All right. Welcome back to the Barclays Tech Conference. I'm Tom O'Malley, semiconductor and semicap equipment analyst here at Barclays. Before I introduce our guests, I just need to read an opening statement. So before we begin, please note that today's discussion may 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 and 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 for non-GAAP financial measures, including reconciliations where appropriate to the corresponding GAAP financial measures. Okay. Now to the good stuff. Great to be here with you both. Michelle Johnston Holthaus.
Yep.
Perfect. And Dave Zinsner, co-CEOs. And very happy to have you here today.
Thanks for having us.
Thanks for having us.
Why don't we start off with the question that everyone's been asking you since Monday of last week, which is change of regime. How do you guys think things will differ with you two at the helm versus when Pat was at that?
You want me to start?
Sure.
All right. Well, I think there'll be a few differences and there'll be a few things that don't change. Obviously, building world-class products in a world-class foundry, we're still highly invested in doing that. And those two things together will help differentiate us in the marketplace. Pat, clearly, Dave and I enjoyed working with Pat. And he left us in a better operational place.
But I think Dave and I will both tell you that as the CEO of Intel Products, we're going to invest more in products, be focused on making sure that we shore up those roadmaps, that we're more competitive in a lot of the growth markets than we have been historically.
And that will then fill the fabs, right? So great products mean more wafers, which means better capacity in our fabs. As well as we're laser-focused, we've obviously made very large investments in our fab footprint. And we need a much better ROI for the investment that we've made. So we're going to be laser-focused on how do we take advantage of the investments we've made, how do we bring customers into those fabs and start to see wafers fill them.
Yeah. The only other thing I would say is we've been asked how we're working together, what roles we're taking on. For most of it, we're co-CEOs, so we want to be in the room together doing it together. Obviously, Michelle is super experienced in the product side of the business, dealing with customers. So probably makes more sense for her to emphasize that. She's out there meeting customers on the product side on a regular basis.
For me, I'm more in the operational financial aspects of the business. And so I'm going to lean a little bit more heavily to that. Probably the only exception will be the customers for foundry. Obviously, they want firewalls. They want to protect their IP, their product roadmaps, and so forth. So I will deal with that part of foundry to separate that from the Intel products business. But outside of that, as we deal with the foundry business, we'll do it jointly together.
Incredible. Why don't we start with Michelle? So on the recent earnings call, you talked about more outsourced silicon in 2025 than 2024. In the future, do you see that trend increasing? Or could you maybe speak to the next couple of years in your mind?
Yeah. The answer is kind of maybe, maybe not. I mean, when we are thinking about Intel products, we're really looking at time to market, performance, area, cost, a lot of things when we are thinking about who is the right manufacturing partner for us for products that we're building. And so maybe to take a step back, when I was running CCG, the way I made those decisions is really looking at, okay, you have a market window that you have to hit in client, right? You have your holiday and you have your SIPP or your commercial buying cycle. So your products have to hit those windows. Obviously, we have a very broad set of product portfolio, and you need a variety of price points and a variety of performance price points as well.
You look at those and you look at what Intel Foundry can offer and you look at what TSMC can offer. What I found was at times picking TSMC was the right decision because where I could land on their performance price curve made the most sense for the ASP that I could get at the time. Then other times it made sense to use Intel Foundry. Now, we are using Intel Foundry for Panther Lake, which is our 2025 product, which will land on 18A. This is the first time that we're customer zero in a long time on an Intel process. Just as you would expect, Apple's experiences with TSMC.
You make progress, you take a few steps back, and then you make progress moving forward. We're quite happy with the decision to be on 18A. But I'll also say TSMC has been a fantastic partner. They're very easy to work with. And myself being very customer-oriented, what I tell everybody at Intel every day is they're the benchmark for what's expected in the industry. And so I feel like working with the Intel Foundry, I can continue to push the bar to make us more competitive and set expectations for what it means to be a world-class foundry.
Super helpful. And I think Pat used to talk about 18A being the kind of keystone for the company, like betting the company on 18A. I think that there'll be milestones coming up that we'll expect to hear from you guys on 18A. But I guess the question is, how tied to the potential separation of the business is 18A? Can you see an Intel Foundry Co and Product Co? Maybe this is a question for both of you. Is that something that is possible today? And maybe any iteration of that that you'd be willing to kind of go into?
