Intel Corporation (INTC) Earnings Call Transcript & Summary

Vivek Arya, from BofA's Semiconductors and Semicap Equipment Research Team, and I'm really delighted to have the team from Intel join us today, Michelle Johnston Holthaus or MJ, as most people refer to her as -- really happy that you could join us.

Thanks for having me.

And what I'll do is I'll start with the safe harbor statement that Intel has given me, then we'll go through my questions, but please feel free to raise your hand if you would like to ask a question, and I'll be sure to get you in. So the exciting statement before we begin, please note that today's discussion might contain forward-looking statements that are subject to various risks and uncertainties and may reference non-GAAP financial measures. Please refer to Intel's most recent earnings release, annual report on Form 10-K and other filings with the SEC for more information on the risk factors that could cause actual results to differ materially and additional information on our non-GAAP financial measures, including reconciliations where appropriate, to the corresponding GAAP financial measures. Okay.

That was highlight of your day.

With that out of the way, I really appreciate you joining us. Maybe let's start with the state of the union, right? Give us a sense for how the demand environment has shaped up. I know a lot of macro cross currents this year. But how is the demand environment shaped up so far versus what you guys have thought at the start of the year?

Yes, it's a little bit different. I mean we haven't talked much beyond Q2, but there's a lot of uncertainty in the marketplace, I think, especially around tariffs. And we certainly saw in Q1, customers buying ahead, trying to prepare for what tariffs were going to mean. I think as I look at Q2, things are a bit more stable, but I think everybody is kind of waiting for what's going to happen with 232, what does it mean? But overall, I would say, actually a pretty resilient market considering kind of the macroeconomics. And I do see positivity from customers where they see buying cycles, they see enterprise buying, small and medium business. It's probably the consumer segments that are the most price inelastic and the most sensitive. But I remain optimistic for the second half. Yes.

Okay. You mentioned Section 232. So I know not --that's...

Everybody's million-dollar question. Yes.

How is Intel thinking about the impact?

Yes. Well, first off, we have to understand what the impact is, but maybe the way we're dealing with it. I mean, obviously, no one knows what it's going to be, but we have a whole team that's doing a variety of scenario planning within Intel. And we're doing scenario planning with all of our customers as well as like how are they thinking about it? Where do they want to build, where are their kind of core focus areas as they react to whatever the tariffs will be. I think the other thing that's really important is we have a pretty broad network of manufacturing capabilities. So if you think about Intel 10, Intel 7, Intel 4, we can manufacture those in multiple sites, both in the United States and outside the United States. We use TSMC, so we can shift volume if need be there as well. And we have 18A is in the United States, our new process node, but we have EUV capability outside the United States. So we would have optionality there if we needed it as well. I mean, I don't think in any one scenario planning, they picked all the right lotto numbers. I guarantee nobody is going to get it perfectly right. But I think we have enough scenario planning that we can pull pieces out. I think what's probably most important to me is really understanding what customers want. Where do they want to build, where do they have optionality by SKU. And we understand all of that. So the second we understand those and we already know what our customers want to do, we really quickly can pivot from a build forecast perspective. So your guess is as good as mine in what will actually get published as the rules, but we're trying to prepare for it as best as we can.

Got it. We'll go through, right, some of the details of Q1 earnings and what Intel saw. But a big change, Lip-Bu joined as CEO. So what have you seen so far in terms of the changes in the organization from whether it's an operational or financial or kind of a strategic perspective?

