Thanks for coming, everybody. I'm Chris Danely, your friendly neighborhood semiconductor analyst here at Citigroup. Next up, it's our pleasure to have one of the big kahunas of semis, very appropriate for a lunchtime keynote, Intel. And we have CFO extraordinaire, Dave "The Hammer" Zinsner, here to answer my questions, "The Hammer" Zinsner, answer my questions. First of all, there is Dave right up there with that excellent picture. Intel made us put that up there, not, not Dave, just to clear the air. I'm supposed to read the disclosure, so let me give it my best shot. 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, annual report, and Form 10-K, and all 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 comparable GAAP financial measures. Thank goodness, that's over! All right. So let's dig in. Now that I'm done auditioning for the auction house. I'd say one of the bigger positive surprises for Intel this year, Dave, has been you guys gaining back market share earlier than some of us would have forecasted on both the desktop and the notebook side. So let's talk about what triggered that and, you know, what we can expect going forward. Do you expect those share gains to continue?
Yeah. So first of all, thanks to everyone who gets a free... I never had a crowd like this. I mean, a free lunch will get a lot of people.
Maybe his wedding.
I'm gonna serve lunch every time I'm speaking now. Yeah, so, obviously, you know, we have had some challenges on the product side and on the process side. You know, Pat came in at the beginning of 2021, you know, job number one was to get that rectified. And on the client side, I think we've done a really good job of getting products to market in a timely fashion, you know, with good performance.
Mm-hmm.
You know, of course, we've got a really strong ecosystem, in particular, in the enterprise side of the client space. And that's helped us, you know, kind of get the share back. And I'm really excited for Meteor Lake when it comes out. We should qualify that this quarter. We'll have our pre-PRQ, and then, you know, ramping thereafter. And so, you know, that product has particularly good performance. It's the first tile architecture for us.
Mm-hmm.
It's also on Intel 4. So, you know, we're starting to get into EUV, like, you know, EUV processes for the product, and it has an NPU. So it actually is our first, you know, product on the client side that addresses AI in some respects. So, you know, I think that's helped out a lot. You know, we've gotten, I think, you know, much more focused in terms of providing value to customers, getting a stronger engagement with the customer base. I think that's also helped a lot in terms of rectifying things. I think at this point, you know, based on where we sit on share, we feel pretty good.
I mean, we feel like, you know, we've really righted the ship on the client side of the business. And now it's a matter of just continuing to execute on the products, you know, continue to bringing them out, you know, with new processes. And I think, you know, that business will largely be defined on how the market grows. We actually think 2024 is going to be a pretty good year for client, in particular because of the Windows refresh. And, you know, we still think that the install base is pretty old and does require a refresh. And, you know, we think next year may be the start of that, given the Windows catalyst. So, you know, we're optimistic about how things will play out, beginning in 2024.
Mm-hmm. So let's dig into that a little bit. I think what's really interesting to note is you guys raised pricing, and then your market share went up.
Yeah.
So do you think that any particular product on the client side really spurred that? And then maybe talk about some additional products down the line you think will, you know, give you another step function higher in terms of, share and performance as well.
Yeah, I mean, Alder Lake was a great part, when we introduced it and, you know, we, we knew it was a great part. We thought it drove good value to the customers. Generally speaking, our customers get a premium with our, with our particular, CPU. So, you know, we recognized that there were some, you know, cost-inflationary costs that, we incurred over the course of the last couple of years, and had not passed that, that on to customers. You know, given the, the performance of the part, the quality, the, you know, the, the... and so forth, you know, we thought it was a good time to, you know, even raise pricing, and still, you know, be able to drive a commanding share.
Mm-hmm. Great. And so switching gears a little bit to the server side, the death of your server business has been greatly exaggerated. You're still pulling in, you know, roughly three-quarters of the market out there. Why do you think, you know, your share has not eroded like some would have predicted, or even like what it's done in previous cycles when your main competitor grabbed the upper hand?
Yeah. Well, I, you know, I have to say, you know, even coming into this year, we would have expected, market share erosion. And in fact, in Q1 and Q2, you know, if you look at the surprises we had, relative to our guidance, I think that was in large part due to the fact that our share did better, on the data center side. You know, I think Sapphire Rapids has turned out to also be a pretty good product. You know, One area where it really distinguishes itself, is in AI workloads.
Mm-hmm.
And as it turns out, that just so happened to be the, you know, the place where, you know, things were really moving rapidly over the course of this year, and so I think it benefited from that fact. And actually, when you look at it, we were thinking we would need to see China come back in a big way, which would help out the enterprise side of the business, and we're better positioned even on the cloud space in China, and yet that has not happened. You know, we have not really seen China come back in any material way, and yet, you know, that business actually has performed better than expectations. That said, I don't think we're out of the woods yet in data center.