Yeah, I think it's a really good question. I mean, we really do already run the businesses fairly independently. Product Co makes their decisions, Foundry makes their decisions. But for both of those businesses, I think long-term to be very differentiated in the market, great products with a great process technology that we have first access to is a differentiator. So pragmatically, do I think it makes sense that they're completely separated and there's no tie? I don't think so, but someone will decide that.
I think 18A itself obviously is kind of our first big step in getting back to process leadership and bringing our foundries back to kind of the historical strengths that we've had and as you said, there's some good and some good/bad that comes every time you develop a new process. But just to give some assurances on Panther Lake, we have our ES0 samples out with customers. We have eight customers that have powered on, which gives you just kind of an idea that the health of the silicon is good and the health of the foundry is good. But we have a lot of milestones, especially a couple over the next couple of weeks to get to, to be ready and on time for the end of 2025.
Yeah. The only other thing I'd add is we already run the businesses separately, but we are going down the path of creating a subsidiary for Intel Foundry as part of the overall Intel company. We're going to have a separate operational board for Intel Foundry, which is getting stood up today. We'll have its own ERP system and so forth. So that's already in flight. That's going to happen. And we're just a matter of time at this point that we lock in the milestones associated with creating that separation. As far as does it ever fully separate, I think that's an open question for another day.
I think you guys have talked historically, customer interaction, customer announcements are the best way that at least people in these seats can track the progress there. Is there anything else that you would offer now that you've had two full weeks under the covers here to talk about milestones in terms of what we could look for to say, "Hey, 18A is on track," or to give us a little more comfort on that?
Yeah. I mean, we've hit milestones. We've talked about hitting milestones in the past on earnings calls. We have a couple of milestones that are coming up here in the not too distant future that we're likely to update investors on at the end of January. Early indications are things are going well there. From a customer perspective, as we talked about in the prior earnings call, we've got a number of RFPs that we're working through, RFQs, whatever they're called, that we're working through over the course of the next few months.
And my sense is we'll probably provide some update at January. But I think you're going to find us, and this is kind of a characteristic of Michelle and I, we are a little bit more on a say-do basis. We're more likely to tell you things as we've kind of accomplished big milestones that are meaningful as opposed to early indications of success. So that's our philosophy. I think that's why the board actually chose us to run the interim role because we are so transparent and operate in that way. And I think investors can expect us to do that as we go through earnings calls and so forth.
Okay. So there's the technology roadmap, and then there's the execution in the near term that helps the profitability as well. And we spent some time last night, Dave, talking about profitability actions in the near term. Do you want to talk about Altera and your Mobileye stake? Those are two actionable items that you guys have talked about in the past. Can you maybe spend a little time on the health of those businesses first and then maybe potential plans for those two?
Yeah. Both of those businesses are in kind of a recovery. So that's good, obviously. Mobileye, we do want to hold a reasonable stake in the company, although we recognize that investors need liquidity as well. And we could use cash. So you will see us probably sell some of our position over time. But we're not under the gun to do that. We're going to do that in an appropriate way to drive the right valuation, the right cash inflow, do the right thing for Mobileye shareholders and so forth.
On Altera, we have kicked off the process. We talked about this at earnings. We kicked off the process to engage with outside investors to take a stake in Altera. Our thinking is we'll get another partner in similar to what we did with the IMS business, get a financial partner in with us. We think that there's a lot of opportunity to drive significant value creation in Altera from where we think it is today, so get a stake in, clean up the business more, put it in a good position, and ultimately take it public.
One of the defining moments of this past year was your initial efforts on costs and the right-sizing of the business. Are we done with that endeavor at Intel? Obviously, it's a painful exercise for any company to go through. But do you think there are additional ways that you could right-size profitability standing still today?
Yeah. Now you're talking to the CFO. We are never done.
That's it. I only want to answer one thing.
Yeah. Look, I think we're done with this big reduction in force that we had to take. I think we're complete with that largely. However, we are going to constantly be scrutinizing where we're spending money, making sure that we're getting the appropriate return. We have a number of different initiatives. I would say part of the way Intel has been run has driven a lot of complexity. And I think there are ways we can simplify what we're doing, what we're focused on. Certainly, the message from the board last Monday is they wanted focus. And I think that will be beneficial to us just in terms of operational improvements, but also from a spending perspective. So I think you will see us continue to be highly focused on spending in general.