Yes. I think we're all thrilled to have Lip-Bu on board. We were talking a little bit about this. Obviously, an industry icon that's done a lot of great work over the last couple of decades. And Lip-Bu has really come in and talk to us about 4 key areas. You heard some of those in the earnings call, which is really around reinvigorating the road map, making Intel relevant in this AI growth vector, bringing customers to foundry and fixing our balance sheet. And so when you think about reinvigorating the road map, he's flattened the organization. He's brought engineering to him. He's doing deep dives. The number of customer meetings that he has had in his first 8 weeks is astonishing to me. He came in with a very strong point of view, but really matching that customer point of view to what he's hearing from the engineering teams. He and Sachin has spent an enormous amount of time really mapping out what the AI strategy is going to be, where we're going to win, how do we use the IP portfolio that we have inside and who do we need to partner with to bring that to market. And you'll hear more about that externally, I would assume, from Sachin and Lip-Bu over the coming couple of months. And then a lot around foundry, like where are we on 18A? What is 18A-P? What type of whale or a large customer do we need to have with 14A to make that capital investment makes sense and really had a lot of conversations with customers that have decided not to use Intel Foundry and really trying to understand that. And the best part about those 3 segments is he's very transparent with all of us about exactly what he hears. So employees know where we're winning, where we're not, where we're falling short. And Lip-Bu is absolutely a person that wants to hear the bad news first. And so he's not looking for the glass half full. He wants to know where does he need to get in, where does he need fixing? And then number four is really how do we address the balance sheet? Where are we? How do we bring our CapEx and our OpEx spending down? How do we get more in alignment with what we see as best known methods within the industry and really drive those through for operational efficiency. So he's done a lot in his first 8 weeks. He's hyper focused. I think he knows exactly where the problems are, which is very good to see. But he's willing to listen, he's willing to learn and he's willing to roll up his sleeves. And so I think employees are very optimistic about the fact that he can help us. There's no one at Intel that thinks that we have it all right. We absolutely know where our shortcomings are. But I think sometimes it's easy to see the shortcomings and not always easy to figure out how to fix them. And so I think he's giving us some clarity for how to have those conversations and how to move the organization forward. So it's been very good. It's been a busy 8 weeks.

I can imagine. Is it fair to think that the focus has shifted from market share to first fixing the balance sheet and the profitability. Is that a fair representation? Or it's not as mutually exclusive as I'm making...

You don't think it's quite as mutually exclusive. I mean I think he knows that we need volume and revenue, obviously, to continue on the journey. But I do think in 2025, we try to be a bit more pragmatic about do we have to do all the deals that we're doing? Obviously, if you looked at our customer deal pipeline in Q1, it was significantly lower than it has been in past quarters. And so I think he's trying to look at it and say, hey, where do we have products that hunt and we should get paid for those? And then where maybe do we have some shortcomings where we have to use ASP to drive market segment share. But he's laser-focused on the fact that we need to get our gross margins back up above 50%. And so we need to be building products that allow us to do that because they fit the right competitive landscape and requirements of our customers, but we also have the right cost structure in place, and it really requires us to do both. And so I think it's going to constantly be a little bit of a tug of war in the short term because it takes a while to make product changes.

Right. As someone who is responsible for ensuring the success on the product side, are you happy with where the manufacturing side is right now? And how much flexibility, MJ, do you have in saying, look, if internal manufacturing is whatever, 70%, right, what if you make it 50% or 60%? So how much flexibility do you have in in-sourcing versus outsourcing to make your business a success?

Yes, it's a very popular question today. Here's how I look at it. At the end of the day, customers want great products. So I don't do anyone any good by building the wrong product on the wrong node. And so I try to be very pragmatic and say, how do I deliver a world-class product? I do it by SKU. I look at the SKU, the market need, the cost structure, what's it going to take to win and I pick the node that way. And we've done that consistently in our client product group, and I'm going to apply that across all the products. And so obviously, I'm very happy with the foundry and where and how far they've come along with 18A, and I'm obviously using that as a happy customer for Panther Lake. But one thing that I constantly have a conversation about internally is, sometimes, an Intel Foundry is going to make sense for my products and sometimes it's not. By the way, even when we think about it from a balance sheet perspective, I can't deploy the amount of capital that would be necessary to build on my products either. And so how do I use both Intel Foundry we use Samsung as well and TSMC in ways that really allow me to optimize. And so I've been very public that come our next generation of product, Nova Lake, I'm using both TSMC and Intel Foundry. And I'm doing that because I think it allows me to deliver a more competitive product for our customers. And at the end of the day, the best product wins. Yes, of course, I want it to be on an Intel Foundry, but if it doesn't deliver the best product, I'm not going to build it there.

Isn't there a cost to engaging with 2 foundries?

There is a cost to engaging with 2 foundries. But if you do it in the right way, the cost is pretty minimal. And what I mean by that is if you do it by package type, et cetera, and you're using more and more of the industry available IPs, I don't have to redesign them. And there's a cost to doing that. But a lot of times, I can get the my partners, my manufacturing partners to bear the brunt of some of that cost as well. And so they're willing to take some of that on. And so -- but at the end of the day, that cost is worth it if you have the right product mix portfolio. And so what I've heard from customers time and time again is I want the best products to solve my customers' needs. We'll leave it to you to figure out how to put that package combination in place. And I think it's really important because at the end of the day, as investors, customers buy the best product over and over again. I mean we've certainly seen that in AI categories. And there are certainly markets where continuity is probably more important sometimes than the best product, but that only lasts for so long. So you may be able to have one generation of product where you may not be unequivocally the best, but you don't get like rest on those laurels. And so at the end of the day, we have on the product side, a lot of flexibility to do that. And so we've been very forthcoming that our mix moving forward is 70-30. And we think that, that's the right mix based on being able to deliver the best products to the customers. But we would be open to that changing if it needed to, to make sure that we could deliver a world-class road map.