Sapphire is a good part, but it doesn't address customers' needs in all cases from a TCO perspective. And, you know, so we still have to get, you know, more products out. Emerald will come out later this year. That will be a good step in the right direction. But really, you know, it's when Sierra Forest and Granite come out in 2024 on Intel 3, that we - I think we really start turning the corner on a product performance basis. And then I think we have a real opportunity to improve our market position from there on.
So do you think that either Sierra Forest or Granite will start to result in you increasing market share? Do you-
I think it's stabilized. I think 2024 will be a year of stabilization, quite honestly. And then, as we get into the 2025 timeframe, you know, Granite is starting to really ramp in terms of customer traction. We start to see Clearwater Forest come out, then I think you start seeing... What we would expect is, you know, a better share position. We start to have real, you know, product performance advantages.
Great. In terms of other hot topics, I would probably be run out of town if I didn't ask about AI and Intel. Sounds like you guys have been talking about some nice traction for the Gaudi product. Maybe give us an update there-
Yeah
And then we'll move to other areas as well.
Yeah. So we have Gaudi 1 and Gaudi 2 out. And, you know, given the challenges in getting GPUs, I think we see more customers taking a look at Gaudi as an alternative. In addition, you know, the price points are better and more attractive. And when you look at in certain models, and in particular, as you look beyond the really big parameter training, you know, that requires scrubbing the whole internet, you start to get... You know, you look at it in a more contained environment, particularly on-prem searches.
It has definitely got a real, you know, performance level that is equivalent to what you see from competitors. So, you know, we think that we have an opportunity to really drive some good traction on that to and particularly together with our with Sapphire on the CPU side. We talked about on the earnings call that we had over $1 billion worth of pipeline.
Yeah.
You know, a pipeline is, you know, what you might imagine. It's, you know, just basically, you know, any customer, that we could potentially see some business from, who's expressed some interest, you know, we're counting. That then has to be converted to real dollars and, you know, of course, not all of that converts, to real revenue. But the pipeline itself is building every day, and it's meaningfully bigger now than it was when we talked about the billion-dollar number. I would suspect it'll be meaningfully bigger by earnings call relative to where we stand, today. So we continue to see, you know, that business be attractive, and that will really be our story in 2024, you know, is driving Gaudi together with CPUs.
Ultimately, you know, we'll have our Falcon Shores product, our GPU out, in 2025, and, you know, you know, a lot of that is building the software ecosystem. We think we provide some real advantages from a GPU perspective with that software ecosystem. You know, we've talked about this one API, this notion of an open, you know, platform-
Mm-hmm
That's really hardware-agnostic. We think that's where the industry wants to go, ultimately, and so, you know, that's, that's how we're going to, you know, support our customers as they build out their AI workload.
So would you expect Gaudi to ramp before Falcon Shores?
Absolutely. Gaudi is ramping today. We'll have Gaudi 3 out here in a reasonable period of time, and, you know, those products should ramp all through next year.
When do you think it will be, I guess, material revenue, let's say, hundreds of millions? You guys are pretty big, so let's say-
Yeah
Hundreds of millions.
I think next year will be material. I mean, we'll have revenue this year. You know, that's not negligible by any stretch, but you know, it takes time to get customers comfortable and having run the accelerators on their particular workloads. So you know, that takes months. But I would say, you know, if you look at—We stood up Boston Consulting Group. They wanted to do a large language model within their own environment, fully secure within their own environment.
But it, you know, it had significant amount of parameters, you know, that that it was searching and, and, you know, I think we did that in about 12 weeks, from the time we started talking to to Boston Consulting to the time they were actually running that. And by the way, I was just at a dinner with their couple of their consultants, and they're super excited about it. You know, they're seeing real opportunity to leverage this with their customer base. So I think, you know, this will be one of those things where it just has to build momentum. Every quarter we're gonna I think we'll be able to say we've built momentum and seen strength, and I think next year will be, you know, it'll be a material number.
Great. And I know you're a finance guy and a margin guy. What should we think of the margin profile for this product, for Gaudi?
Oh, margins are accretive. Yeah, this is a good business for us.
Great. Sorry, I'm just scribbling all this down. And is that manufactured at Intel now, or is that on foundry?
It's not, it's not... We're using Foundry,
Come over to Intel itself.
Yeah, I mean, really, what will happen is Gaudi will converge with Falcon Shores. It'll be one product offering. And I think, you know, you'll see over time our foundry will compete with, you know, other foundries for that business, along with others.
Yep. Yes, let's talk about Falcon Shores, and then graphics, as particularly as it pertains to AI. You know, maybe just give us an update on that, and then when would we expect, I guess, meaningful revenue in discrete graphics from you guys?