As far as the $17.5 billion OpEx number is concerned for next year, we're in a good place there. What we want to do is kind of bring it down a little bit more the following year. I think still early in that process, but I'm confident that we'll get there. We have done a lot of work on capital spending to bring that number more in line. We're now at this point focused on a number on a gross basis that's 20-23. Obviously, we expect offsets. The net number is a lot lower. I think there are more opportunities to scrutinize capital spending over time, greater reuse of previously purchased equipment. I think there are opportunities not to get the gold-plated tool and save money in that regard. We're going to be hyper-focused in terms of spending on capital.
And then lastly, we want to take a billion dollars out of Intel's OCOS, which is an acronym for other cost of sales, which is essentially period cost in the cost of sales line. There's a big number there. I mean, it's billions of dollars of spend in other cost of sales. And I think we will certainly be able to take this billion dollars out. We have a good line of sight. But I think there's more opportunity there. And quite honestly, with Naga coming in to run manufacturing, he thinks there's opportunities even on what we call the P cost, which is more the variable cost that we spend for wafers. And so we'll constantly be focusing on this to drive a much better level of profitability, which we understand is the key aspect of driving value creation for shareholders.
The only thing I might add for products is we're going to be laser-focused on where do we sit versus other competitors from a best-known methodology in designing products. But one of the things I know that we can do to drive better efficiency is using IPs from the cloud all the way to the edge. Today, those are three independent teams. They all design their products independently. And I think we have a large opportunity to really think about the way we can use the IP portfolio across the entirety of our product portfolio. And by the way, that's something customers would like to see as well. So it's a large opportunity, I think, for us.
I want to go back there, but just one more on Dave's comments. So the offsets are obviously a big deal as well. You're having a regime change. There's been, I think Vogel is tweeting, so I guess tweets and conversation about potentially eliminating some of that funding. How would you view spend if you were to see a change in that type of support?
Yeah. Well, first, I would say we signed the CHIPS agreement. It's an agreement. Because it took a while, we have hit a lot of the milestones associated with that agreement already. And so we're actually now in the process of, obviously, a lot of documentation that goes towards the submissions to get payments. We're already in that process. So I think we're going to see meaningful inflow in the not too distant future, even before the transition in government. But it is a key component of our calculus in terms of spending. So obviously, if we think that there's an adjustment happening, we will adjust spending accordingly. I would point out one other thing. If you look at our offsets, clearly, there's the SCIPs, which are offsets as well.
But if you look at the offsets that are pure kind of government offsets, it's a combination of grant money and the investment tax credit. And in reality, the investment tax credit is three times the value of the grant. And it tends to get ignored, but it is the more significant part of our offsets over time. Early on, the grants are more, but over time, it will be the investment tax credit that really drives a lot of that. That's tax law. Of course, tax law can change, but I'm assuming that that will not happen. And it's actually relatively straightforward in terms of how you get it. You essentially spend the money. There's a time at which you can claim it on your tax return, and then you get a refund in the subsequent year when you file your taxes.
Okay. On the product side, what I wanted to touch on was the AI PC. So a lot of debate as to what is an AI PC. You guys have talked about it from a product line perspective. So when you sell a certain type of product that is an AI PC. Can you talk about 2025? Is this an inflection year in what you guys define as AI PC? Would you define AI PC differently? And is that something that helps Intel in terms of unit volume, or is this an ASP uplift, or is this net-net just a redefinition of the existing market?
Yeah, it's a good question. So AI PC obviously started in 2024, and we've openly stated we'll sell about 40 million units this year. But I'm also a pragmatist. Most of those are being used in the same way you're all using your notebooks today. We've done a large investment in working with the software ecosystem to bring over 400 ISVs to the marketplace because you need to start seeing those usage models that really drive why you would buy hardware. And so when I look at 2025, between all the work that we're doing, the work that you've seen Microsoft doing, I do think you're going to start to see more demand for, "I need an AI PC." There's a lot of versions of what an AI PC is.