Got it. And then on the competitive landscape, even though Intel has access to TSMC and obviously, your internal manufacturing, what has enabled your competitor to take the lead in average selling prices? Because one difference that was clear from the kind of the respective Q1 earnings calls was that you saw mix kind of downshift, right? They are not noticing that, right? So their ASPs have gotten better. So even though both have access to the same foundry conceptually, what is helping them gain the upper hand on ASPs, at least in the near term?

You're really talking about ASP growth?

That's right? Yes.

Yes. So if you look at our primary competitor, they have generally historically sold N minus 2 through N minus 4 products. And so really, what you saw isn't them selling a lot more of the end node, but more N minus 2 and N minus 3 products. And so really, it's a slight mix shift. And for each of those process node mix shifts, it's about you get more ASP because it's a better product. So I think it's less about the way that they've used the manufacturing capabilities that are in the marketplace and more about how the mix overall has changed. And I do think it is credibility to the fact that their product road map has gotten much stronger. And so as your product road map gets stronger, you have an easier time selling into those, I guess, you could say, more current products. And when you're selling N minus 4 and you move to N minus 2, you get a pretty significant ASP uplift, and we've continued to see that in competitors. We do too get ASP uplift. But in Q1, particularly, we did see people buy older node products because they wanted to balance their portfolio mix for tariffs. And really, what that means is in a lot of -- if you have an older mix of product inventory, you can hit more ASP price points, particularly in the client space because your overall cost structure is lower. So depending on the size of tariff, you can still fit in that $6.99 to $7.99 price point where the majority of client volume is, whereas like a $12.99 product for an OEM with tariffs would jump to $17.99. And that market TAM is very small at $17.99. So there's a much larger risk in carrying that inventory. Remember, the OEMs don't have a lot of cash on their balance sheet. They don't make large margins per unit. So it's much easier for them when they're thinking about making sure that they have the right inventory for it to be lower cost inventory. And so that both advantage our competitor and maybe a little bit of a disadvantage from an ASP perspective for Intel, but I'm happy to sell those products because I think great gross margins on them as well. So -- and we see -- we continue to see that in Q2 where demand for those products remains quite strong. In fact, I'm supply constrained.

Got it. And then one other thing we noticed was that this year, Intel did not make the incentive payments, right, to your PC OEMs, which I think we have seen in prior years. So what made you kind of go away from that?

Well, I think most of us are more in the mindset of you want the business to land in the quarter for which it needs to land. We do have a more competitive client road map than we have had historically, and we should get paid for that as well as particularly in Q1, you just saw strong demand because we did see pull-ins. And so you don't want to go pay people to pull in product if they're naturally going to do it anyways. But I think Lip-Bu, Dave and I have maybe a little bit different philosophy of we really want to land the business that's in the business within the quarter. And when it makes sense to do a customer incentive, I'm not saying we won't do it, but I think we just want to be very balanced and pragmatic about the way that we go about that. And as our road map gets more competitive, we should have to spend less dollars in that regard.

I see. Have you seen any change in the competitive situation on enterprise PCs in that -- again, the competitors' claim is that, look, we are gaining share. We have all these SKUs at Dell and whatnot. So have you -- and this has been an area of real strength and dominance, right, for...

And I believe that enterprise will remain a real strength. This is one of the categories where when you're a CIO and you're thinking about managing the fleet of iPads or notebooks or whatever you deploy as your edge unit, ubiquity in those units is extremely important. Things like vPro technology manageability and security are extremely important, and we've done a lot of upgrades for enterprise, particularly around vPro this year. And so there are a lot more competitive enterprise SKUs. And I think the question that we need to watch and we need to be kind of laser-focused on is how much sell-in do you have. It's one thing for an OEM to offer the SKU. It's another thing to actually gain traction. And what I hear from CIOs is that they love working with Intel. We deploy engineering talent at the second they have a problem. We understand how to manage their fleet, and we're very good at working with them to deploy real-time updates to their fleet. And we know what it means to manage a fleet. And so it is one thing to build an enterprise product. It's another thing to manage an enterprise portfolio. And we have seen that to be a core strength. And you can be assured, I know that's a core strength and an area we'll continue to invest and an area where I absolutely want to protect market segment share because it is sticky share. There's some share where you can lose and bounce right back just because you have the next best product. This is a place where you don't bounce back really quickly from. And so I'm going to do everything we can to protect share in this space. And I actually think most of the OEMs are on board in wanting to deploy Intel into enterprise because they see good traction and stickiness for themselves as well.

