Seagate Technology Holdings plc (STX) Earnings Call Transcript & Summary

Welcome to the Seagate Technology Fiscal First Quarter 2027 Conference Call. [Operator Instructions] Please note, this event is being recorded. I would now like to turn the conference over to Shanye Hudson, Senior Vice President, Investor Relations. Please go ahead.

Thank you. Hello, everyone, and welcome to today's call. Joining me are Dave Mosley, Seagate's Chair and Chief Executive Officer; and Gianluca Romano, our Chief Financial Officer. We've posted our earnings press release and detailed supplemental information for our September quarter results on the Investors section of our website. During today's call, we'll refer to GAAP and non-GAAP measures. Non-GAAP figures are reconciled to GAAP figures in the earnings press release posted on our website and included on our Form 8-K. We've not reconciled certain non-GAAP outlook measures because material items that may impact these measures are out of our control and/or cannot be reasonably predicted. Therefore, a reconciliation to the corresponding GAAP measures is not available without unreasonable effort. Before we begin, I'd like to remind you that today's call contains forward-looking statements that reflect management's current views and assumptions based on information available to us as of today and should not be relied upon as of any subsequent date. Actual results may differ materially from those contained in or implied by these forward-looking statements as they are subject to risks and uncertainties associated with our business. To learn more about the risks, uncertainties and other factors that may affect future business results, please refer to the press release issued today and our SEC filings, including our most recent annual report on Form 10-K and quarterly report on Form 10-Q as well as the supplemental information, all of which may be found on the Investors section of our website. Following our prepared remarks, we'll open the call up for questions. With that, I'll hand the call over to you, Dave.