At the end of the day, I think the one thing the entire industry aligns on is you need a neural processing unit to do the low-latency work in the background that allows you to have battery life but to be doing those AI workloads all day long. And so we're all aligned on that. So when I think about 2025, I think it's volume. You'll certainly see ASP uplift that comes with AI PCs as well. And you're going to start to see a lot of the CIOs making investments for future-proofing. So they may not know exactly how they want to use an AI PC today, but they know that the longevity of that purchase is over a three to five-year horizon. And so we're already seeing CIOs coming to us saying, "Hey, what do I need? What is future-proofing?
How do I need to be thinking about those purchases?" which I think is a very good indicator that a lot of times either software proceeds or hardware proceeds, and in this case, much like when you all got Wi-Fi on your machines there, you had to put the capability in the machine, and then you had to deploy all the hotspots, and so in this case, now we have the hardware and the machines, and we really need that software ecosystem to really start the innovation engine and to show people what they never knew they wanted to do on their PC is now possible.
So there's two threats, I think, to the PC market that I think I'd just like to hear your opinion on. Obviously, there's the Arm ecosystem. So you had Qualcomm talking about by 2030, 50% of the market moving to Arm PC, which is an aggressive number, I think, by many people's standards. But anyway, still something to notice. And then two, obviously, you've been on the road for the last two weeks probably talking to customers and your competitive dynamic with AMD, where if you hear about a leadership change, the instant reaction from some customers is, "What's the roadmap? Where are we going here?" So can you talk about those two threats and how that may impact the next couple of years?
Absolutely. So when we think about Arm, obviously, the Apple machines are all based on Arm, and they've been 8%-9% pretty relatively stable from a growth perspective. And Qualcomm makes up about less than 1% of the PC market today. And if you look at the investment in Arm, you look at the work that Microsoft's done, I mean, there has been a very large push to make Arm ubiquitous in the PC. And there's some real challenges to Arm being ubiquitous in the PC. You will never hear me say that it will not happen because competition makes us better. And as long as you're constantly worried about who's knocking on your door, you're going to constantly be innovating and making sure that you don't have blind spots. But we do see that there's still a lot of incompatibilities.
I mean, if you look at the return rate for Arm PCs, you go talk to any retailer, their number one concern is, "Wow, I get a large percentage of these back because you go to set them up, and the things that we just expect don't work." And Apple did a lot of that heavy lift for Arm to make that ubiquitous with their iOS and their whole walled garden stack. So I'm not going to say Arm will get more share than it gets today, but there are certainly, I think, some real barriers to getting there. And I think another barrier is we took too long at Intel to become performance and power-oriented. And we made a massive leap with our Lunar Lake product last year. We are as performant on performance and battery life as most Arm devices out there.
For our customers, a lot of them are saying, "Okay, you're finally in the ballpark of being focused on all these right things. Therefore, I believe I can bet on an x86 architecture." When it comes to AMD, we both kind of have the same bet, right? When we think about the work that Lisa's doing and the work that we're doing, we believe that x86 is the best overall basic architecture and will continue to build upon that. I've spent a lot of time with customers in the last two weeks, as you can imagine. Probably the thing that is the most exciting about the last two weeks, despite a lot of very, very difficult conversations, is customers want us to be successful. Our customers have decades of relationships with Intel, and those don't go away overnight. I've seen customers lean in.
I've seen customers change their roadmap. I've seen customers say, "Michelle, I need you to look me in the eye and tell me that your say-do ratio is going to continue as it has for the last three years and that you're going to tell me if something changes," and so there's a lot of trust build-up there. We have a lot more trust to continue to build, and I'm not saying it's not going to be bumpy, and I'm not saying that others won't take advantage of certainly the few potholes that we've had in the last couple of weeks, but I feel good about where we are, particularly on the client side, but my say-do ratio for my customers, I want it to be perfect. I ran sales for a long time. Nobody likes to send a dear customer letter ever.
It's the worst thing you can do to your customers because they bet on you. They bet their business on you. We have a lot of work to do on the data center side, which I'm sure you'll ask me about, but we have a lot of work to do there. But on the client side, our say-do ratio for the last four years has been very good. We've met our schedules. We've met our performance. So you can expect that to continue. But everybody is really excited about the PC market, as you said. So we have more competitors than we've ever had. You will see more competitors enter the marketplace in 2025, and we're going to have to be on our toes and making sure that we're winning.