You mentioned Panther Lake is up and running. You're happy with that success. So if that progress is going per your expectations, why go back to TSMC for Nova Lake?

Yes. So maybe just to baseline everybody on Panther Lake. So Panther Lake is a product that's going to launch in the second half of this year. And it is all built on Intel 18A. And really -- so Panther Lake is an all-mobile stack. When you get to the next-generation Nova Lake, it is both a mobile stack and a desktop stack. And so one of the things about the desktop market, which is a place that we have lost market Sigma share, it is a very elastic market. The best product at the time of graphics card launch is really how you kind of take advantage of that TAM. And so being able to land on a node that is already ramped is at very high performance plus yield is very important. So you can imagine, I'm looking at how much yield and product can I get in a very short amount of time. And so when you look at that, you might actually pick maybe not the latest dot of a node at TSMC, but you know you can get a lot of wafers and a lot of product in a really short amount of time. And so you'd put that SKU on TSMC. And so when I say I'm pragmatic, I literally look at it by SKU and where it makes the most sense. And so I like personally a portfolio where I use both boundaries because at times, I want to be cost. At times, I want to be about volume and at times, I want to be about performance. And depending on which is most important for the customer and the segment, that's what I pick. And so it isn't anything about Intel doing something wrong. It's more about optimizing the mix to be able to deliver best on behalf of my customers and hit a particular market window with a particular amount of volume.

Makes sense. On the server CPU side, when do you think investors should say, all right, this is the inter product that's going to help start taking the market share in the other direction?

Yes. So when we look at the data center market, what we've said is like in 2025, we want to stem the share loss. We want to kind of get to the bottom and then start to move up. So obviously, customers are deploying Granite Rapids right now, which is a good step function. And then we've got Clearwater Forest and Diamond Rapids, which is our E-core and our P-core product lines coming in '26. And so I think we'll stem the tide and then we'll start to see market segment share start to build back up with those next 2 generations of products. But it's a journey. It's not like this large step function where you're going to see 10 points of share added where -- because these are more inelastic markets, and it takes time for customers to ramp. And in some of these cases, you have to move an entire data center. to your product, where it may have been deployed all on a competitor and you have to make like a lock step shift with some of these customers. And so it takes time. So the journey back in data center is a little longer than it would be in client as an example. But I feel good about the products we have. I feel good about the execution on Granite Rapids, and now we need to repeat that on Clearwater Forest and Diamond Rapids. We have good feedback from customers on both of the products. And so it comes to execution. I think it's not a surprise. Our execution on the client side has been more consistent than it has been on the data center side. And so I'm really trying to bring a lot of the best known methods and things that we've deployed on the client side to data center, but we've just got to continue to build products that our customers want to bet on. Because, obviously, not just for us, but for our customers, data center is a high-growth category, high-margin category, and you can't afford to miss the market window. And so we need to be the trusted partner for them in that. And I think we're doing a better and better job, but it's a journey back to the Intel, I think, that all of you would expect.

And in terms of the wallet share shift that we have seen in the last few years towards AI, right, and away from traditional compute, are we on the other side of that? Like are we at a point where that trend has stabilized? Or -- but I guess the roundabout way that I'm trying to answer -- ask this question is that if Intel's AI product takes some time, right, to become competitive, does that influence how quickly you can rebound on the server CPU side, if there is a correlation, then?