Yeah, you know, 2025 is when it'll be out. You know, obviously, it'll take time for it to ramp. And, you know, obviously, the software has got to be there, which is what we're developing-
Mm-hmm
As we speak. And, you know, obviously, that takes time as well. And, you know, as it relates to... And you mean discrete graphics in, like, you know, the client side? Is that what you're asking about?
Yeah. Exactly.
Yeah. Um-
Ponte Vecchio, and whatever follow-ons-
Yeah
Are after that.
So, on the discrete graphics side, you know, obviously, what we have strength in is, you know, integrated graphics-
Yeah
Even today. So we actually have a good, good position. You know, our discrete—we have a plan to build out the discrete graphics business over the course of the next couple of years. And, you know, I think we're leveraging the GPU technology across both the data center and the client space in a way that I think, you know, should be a pretty, pretty efficient way of developing those products. So, you know, all looks good.
Sounds good. So, a comment that your boss made last week, I would feel remiss if I didn't ask about that. You guys are-
He made a lot of comments.
Yeah. That's why you love him.
He was excited.
So yeah, he stated that the quarter was progressing better than expected, and that was, like, halfway through the quarter. So it must be a lot better than expected if he's already saying that.
Yeah.
Could you just talk about where you guys are seeing the upside?
Yeah. So what he said was that we are tracking above the midpoint. Of course, you know, quarter's not over yet. We gotta finish strong in order to be above the midpoint for the quarter. But, you know, things are going better than we had planned. We had expected client to be up, and client is, you know, up. So it's relatively consistent with what we thought coming into the quarter. I'd say on the networking business, the NEX business, we expected that to be weak, and that's played out largely as expected. Really where the surprise has been has been in the data center business, the DCAI business.
Now, we do expect that to be dipped down quarter-over-quarter, but I think it's safe to say that it's at least tracking better than we had anticipated. So it's maybe not as down as we had originally thought, coming into Q3.
And-
You know, why is that?
I was about to ask.
So, you know, China's still, you know, relatively soft, so it hasn't been from that. But, you know, Sapphire Rapids as a product has done better. And, you know, we'll see how as we exit the quarter, how we do in terms of market share. I suspect we will have done better than we had anticipated for the quarter, once everything is set. I think that's part of it.
Mm.
You know, and as I said, you know, Sapphire is, you know, a product that happens to be pretty good suited for, for AI workloads. So, you know, I'm sure we're seeing some benefit from that across both enterprise and cloud.
Great. So just so I have it straight, the better-than-expected was entirely from data center, and basically, everything else is about what-
Everything else is pretty much as expected-
Okay
Although doing different things, obviously.
Has this, you know, strength flowed into bookings as well for the future? Have you guys seen bookings start to shift?
Yeah. I mean, probably early to say how Q4 is evolving. I would say maybe the best thing is we felt like the second half of the year for 2023 was gonna be-
Yes
A strength for us relative to, the first half, and that is playing out as we anticipated, and even in, you know, surprising us a little bit to the upside.
Where would you be most excited about for 2024? You mentioned that, you know, you think your server data center market will stabilize. You expect to keep gaining share in desktop and notebook. How about the rest of the businesses? Let's leave Q4 out, but let's talk about next year.
Yeah. Yeah, I think, you know, data center, you know, given the new products, and the opportunities within AI for Gaudi, I'm excited about the data center business. On the client side, you know, that business doesn't necessarily, you know, grow at breakneck speeds, but given the opportunity for a refresh, you know, midway through the year or so, and, you know, how good the Meteor Lake product is, and, you know, and just, you know, the opportunity given its application to AI, I feel really good about the client business as well. Maybe go back. The other thing on the data center side is, you know, we have suffered through inventory reductions, you know, for several quarters. That's likely to continue through the rest of this year.
But I think after that, we'll have kinda worked our way through inventory, and so that business has I think an opportunity for a tailwind like client is seeing right now, where it just kind of moves up to what, you know, basically, instead of bleeding through inventory, our customers now have to buy everything from us, so we get that lift from that as well. NEX will likely have that similar dynamic. It might be a quarter or so later-
Mm-hmm.
Than data center, but you know, sometime next year, we should start to see a little bit of a lift from them. And I have to say, you know, the other one that I'm excited about is the foundry business. You know, while it's probably early to get a lot of loadings on wafers, particularly since a lot of it's gonna come.
Down payments count for foundry revenue, right?