Thank you, Shayne, and hello, everyone. Fiscal '27 is off to an outstanding start. September quarter revenue grew by more than 60% year-over-year and non-GAAP EPS expanded by nearly 3x to a record $7.68. We also achieved record gross and operating margins, demonstrating the earnings power of our business and ongoing demand strength for data storage. Both the data center and edge IoT markets delivered double-digit sequential revenue growth, up 18% and 13%, respectively, reflecting ongoing demand and pricing strength. These trends support our expectation to deliver strong sequential revenue growth and gross margin expansion throughout fiscal 2027. Our confidence is supported by an extremely strong demand environment with visibility extending beyond 2030 and an aerial density-driven technology road map poised to capture this growth. In the September quarter, we shipped a record 198 exabytes nearline capacity into data center customers, and we expect a meaningful step-up in the December quarter, Supported by the continued ramp of the Mosaic 4 as we stage new generations of Mosaic products and ship to a broader customer base, quarterly exabyte growth may vary. At the same time, we expect our technology road map to support nearline exabyte supply growth in the mid-20% range on a compound basis. Overall, we delivered another very strong quarter, including record free cash flow generation of nearly $1.5 billion, driven by durable demand our continued technology innovation and disciplined operational execution. Additionally, we significantly strengthened the balance sheet, achieved our primary debt reduction objectives and regained investment-grade status. With those goals largely achieved, we are shifting our focus towards capital returns with share repurchases serving as our primary vehicle complemented by dividends. This month, our Board approved an increase in the quarterly dividend to $0.76 per share. Importantly, this capital allocation strategy reflects our confidence in the cash flow potential of the business as well as the multiyear demand visibility afforded by our customer agreements. Demand strength continues to be driven by data center customers, which accounted for 90% of our ship to capacity. Within data center, Cloud nearline products achieved a 12th consecutive quarter of revenue growth, while enterprise OEM revenue reached its second highest quarterly level in over a decade. Data growth is profound across a wide variety of applications. Behind these trends is a growing recognition that the value of data extends well beyond its original use. In a recent discussion with Seagate, a leading hyperscale customer observed that their customers are increasingly discovering new ways to extract value from data long after it's initially stored, strengthening the incentive to retain it and AI innovations are amplifying this dynamic. As AI adoption expands beyond model training and toward inference and agentic applications, persistent data becomes increasingly important. These systems rely on vast amounts of historical data to make informed decisions and take effective actions while constantly generating new output that must be governed, retained and reused. Estimates suggest that billions of AI agents will be deployed worldwide over the next decade, executing hundreds of billions of actions daily. We believe future storage demand will increasingly be tied to this new data life cycle and compound again as physical AI applications proliferate. As the volume of persistent data grows, storage demand can extend well beyond the initial infrastructure buildout. These positive demand dynamics play directly into the strength of hard drives which deliver the scalable, reliable and cost-efficient foundation for data retention. Hyperscale customers have long recognized these benefits and continue to pursue higher capacity storage architectures. Our engineering teams work closely with leading cloud providers to align our product road map with future data storage requirements and simplify the adoption of higher capacity drives. A recent example is Microsoft's qualification of a joint Seagate and Broadcom solution that combines our HAMR-based Mosaic drives configured with SMR technology and Broadcom's connectivity infrastructure. Microsoft already deploys HAMR drives, and this recent qualification validates the SMR enabled solution at hyperscale. Customers can benefit from additional capacity gains that SMR provides. However, broad adoption at higher drive capacities have been limited given the meaningful architectural and software changes required. This joint solution provides a blueprint for implementing SMR storage to unlock additional capacity gains on top of our HAMR-based Mosaic platform. Beyond hyperscalers and CSPs, demand is expanding among a broader set of enterprise OEM customers that are managing increasingly large data sets. We are engaging more directly with many of these customers and developing reference architectures that pair proven compute and storage hardware with flexible customer selected software. These solutions offer a path to scale persistent data capacity improve storage economics and drive platform efficiencies. The strengthening demand environment is extending our customers' planning horizons, which is reflected in the supply agreements we now have in place. The vast majority of our nearline [indiscernible] supply is now allocated under long-term supply agreements through calendar 2028 with some cloud customer agreements extending through calendar 2030. Our strategic planning discussions extend even further. We are working closely with customers to address our growing stores need across the time horizon by executing our Mosaic product road map. As we transition to HAMR-based Mosaic products, we are investing in additional tools and technologies necessary to offset longer manufacturing cycle times and maintain stable unit production. For Seagate, increasing capacity per disc is driving exabyte growth. Our build-to-order contracts established specific configuration and pricing terms for nearline shipments through calendar 2027. As we move through the next few quarters, we expect to begin finalizing configuration and pricing terms with customers for the exabyte supply that we've allocated to them for calendar 2028. Pricing continues to improve, reflecting both the execution of our value-based pricing framework and higher external component costs. We expect a stronger demand environment evident in customers' multiyear commitments to continue to support these healthy pricing dynamics. Taken together, these agreements provide customers with greater supply assurance while increasing visibility for both parties. They also reflect customer confidence in long-term demand growth and the increasingly strategic role that hard drive storage plays in the modern data ecosystem. Supporting this demand growth requires storage solutions to deliver scale, efficiency and economics. These very attributes serve as the foundation for our technology strategy. Increasing storage density within the same data center footprint is the most cost-efficient way to maximize bites per square foot as one large hyperscale customer put it, and our HAMR enabled road map is designed to deliver on that objective. Advancing aerial density is our North Star. We are scaling drive capacity through aerial density rather than relying on higher drive unit production. We believe that this is the most efficient way to help customers manage their extreme data growth and retention needs. The same approach supports profitable exit growth for Seagate. We are executing well on the product road map milestones that we've outlined. We continue to expect HAMR to represent roughly 50% of our nearline exabyte shipment run rate exiting calendar 2026. With Mosaic 4 representing roughly half of those HAMR shipments. Mosaic 4, which offered us to 44 terabytes of capacity is continuing to ramp with 2 leading CSPs. We are on track for additional qualifications to complete over the next couple of months across multiple global CSPs and OEM enterprise customers. Product development for our third generation was 5 product remains on track. These drives can deliver upwards of 50 terabytes of capacity and leverage our highly efficient tenders platform. We continue to expect customer qualification shipments to begin towards the end of calendar 2027. Reflecting on our ongoing advancements in Photonics, we recently demonstrated 8 terabytes per disc in the lab, providing further evidence that HAMR and the Mosaic platform can deliver on customers' growing capacity needs and reinforces our conviction that increasing exabyte supply through aerial density growth is the right long-term strategy for both our customers and for Seagate. Wrapping up, our results demonstrate both the durability of demand and the effectiveness of our strategy. data creation, retention and reuse are compounding demand for mass capacity storage. And based on the long-term visibility we have with customers, we believe we're still in the early days of a multiyear growth phase. Our technology road map, combined with disciplined supply management and a value-based pricing strategy and the operational leverage in the business gives us confidence in our ability to deliver strong profitability cash generation and shareholder returns. I'd like to recognize our global team for their execution in delivering a record quarter and a strong start to fiscal '20. I'd also like to thank our customers, suppliers and shareholders for their continued trust and support. Now I'll turn the call over to Gianluca.