Maybe one thing I just want to add is I'll brag for Michelle since she probably won't brag for herself, so keep in mind a couple of things because the start of this question was around the change in leadership. Michelle and I are co-interim CEOs, so that's a temporary role. Somewhat lost in all of that is Michelle got promoted to be CEO of products, and that is a permanent role. I mean, she is going to be the CEO of the products business. We've never had that kind of across-the-board leader on that business, which I think is important.
It helps frame the roadmap, thinking about all the dynamics between the various businesses, thinking about how the functions work together to be successful, so all of that is under Michelle's purview, and Michelle is very, very good with customers, very good with customers. Universally, the feedback, even in this last week and a half from customers, is very positive around Michelle getting that role and they being able to interface with her not only in the client side of the business, but also in data center and in the networking part of the business. So I actually think we have improved that dynamic with customers and put us in a better position relative to competition given this new structure.
Thank you.
Yeah, you're welcome.
Yeah. Thanks for the title. Forgive me for not also saying CEO and so forth.
Not so important.
But I am remiss not to mention. So following up on the PC side, just jumping right back into it, the data center side, it seems like you're more comfortable on the PC side than you are on the data center side. Could you talk about the x86 dynamic in the data center first, and we can move to AI in a little bit? But where are the differences versus where you are in PC, and why does your feeling seem a little bit more, I'd say, skeptical about where you're at today?
Yeah. I mean, I don't think it's a surprise to anybody. We've had some big challenges on the data center side from a market segment share perspective, from a competitiveness perspective. We've talked about those. And as you look at kind of data center market moving forward, there's not a lot of TAM growth. It's really about core count growth in that. And for us, 2025 is a year of stabilization. We've been talking about that since Q2, about stabilizing that market segment share loss, being really laser-focused on building the right products to regain share. We have two products coming with Clearwater Forest and DMR. Now I'm going to blank on what that stands for.
DMR?
Yeah.
Diamond Rapids?
Diamond Rapids. Thank you. We talk in acronyms all the time. So Diamond Rapids, which is our P-core products moving forward. And so we have a lot of work to do there, a lot of work to do there. And I think as much work to do on the product side as we do on the customer side. And that business has changed a lot. If you look at the CSPs, most of them have custom parts. We obviously have our new x86 consortium where we'll work to give people head-node IP so that they can build it based on x86. So there's that piece. But obviously, AI has also come in, and the shift from CPU to GPU has been probably the largest shift we've seen in decades in regards to the shift in the way that compute is used.
And so I think that has really caused, at least me in this job, to kind of stop, reflect, and listen, trying to understand why did we think we were going to be successful? How do we learn from that? And then how do we put those learnings into what we do moving forward? I feel good about where we are from a Xeon roadmap perspective, but we have a lot of work to do on the AI side. I'm sure you'll ask me a question. Everybody wants to know about AI. Obviously, it's this massive market opportunity. So for me, my first couple of weeks are just about listening. What's working for our customers? What's not working? What are they happy about? What are they not happy about? Then we are going to come back. We're going to huddle. We're going to reevaluate. What are we doing?
Is it actually going to win? If it's not, what pivots do we need to make? I don't know what those are yet, if I'm just blunt. And most importantly, who's with us on this journey? We need to find a few key partners that want to win with Intel. There are a lot of advantages to the architecture that we have, and we see that. But up until now, AMD has been doing a better job of servicing those customers. So those were my customers before when I ran sales. And so I'm going to be laser-focused, and how do I switch them back? But it requires not just relationships.
It requires great engineering and great products. And so part of the shift in our strategy is, you want my opinion, we weren't making enough investments in products. We just weren't. We weren't moving fast enough, and people were leapfrogging us, and they were being more innovative, and they were more willing to disrupt themselves. What you'll see from me is I'm willing to disrupt myself. I'm willing to have a year that isn't so great if the next year is even better because at the end of the day, if we build world-class products that allow our customers to win, we'll win. That's, I think, where we really fell short.
