineers assigned to those specific accounts. The benefit to that is we're actually testing both our new generation drives early in the process in the production environment. So we're ironing out bugs. It gives customers -- it leads to faster qualification cycles. It gives customers a lot more confidence in the reliability of the drives when we do put them at scale into production data center environments. So it's played a massive role. And there's a third customer now looking to us to replicate that dedicated environment in our facility as well.

And I would think that this is going to help to risk mitigate the HAMR ramp as well. Are you -- so are you operating HAMR drives in those facilities?

Absolutely. Yes, we've been operating HAMR drives in those facilities for over a year now.

And is that part of what led -- or was it a particular customer because of what's happening in that location to push you to qualify HAMR a bit earlier?

Yes. We've been sending HAMR drives engineering samples to those 2 facilities for well over a year. We have quarterly feedback from our customers on the performance of the drives, what enhancements they would like to see. We take that feedback on board. Every quarter, we ship them new iterations of those drives and the progress that we've been making there has given us the confidence to accelerate the qualification of our HAMR drive.

And Kris, you've expanded gross margin 400 to 500 basis points over the past 12 months, which is obviously great. How much of that comes from better pricing versus better utilization versus other factors?

Yes. I'm very pleased with the progress in gross margin expansion that we have seen even in the last couple of years, moving from the 20s into the 30s and now into the mid-40s. So great progress has been made there. Also, at the last earnings call, I've indicated that incremental gross margins, I'm very confident to see them at 50-plus percent, right? So even if we look at last quarter, the incremental gross margin was actually 75%. And in this quarter guidance, at the midpoint, it was implied at 65%. So really good progress. And so how are we driving gross margin improvement? There's mainly 3 elements to it. First of all, is pricing, and we already talked about that, right, compared to historically where we've seen ASP erosion. Now we are in a stable price environment, actually with some modest increases in ASP per terabyte. Secondly, of course, there is a mix element. And as Irving has indicated before, working and collaborating with our customers moving to higher capacity drives is a win-win for our customers as well as for us. Also moving to UltraSMR that gives us higher capacity at a very little cost because most of that is done in firmware and software. And then third, of course, we are focused on driving down the cost in our own manufacturing facilities, leveraging automation, machine learning and AI tools in our own factories, and we drive cost reductions through our supply chain as well. So when you combine all of that, good progress and still further gross margin expansion to be expected.

So where do you think gross margin can go? I mean, obviously, asymptotic gross margin, if the incremental gross margin is 50%, then the asymptotic gross margin is 50%. But do your customers care what your gross margin is? Like do they -- if it was above 50%, would they start to say, "I don't know, this whole -- I'm going to finally force your hand and make you expand unit capacity?" Or how does that all play into it?

Well, maybe a couple of things. I said 50 plus. I didn't say 50%. So that's one. Secondly, I mean our customers for the vast majority of the infrastructure, they pay 60%, 70% gross margin, right, on the GPUs, the CPUs, the memory, all of that. So I think they're used to that kind of environment. I think what fundamentally has changed is that our customers recognize how important data storage and HDD is in their overall infrastructure build-out, right? It's a key element. It might only be $20 billion out of $500 billion, but they need to have that $20 billion right. And there is a tight balance between supply and demand. There's a lot of technology innovation going on, new technology that's being ramped up, and they understand the value of that technology.

There's a point where -- I mean, we've seen this movie before where I'm just an old school semiconductor guy. And whenever you tell these super rich companies that they can't get a product for 60 weeks, they're going to do whatever they have to do to get what they need as fast as they can. So they're going to place orders and they're going to book out into '27. But when push comes to shove, they might not take those drives. And if that were the case, how do you protect against that? Do you -- like do you look at bookings? And is bookings the leading indicator where if there's this big, huge surge of bookings and then it completely falls off, you're like, okay, that's the first indicator of maybe they've just placed a bunch of orders just for the sake of getting in the backlog and they might not take the...

Yes. I mean we look at the demand signal that our customers give us as one data point as an input data into what we do to determine what supply output we generate, right? So customer demand signals is one. We look at what is the CapEx spending over the next few years. That's another data point. We are looking at data center builds, when are they coming on stream, what is the ratio of data center builds to how much storage they need. We look at also refresh cycles. Typically, customers refresh the hard drive fleets every 5 to 7 years. So we factor all those data points in and we come up with our own internal view of what we think the supply environment should be. And so that's the decision that we make. That may be sometimes slightly below what the customer expectation is, but we form our own opinion on what's required.

Right. And so currently, we don't see any inventory build in the channel or at the customer level. Most of the hard disk drives that are shipped to the customer are deployed right away.

And what -- just thinking in the past, what -- like what would be the first sign of if X happened, you would begin to worry that maybe demand is faltering? Is it -- I mean, is the first order -- what the customers are saying about their CapEx? Is it just that simple?

Yes. I mean we start to see potentially maybe data center growth starting to slow down. We've had instances where they may have -- they've had some delays and those are temporary issues. In fact, 3 quarters ago, we had a slight dip in revenue because there was a delay of one of our customers' large data centers, but that -- the rest of what they had planned in the pipeline was still on stream, so that didn't change it. But we look at all these factors as far out as we can get, and we factor that into our planning process in terms of the CapEx investments that we make.

Great. And Kris, I just wanted to ask you about share repo and free cash flow is obviously growing quite nicely. Can we expect to see a more aggressive share repo going forward? And -- or does it make sense to maybe pay down debt or substantially raise the dividend as well?

Yes. So I mean, first of all, I mean, we're experiencing very strong demand that translate into strong revenue growth, gross margins and operating margin expansion that also translate to very strong free cash flow and free cash flow margin, right? Just the last couple of quarters, we had more than 20% free cash flow margin. And here as well, free cash flow margin can move from the low 20s into the mid-20s. So we're doing -- we're executing really good there. And what do we do with the cash? Well, of course, we continue to invest in our business. As Irving indicated, we continue to invest in our technology and product road map. We continue to invest to a certain extent in our manufacturing footprint with CapEx on or about 4% to 6%. But despite all those investments, we delivered very strong free cash flow. And that free cash flow is being returned back to the shareholders through a combination of our dividend program and a buyback program. The dividend program was installed 6 months ago, and we started really low. But 6 months into it, we already increased the dividend at our last earnings call by 25% now to $0.125 per share per quarter. And so there is still plenty of room to further grow our dividend over time. In addition to that, on or about 6 months ago, we installed a share buyback program. We got a $2 billion authorization from the Board, and we've immediately started using that. In the first quarter, we used on or about $150 million. In the second quarter, we used $550 million. And actually, when you look at it in the last quarter, we returned all the free cash flow using those 2 programs back to the shareholders. And we will continue to do so in this quarter. There's no hesitation that -- even at current valuation, there's no hesitation to return the cash flow back through the combination of dividend and share buybacks.

Would you like to touch on the [ retained ] statement?

Yes. So we -- in addition to that, we have a pretty healthy balance sheet by now, right, on or about $4.7 billion, $4.8 billion of debt. We have $2 billion of cash. So net debt is down to $2.7 billion, $2.8 billion. In addition to that, we still have 7.5 million of SanDisk shares. At today's price, that's about $1.5 billion. And we've said publicly that it's our intention to monetize that stake in a debt for equity transaction before the 1-year anniversary. So that's coming up as well that will further reduce the debt on the balance sheet.

Great. Thank you for the time. We're out of time. Thank you again.

Thanks, Tim.

Thanks.