Thank you, Dave. Seagate posted very strong results for the September quarter, exceeding the high end of our guidance range for both revenue and earnings per share, while continuing to establish new profitability record were translated into our highest ever free cash flow generation. September quarter revenue was $4.2 billion, up 17% sequentially and up 61% year-over-year. Non-GAAP gross margin came in at 57.8%, up 510 basis points sequentially. Non-GAAP operating margin increased 610 basis points sequentially to 50.7%. And non-GAAP EPS was $7.68, up 35% quarter-over-quarter and 194% year-over-year. Sustained data center demand continues to drive strong company performance. In the September quarter, we shipped a total of 120 exabytes, up 21% year-over-year with data center representing 90% of that total. Data center revenue was $3.4 billion, up 18% sequentially and 63% year-over-year and representing 81% of total revenue. Data center shipments totaled 198 exabytes, up 24% year-over-year and represented 90% of the company's overall exabyte volume. As Dave noted, we are staging Mosaic 4 products higher than anticipated customer qualification and we expect capacity shipments to meaningfully step up in the December quarter. Global Cloud customers accounted for the vast majority of data center revenue and exabyte demand. a trend that we expect to continue as we ramp Mosaic for volume. At the same time, demand across the enterprise OEM data center market has continued to strengthen. As AI application transition more toward inference and agent-based workloads, customers are seeking scalable, high-capacity storage solutions across hybrid cloud data center environments. In the September quarter, revenue from these markets increased for the sixth consecutive quarter with double-digit growth both sequentially and year-over-year. Complementing the strength in our data center business, our edge IoT markets also delivered solid revenue growth. Revenue totaled $784 million representing 19% of total revenue. This was an increase of 13% sequentially and 52% year-over-year. Moving on to the rest of the income statement. Non-GAAP gross profit increased significantly to $2.4 billion, up 28% quarter-over-quarter and more than doubling from $1.1 billion in the year ago period. Non-GAAP gross margin expanded to 57.8%, up from 52.7% in the prior period, reflecting favorable pricing execution and an improving product mix as we run Mosaic 4 products. These factors more than offset cost headwinds associated with certain externally sourced components. Looking ahead, we expect continued benefit from our value-based pricing strategy and product mix to support quarterly margin expansion for the rest of fiscal 2027. Non-GAAP operating expenses were $300 million or about 7% of revenue. The strong demand environment, combined with disciplined expense management, drove substantial operating leverage with non-GAAP operating profit up 33% sequentially to $2.1 billion, representing 50.7% of revenue. Other income and expenses, totaled $38 million in the September quarter. Following the debt reduction we completed early in October, we anticipate other income and expenses to decrease further in the December quarter to approximately $25 million, benefiting from the lower interest expense. Non-GAAP net income grew to $1.8 billion with corresponding non-GAAP EPS of $7.68 per share based on that expense of $328 million and a diluted share count of approximately 232 million shares. Turning now to cash flow and the balance sheet. Operating cash flow was $1.7 billion, and capital expenditures were $211 million for the September quarter or 5% of revenue. For fiscal 2027, we continue to expect CapEx investment we remain within our target range of 4% to 6% of revenue, as we support the ongoing transition to MR technology. Higher profitability and in-line CapEx resulted in free cash flow generation of $1.5 billion, up 32% from the prior quarter and more than tripling year-over-year. We expect cash generation to further improve throughout fiscal 2027, supported by sustained demand, operational efficiencies and CapEx discipline. During the September quarter, we retired $1.2 billion in debt, including the remaining convertible notes and approximately $1 billion in higher debt. We exited the quarter with gross debt of approximately $2.4 billion and a net leverage ratio of 0.2x. Following the close of the September quarter, we retired an additional $750 million of higher senior notes reducing principal debt to approximately $1.6 billion. As a result of our actions, we expect to be net cash positive exiting with calendar year. We have now delivered on our goal of strengthening the balance sheet and are now focused on returning more of our free cash flow generation to our shareholders. During the September quarter, we returned $568 million to shareholders through dividends and share repurchases, roughly 2x the amount of return in the June quarter. We are also increasing our quarterly dividend by approximately 3% to $0.76 per share, reflecting our confidence that strong cash generation is sustainable. We exited the September quarter with $4.4 billion remaining on our existing share repurchase authorization and expect to significantly increase share repurchases starting calendar 2020. We ended September quarter with cash and cash equivalents of $1.3 billion and total liquidity of $2.6 billion, including our undrawn revolving credit facility. Turning now to the December quarter outlook. We expect another quarter of strong sequential growth, underpinned by continued demand momentum, a significant increase in exabyte shipments and ongoing execution of our value-based pricing strategy. We expect December quarter revenue to be in a range of $4.8 billion, plus or minus $100 million which represented a 7% year-over-year improvement at the midpoint. Non-GAAP operating expenses are expected to remain flat at approximately $300 million -- based on the midpoint of our revenue guidance, non-GAAP operating margin is expected to expand towards the mid-50% range. Non-GAAP EPS is expected to be $9.20 in plus or minus $0.20 based on a tax rate of about 16% and non-GAAP diluted share count of 133 million shares. Our September quarter results and December quarter outlook reinforce our expectation for annual revenue growth to meaningfully outpace fiscal 2026. I durable demand, increasing MOSAIC adoption and healthy operating leverage are supporting ongoing revenue growth, margin expansion and strong cash generation. Together with our increasing focus on capital returns, we are well positioned to increase long-term shareholder value. Operator, let's open the call up for questions.