With this products-first mindset, and forgive me, it's the question everyone asks as well, you've been more open about outsourcing in your PC business. If you get to a point where you feel like for your products to be competitive, you need to move in that direction as well, is that something that the two of you would feel comfortable moving forward with, or would you wait to have someone move in? I know that's a tough question, but how would you think about that?
Oh, I would feel very comfortable with it. Because at the end of the day, bluntly, I can be attached to my Intel Foundry, but if I have a losing product, that doesn't really help. I need to build world-class products that my customers are excited about and they want to buy. and if that means landing something on our data center roadmap on TSMC, I'll do it. Do I think long-term it will come back to Intel? I do because I do think we have a good roadmap, and I do think we're making positive progress. but in the short term, if that needs to be done, I'll make that decision.
I think that's a very refreshing attitude. So next question is on the data center side for AI. So I think that one thing, I think there was a lot of debate about the last several years at Intel, but one area where I think that people pick up the most is the data center product, and particularly the Gaudi product line and then the GPU efforts as well. Can you just talk about, do you need a full reset here? Do you need to pivot in a different direction? Because when you see different iterations of product all have struggle, you feel as though maybe we need to start from scratch. Is that the direction you're going to go in AI, or are we going to keep kind of going down this path?
Going back to scratch means I have nothing to learn from. And so I look at Gaudi as kind of step one. There's some really good things about Gaudi that we're learning, particularly from the software and the platform level. But Gaudi does not allow me to get to the masses. It's not a GPU that's easily deployed in systems around the globe. And when you think about those that deploy Gaudi, it's from the largest hyperscalers to smaller customers that are deploying at the edge. And so we really need to think about how we go from Gaudi to our first generation of Falcon Shores, which is a GPU. And I'll tell you right now, is it going to be wonderful? No.
But it is a good first step in getting the platform done, learning from it, understanding how all that software is going to work, how the ecosystem is going to respond, so then we can very quickly iterate after that. If you just stop everything and you go back to doing an all-new product, products take a really long time to come to market. And so you're two to three years out from having something. So I'd rather have something that I can do in smaller volume, learn, iterate, and get better so that we can get there.
And obviously, AI is not going away. Obviously, training is the focus today, but there's inference opportunities in other places where there will be different needs from a hardware perspective. And so I'll be focused as well on where and how can we be competitive, where can we get our first foothold in that market, and then how can we grow from there. But I'll also be very honest with you. We'll be very pragmatic about how we do it. We're not going to throw hundreds of millions and billions of dollars at things that don't get traction. We need to fail quickly, learn, and iterate.
You talked about holding hands with several partners on this journey to start on the product side. On the foundry side, increasingly, we're hearing more announcements from you guys. Could you give us an update on the pipeline right now and how things are kind of trending over the last several months?
Yeah. I think we'll talk. We're not going to have these frequent long lifetime deal value updates because really they don't mean anything. I mean, unless we're getting revenue from customers, it's not that important in the greater scheme of things. But obviously, getting customers signed up, having them evaluate the process, getting a roadmap with them is important. As I talked about and we talked about on the previous earnings call, we're making progress there in terms of signing up customers. We do have additional customers on 18A. We're working through some, like I said, some RFQs with customers on 18A. So I think the progress has been relatively good. What tends to get missed in all of this is we also have the packaging business, that advanced packaging business that we're also trying to drive growth in.
We've actually gotten some early wins that are actually starting to show up in terms of revenue. That will be a meaningful part of revenue for Intel Foundry for next year. And we think we've got good differentiated technology there that customers want. So now it's about building that customer base. And then eventually, what I'd like to see is a lot of the customers we win on advanced packaging transition them to have 18A volume with us as well. So it's going to be a long road. I'm not suggesting that we got panacea here.
We've got to execute. Obviously, the profitability of that business is not where we want it to be. I think we will see some meaningful improvement next year, both in terms of gross margins and operating margins on Intel Foundry. As we kind of start to see these customers roll in in subsequent years, I think we'll start to see the business get to where our real goal, which was to get to break even midway through now and at the end of the decade.
With that, there's time, but I know that everyone in this room is wishing you the best. Thank you both very much for being here.
Yeah. Thank you. I appreciate it.
Thanks, everybody, for coming.