If there's a correlation. We're actually doing quite well on server head nodes for AI. So I mean our market segment share there is actually quite good. So the more AI grows, I think that's positive for us. I think we're just at the beginning of the AI explosion. I mean this is a $1 trillion market that's going to continue for a while. I think the million-dollar question for us is when are we going to have rack scale architecture products that we can deploy in meaningful size so that billions of dollars of revenue. I think it's in the next few years. I mean I think that's really for Lip-Bu and Sachin to come talk about. But I do think they have ways over the next -- each year of coming to market with new products that will help us learn and continue to be there. But I mean there's no one at Intel that's going to tell you that we're late here. But I'm optimistic that we have the right pieces. We have the right APs in CPU. We have the right APs in graphics in interconnectivity. We know how to do rack scale. But we also need an infusion of talent. And this is a huge focus for Lip-Bu of bringing in some new talent to help us here. And I think we also have to look at delivering products in these categories that fit in some themes in the marketplace, maybe nontraditional places where we can differentiate ourselves against the market leader NVIDIA. And so that will be a big focus for us as well. The one thing I can tell you, and I'm sure as you've had your meetings throughout the day, people want alternatives. People want choice. They want open. Those are all things that Intel brings. Now we have to bring them in a meaningful way. And I can tell you all that work has been happening. I've seen product road map plans. And so I think there's a meaningful opportunity there that we can go and deliver. But at the end of the day, we have to prove it to customers, and we need partners that want to deploy with us in mass scale because that's what it takes. But I'm optimistic we can get there.

But -- and you expect it to be more organic or inorganic?

It probably depends on the segment. So it's probably a little bit of both. Like if you're thinking about like training, it's very different than if you're thinking about inference as an example. And I don't think we can do it all by ourselves. Like I think we have to use industry partners, et cetera. And so I'll let Sachin and Lip-Bu talk a bit more about that. But I can tell you in the first 8 weeks that he has spent a lot of time here, spent a lot of time with industry partners about how to do this. And it's the #1 request of every one of our customers, right? Like get in the AI space, I need alternatives, I need partners and you're missing. And so it's a good opportunity for us. And we certainly have all the pieces of the toolkit. We just need to put the Rubik's cube back together.

Got it. How do you see competition from ARM? Because outside of the Apple ecosystem, they don't show up in very big numbers in Mercury, right? But then we hear of these large hyperscalers talk about, oh, half of my instances are based on this ARM product or somebody else saying, all my incremental instances. So what do you actually see, MJ, from your customer discussion? How important is ARM to them? And let's say, if it is more important than you think, then what is Intel's strategy to deal with it?

Yes. I think the answers are very different depending on your client or data center. I think in client there was this large push for diversification over the last 18 months. You saw, obviously, one of our competitors come to market with an ARM-based solution. But I think what we found is no matter how much money you throw at it, I actually think to fix the ARM ecosystem pieces in client, it's really about volume. So it's about shipping billions of units and getting the software ecosystem that x86 has built over the last couple of decades. It's really getting that to port to the ARM ecosystem. And that is not something that's going to happen overnight. It's going to take a lot of time. Data center is very different. In data center, we do see large deployments of ARM, particularly at a lot of the CSPs. My understanding is that their deployment of ARM is more their internal workloads. So the workloads where they can do the software work and they can deploy that. What I'm not saying is when you think about somebody who's going from on-prem and moving a workload to the cloud, if that's a customized workload for them, I am not seeing a lot of that being deployed on an ARM-based workload. I'm seeing it being deployed on an x86-based workload. And so this is where our open x86 architecture and building chiplets -- and I think, albeit we could have done this earlier. But the fact that now if they want to go build a custom CPU for their data center and they can do it on x86, that will be easier for them to port the software ecosystem outside of their custom workloads there. And so I think you'll see a deployment of both. But I'm optimistic that by opening up x86 and making it available, if it's easier for the software ecosystem to port, I think we put to bed that we can't be -- we can't match them in the performance per watt. I think we've taken that off the table. So the fact that there are no intrinsic things that would stop somebody from deploying, now I think we need to go enable a few of them to build a custom x86 chiplet with their own configuration and deploy it and see how it works. But -- but ARM is a formidable competitor in data center and they're a formidable competitor, hey, they have a product, but they were also much better at listening to the customers' needs and building a customized -- allowing them to build a customized product. So I now have that, albeit I might be a bit late to the market, but I still think I have the software ecosystem in favor. And so now I need to get some of those large CSPs to be building some custom chips with an x86 core instead of an ARM core. Interesting. And we do see interest. I don't have any design wins to report, but we do see any new design wins to report, but I do see some strong interest in really understanding -- I don't think there's a customer in the marketplace that would tell you that the value of the open x86 software ecosystem it's very, very strong. And I think you'll hear that from every competitor that we have. And so we just haven't necessarily utilized that strength as well as we should and now we are. And so I think now it's on us as a product team to go and get some design wins and show the world what it can do.