[Operator Instructions] The first question today is from C.J. Muse with Cantor Fitzgerald.

I guess first question regarding exabyte deceleration in the September quarter. I'm curious, given your inventory days moved up to 97 from 83 million -- was there kind of staging involved prebuilding that's supporting your outlook for the accelerated exabytes into December? And then maybe as you think about calendar you were kind of in the mid-30s range. You're reiterating kind of the mid-20s long term. Curious, given the ramp of Mosaic, whether we should be kind of in that 30%-plus range? Or do you think we will revert to that compounded number at some point in calendar '27.

CJ, thanks for the question. Yes. If you think back in the June quarter, we had a pretty appreciable step-up in the September quarter, we were relatively flat to your point. and the quarter we're in now, the December quarter will step up again. So we are on a trajectory and I think 25% is a really good number for that. we're driving product transitions, and we have to stay with them quite a ways out in time. We make sure that we lock in with our customers on what they need to when those things have to hit their dock and sync up with all the other parts that they need to build the data center. So pretty happy with the progress and still staging very aggressive aerial do and product transition technologies so that we can get to those higher exabyte levels.

Yes, C.J., we have a few customers that we qualify most like for in the next quarter. And therefore, we are preparing manufacturing for those calls and the volume that we can shift to them in the December quarter and data.

Perfect. And as a quick follow-up, Gianluca, on the operating margin guidance in the slide deck approaching or towards the mid-50s -- should we be modeling 50%? Or are you suggesting something else?

No, I think that's .

5%, sorry. 55, sorry. .

Well, we are suggesting something that is between 50 and 55. So of course, it's another improvement compared to what we have done this quarter that was already above the 50%. So we are very happy on how we are managing the operating margin in particular. We have seen continuous improvement in revenue in gross margin, very stable OpEx. So this is resulting in sequential strong improvement in operating margin, and we are approaching even higher results that we have guided for this December quarter. And as we said in prior calls, we expect to improve even more in the rest of the fiscal year.

Next question is from Erik Woodring with Morgan Stanley.

Really strong pricing, obviously, in the quarter, I'd love any commentary that you could provide there just on how to look forward on that price per terabyte metric -- but separately, also, just to address the cost per terabyte declines over the last 4 quarters, you've done something like 15% annual cost declines or year-over-year cost declines. You just did kind of mid-single digits. So -- can you maybe help us understand any temporary factors that are impacting that related to the Mosaic transition? Or when we should expect that metric to accelerate? And then a quick follow-up, please.

Yes. Thanks, Eric. It's a complicated question. So on the pricing side, I continue to believe that, that's very dependent upon demand and demand continues to be very strong, not just in quarter, but look 2, 3, 4 quarters out, people are demanding that we go higher in exabytes as well. So it's a great long-term relationship that we have with the customers. And we're able to predictably go through these product transitions and accelerate our technology the way we talked about. And I don't see that changing. -- the past years ago, we would see product swaps and things like that, people saying I don't need product in quarter. We're seeing none of that right now. Absolutely. Everything we can make above and beyond our plans as being spoken for. On the cost side, you know that we control our manufacturing efficiency of our critical technologies. And that's the largest driver of our costs. We get a lot of efficiency, especially in cost per terabyte if that's a metric as we transition platforms. But during some of the transitions, we get early cost things like scrap that go up. It doesn't mislead us from driving the transitions. We continue to drive them very hard and then we go work those things down. Right now, I see more opportunity because the future transitions look even brighter than I would have said 90 days ago. So I'm pretty excited about that.