Right. You mentioned the move towards 50%, that's kind of an intermediate type target. What needs to happen? Is it a product issue? Is it a pricing issue? What is the time frame over which you think Intel can get to those kind of gross margin levels?

Yes. We haven't actually forecasted a gross margin, but we all know we should be at 50% or above. And we've been there before. And so I think our future products can all get there. I think really what it comes down to is you have to have a lot of discipline in your product life cycle planning to build products from day 1 that hit that. And so there's a lot of things that I talked about when we talked about Lip-Bu coming on board of getting our OpEx and our CapEx in line, getting the types of products that we're going to build in alignment, really understanding market and ASPs. And so when we'll see that across the board, but future products that are already on my road map show me that we can have a 50% gross margin like across that entire product line. And so that's the expectations that our shareholders have. That's what our competitors can do. There's no reason we can't do it. But there's some things that we have to change internally, I think, particularly around our spending, the way we build products, the efficiency for way we build products. One of the things you've heard Lip-Bu say a lot of we need to go to market a lot more on a stepping. It go -- every stepping costs a significant amount of money. And so we needed to be doing a lot more upfront silicon validation and things like that prior to tape-ins and things like that. And so...

More EDA.

More EDA. Yes, I imagine. He likes that. But there's just a lot more of those types of things that we naturally need to build into our product processes that we had already started doing. And now, obviously, we'll do it more rapidly. That will help us get there. But I think the most important message is we all know that that's the expectation. And moving forward, if you have a product and you're going through like our decision matrix, you actually can't get approved if you're not a product that can show me that you can get above a 50% gross margin. Base on a set of industry expectations and ASP, which is something that we probably should have had before, but we have it now. So that product doesn't move forward. You actually don't get engineers assigned to it if it's not 50% or higher gross margins moving forward.

Okay. And then the last question is, in the theoretical scenario where Intel is not able to get a large external customer for 18A or 14A how does your product strategy and in-sourcing versus outsourcing decision and the cost structure and margins change?

I don't know if they really changed because when you think about what we've said about foundry and profitability at the end of 2027, it has a very small external footprint. So I think part of this is just getting the right mix in the factory, the right cost structure, et cetera. But I think as you -- so I think on the product side, it's probably less on the product side. I think more it's about foundry and thinking about what kind of capital deployments are they going to do? And what kind of customers do you need to ensure that you have so that those dollars are spent well on behalf of shareholders, right? And so if you're thinking about 14A, in 18A, we started that process not as a foundry process and then we tried to make it a foundry process. 14A, we're starting from day 1 with being a foundry process and having all the right PDKs and IPs in the industry. But I -- if I were going through that process, I would say, to make those investments, you need to know that you have a customer besides Intel products that's going to fill that factory because you just -- you need that diligence, I think, to make go and make that capital investment. And if you don't have it, then you do iterations of 18A, 18A-P and whatever after until you can show a customer that you can get there and then they come on board. And so I think those are the level of discussions that Lip-Bu is going to need to be having as he's thinking about the overall balance sheet and the way that we deploy capital. But even before then, there's a lot we can do. We have a lot of assets under construction over $50 billion. how do we get all that deployed before we build anything else as we think about deploying our capital dollars. This year, we'll actually have the lowest number we've had in the last 4 years of AUC. And I just think there's a lot of financial discipline that we can have there that we're laser-focused on. You kind of got to do all these things in parallel, right? And so -- but I feel good about where the foundry is. I feel good about where they'll get to profitability by the end of 2027. And I think to continue to make investments, we've got to show we can bring other customers in.

Got it. And last question, I know we are at the end of our time. What are you most excited about the second half, right, of this year, right? I know there's a lot of macro crosscurrents, but what...

Well, I think there's a great opportunity in the marketplace. I mean, like, obviously, I come from the client side of the business, but with Windows 11 refresh, with the explosion of AI, I mean, AI is going to come to the edge at a rate and speed that I don't think any of us have predicted. Like the excitement that I hear from customers about local AI, all the verticals for AI. I mean I just think there's an explosion of opportunity that we haven't even seen unleashed yet. We just need to make sure that some of the other challenges that we're all balancing around tariffs and things don't slow it down. But I'm actually quite optimistic about the market, AI, where we are from a road map perspective and new places that you'll see AI where I think very much play to Intel's strength. So I remain very, very optimistic about the market in general and the way that Intel can play and service our customers there.

Terrific. . Thank you MJ. Thank you for your insights. Really appreciate it.

Thank you, everybody, for coming.