Yes . Eric, I would say our cost declined for what we produce. So as a media is continuing to be very, very strong. So we are happy with that. The cost of the external component, as you know, compared to a year ago is now much higher. And therefore, over is at least a partial offset of our internal cost reduction from a higher external component cost -- so we have to see where the external component costs will go in the future. But what we can control is actually performing very well. Now when we have enough visibility -- and we have the time to include that cost increase for the external component into our pricing, of course, we do it. And this is also why we have seen a little bit higher cost increase -- sorry, a price increase to our customers in the current quarter. But sometimes it's not so predictable. So we need a little bit of time to adjust.

Okay. And then Dave, maybe just going back to CJ's question at the top of Q&A. I just want to make sure I understand the kind of product staging dynamics in the September quarter. . I just want to make sure there's no kind of right tail risk here in terms of any yield issues that we need to be thinking about anything in terms of data center readiness. I just want to make sure there's none of those that are impacting and this was -- this was generally just product transition, leading to the slower sequential exabyte growth as opposed to anything. Maybe that would be a little more concerning.

Yes, there's none. The big step up in the June quarter, it tapered off a little bit in the September quarter, just timing and then it will be another big step of the December quarter. And then as we get through these product transitions, we'll see exabyte growth.

The next question is from Amit Daryanani with Evercore.

I guess the first one was really, you folks reiterated the 4% to 6% of revenue for CapEx in fiscal '27. Is that a sustainable framework to support mid-20% exabyte growth longer term beyond fiscal 2017 as well, especially given the fact that you need -- you talked about longer tooling needed to offset the longer HAMR manufacturing cycles. Could you just talk about like what the right framework would be longer term. . And to the extent you need additional heads or media capacity, how do you kind of weigh between building it organically versus acquiring it?

So first order, there is no way we can acquire the HAMR technology, the leading-edge technology. So we have to do that ourselves exactly to your point. Yes, we see that 4% to 6% model being plenty for us to continue to grow according to our plans. Actually, the latest acquisition of new tools is really helping us drive aerial density because we get to buy remarkable no tools that are maybe more capable than what we're replacing in the factory. So that's cause for optimism. We are using some legacy technology from outside as on a case-by-case basis that will help us through some -- help us pivot some of our manufacturing technology to HAMR. But I definitely think that 25% is the target for our exabyte growth. We can do that predominantly through the aerial density transitions that we're driving. Some of those components get -- we've talked about this in the last couple of quarters. Some of those components get a little bit longer time in the factory. And so therefore, we have to adjust our output and accordingly, but we can do that all within the 4% to 6% window.

Got it. That's helpful. And then as HAMR ramps up, especially with Mosaic 4 becoming a bigger percent of exabyte for you folks, what's the wise to think about cost per bit declines as you get through that transition? I get that the 4% to 6% range right now isn't the optimal one. But I'm just curious as HAMR ramps up for you, what's the right cost for bit decline longer term.

Yes. As I said before, the cost that for the components that we produce, of course, is continue to decline. And when you go from 3 praised to 4 terabyte per disk major impact in terms of cost per terabyte decline. As I said before, right now, there is also an impact from external components. So you need to separate the 2 between internal efficiency and technology improvement that is driving cost reduction and the overall cost per terabyte that is, of course, also impacted by the external supply chain and external component costs. So you have seen a very good decline in the prior quarter, a little bit less in the September quarter. I would say we need to understand the various drivers for the different components because we buy a lot of different components externally. But overall, when you go even longer term, the technology is what is driving cost per terabyte and having a technology we continue to generate more terabyte per disk is exactly the way that we want to follow in order to reduce the cost per terabyte and improve our profitability.

The next question is from Krish Sankar with TD Cowen.

Dave or John, I just want to follow up on an earlier question. I understand the dynamics of the cost with the product transition. But if we just look at it from a high level, 50% of the exabytes exiting the sales going to be hammer, which is gross margin accretive unlike your peer, you're sticking to 10 platters not going to 12 or 14, which, in theory, I think you said increases cost. So I'm just wondering why will the cost down be even better going forward. Or would you consider moving beyond the 10 platters -- was there any technical reason for not doing it? And then I have a follow-up. .

I'll start at the end. There's no technical reason for not moving beyond 10 platters. I mean I think we like the platform and we get leverage, cost leverage, to your point, going from product generation to product generation, we don't have to transition the mechanical piece parts and paper new tooling and things like that. So we know the platform, we like it. We're going to stay on it. But there's no technical reason why we can't start adding platters. Actually, from a qualification perspective, it's much simpler, especially with our customers to just qualify 1 thing that they don't have to go through any new risks on except for just the technology transition. They don't have to changed the dynamics of how many heads and disks are in the drive. From my perspective, I probably should have said this on 1 of the earlier comments as well. there's not a significant cost increase in the HAMR technology from 1 generation to the other. So cost per terabyte take down, you can pretty much calculate it, if you go from 4 to 5, for example. And we are driving very, very hard right now.

Got it. Got it. And then as a quick follow-up. I understand you're maintaining your capital intensity 4% to 6%. Obviously, your peers talk about going into like 8% in the next couple of years. So obviously, the dissent third player talking about adding capacity in Japan. . So I'm just wondering, is the 4% to 6% still a good bogey to use for the next few years? Or should we worry about any potential share loss? Or is the demand so much greater than supply that it should not matter.

Well, I don't really speculate on what our competitors' plans are. I'd just say relative to ours, I think 4% to 6% is just fine. I have a lot of respect for everybody in the industry that their technology solutions may differ from ours. And you may see that over the next few years. From our perspective, demand remains really strong and from a supply side, we continue to believe that bringing meaningful HDD capacity to market is a really complex process that requires substantial investment in R&D and technology. It's an intricate supply chain balance. And the idea of just plugging in a few machines. I mean it's orders of magnitude more difficult than that. It really requires tight coordination throughout our supply chain. So most of our -- as I said in the prepared remarks, our nearline exabyte supply is already allocated through calendar year 2028. And we're having conversations much further out than that. So we have very -- a lot of confidence that our customers are going to be there, and we believe supply will remain fairly tight for our products over that period. So we're just going to keep focusing on our technology road map and try to get the products out and up and meet exactly what the customers need when they do, and I'm not really worried about anything it is.

Next question is from Asiya Merchant with Citi.

Great. You mentioned a lot about KV cash adoption contributing to demand. So given that, how should we think about unconstrained demand and the impact to exabyte demand given just the trend that you're seeing on KB cashing?

Yes. Thanks. I do think we are watching a lot of new technologies come out that are increasing the value of data very quickly. And I know a lot of people toss around the letters AI. It gets really confusing. What I would say is the data sphere is growing weeks and bounds anyway. Just look at video on the Internet or video analytics, for example, in places like our factory, it's growing very quickly. And then AI is an accelerator of that dynamic. The ability to take data from the application layer and move it down to the storage layer and then promote data from the storage layer back up. We've never seen a the systems be as efficient as they are right now with software, not in the way with compute technologies that are much faster than they were in the past. And architecturally, things like the cash plug between those 2 layers only accelerates it even faster. So very excited about that. I don't think any of us is overlooking a world where things are this efficient. Above the storage layers, and I do think that will drive a lot of future demand for us.

Yes. We are seeing demand for storage in the different part of the cloud growing very rapidly, even a little bit faster than what we were expecting. So as we were discussing before, our demand is stronger and stronger, how sales a supply demand gap is probably growing at this point. . But as we said, we are really focused on our technology transition and we will ramp more of our 40 terabyte drive in the next few quarters, and this is how we want to address demand.

Okay. And just as a follow-up, could I just even ask about physical AI? I know that's probably a little bit longer term. But are you seeing any of that demand currently translate into meaningful inflections for your exabyte.

I would say not yet directly, but we're hopeful. I mean, we've had conversations for a long time with autonomous vehicles companies. And people that are moving data around in the world. The interesting thing to us about physical AI is not only the need to store the data for training, but it's also the need to store the data to make sure we made the right decision. And that storage will probably be closer to where the data is actually created, which is the edge. So there's maybe new products that are coming on the horizon. It's still very early days in this, but I think there's a lot of innovative companies out there driving for value-add solutions out at the edge, and we're watching them carefully. I think this could be a further driver of storage that we haven't really contemplated yet.

Next question is from Mark Newman with Bernstein.