But, you know, packaging is looking very good for us. And packaging, we can turn to revenue, you know, in a relatively short period of time, measured in months, not years, you know. Wafers, you know, it's years before you turn that into revenue, and we are getting a lot of traction. And I like the fact that, you know, not only are we getting revenue from it, which pays the bills, but also it's a great kind of lead-in for customers, where we, you know, get them engaged with us on a packaging side and then, you know, kind of convert them over to be customers on the, on the wafer side. And, you know, so I would. You know, it's, you know, relative to the size of our business, is it gonna move the needle significantly? No.
You know, but it's a good bit of business, and it definitely, I think, gives us inroads for later on the wafer side.
Great. So just to run through some of those, I guess, you know, how do we, reconcile, what, you know, certain large semiconductor companies are saying about AI, and all of the budgets are going towards graphics, and then, you know, obviously, you guys are seeing a little bit better-than-expected trends and-
Yeah.
You know, we don't really see a lot of the, you know, CSPs increasing CapEx. I know you guys talk to them all the time.
Yeah.
What, what, what do you think is going on? You think it's maybe, you know, not, not as bad as some had feared, or what's your sense?
Well, you know, look, I think that at the moment, you know, given these big large language models, 100 billion+ level parameters, you know, there's a need for heavy-duty GPUs to do that work. You know, we, I think, are a beneficiary of that because of the CPU that we have, but clearly, you know, the biggest beneficiary is the company that sells those GPUs. But, you know, over time, we view AI as a workload. It's not necessarily a market. And so, you know, if it's a workload, it's a workload over every market we serve. It's, you know, yes, on the cloud side, but also on the enterprise side, the clients side, at the edge, in the network.
You know, as you move farther away from the cloud, the opportunities for it get bigger, because, you know, the architecture that works for a cloud service provider is not the architecture that works for an enterprise customer in terms of the cost of these systems and the power usage of these systems. And so we think there's a lot of opportunity, for us to play in this space and generate revenue that, that will - that, that makes this a meaningful tailwind to our business, regardless of what, what market we're talking about. So I, I, I actually think that, this is a huge positive, for Intel.
Now, you know, in the near term, we acknowledge that, you know, there's a lot of dollars going to GPUs, you know, for large language models, and, you know, that definitely takes a little bit of the wind out of the sails of our data center business, and a part of the reason why we think, you know, Q3 and Q4 will be more muted than they have been in the past. But I think as we go into 2024, I think there's a real opportunity for us to see meaningful growth from AI.
Great. Before we dig into manufacturing, another announcement you guys made recently was signing up a foundry. I think it was a prepayment, a down payment, something like that, layaway, whatever, for 18A. So, can you expand upon that? Where—what kind of customer is it? When can we expect this to be delivered, and are they, you know, locking into a contract, or is it just we see the capacity coming on, and we wanna get first dibs? How exactly is-
Yeah, I would characterize it as, you know, we're engaged with several customers, and as we've talked about, we thought we could get a, what we call a whale, done.
Would this be a whale?
By the end of the year. This would be a whale.
Okay
By the end of the year, on 18A. And, and so one of these particular customers, was pushing us to, move a bit quicker, and yet, you know, we're not all lined up in terms of the definitive agreement and all that stuff. And so, the request was to what, what ultimately the request was driving was an acceleration of our Arizona build-out, you know, which, you know, of course, requires, you know, capital, but also a commitment you know, we needed to see a little bit of commitment from the customer to make sure that they were, they were really serious about, about the demand.
So they, as a kind of, I call it a show of good faith and, and to support us in terms of accelerating Arizona, you know, put up what Pat said was a meaningful prepayment. You could call it a down payment, too. I guess you could, just as, just as right. And so, you know, we're, we're gonna get going on on accelerating Arizona to get that out. 18A, you know, really isn't in, you know, in production, you know, for for foundry customers, probably until the 2025 timeframe. So, you know, it would have to be out a few years before we actually see the revenue from it. But, you know, in order to get going now, to get production out, we gotta accelerate what we're, what we're planning to do.
I look at it as a great endorsement of 18A. Obviously, they had to put 18A through its paces a bit to get comfortable that, you know, we would be the right partner. So I think it's a good validation of where we stand. That said, you know, we still don't have our .9 PDK yet out. That should be out pretty soon, and that, I think, is really where you start, you know, seeing customers get real serious around, "Okay, you know, now we've got a .9 PDK, we can kind of, you know, track our way to one.
It's time to start, you know, locking in agreements for the volume. And you know, the good thing about this business, I think we are, you know, it's you know, we're well-situated in terms of the dynamics that are going on across the industry. You know, we got geopolitical concerns. We went through a lot of supply chain issues in the past that, you know, weigh on the minds of customers. You know, there's some you know, competitiveness, you know, challenges that purchasing departments have. And so, you know, our goal is to, you know, help alleviate that in the industry and be, you know, kind of the Western supplier for wafers, for you know, for customers.
I think earlier in the year, you guys mentioned you were working on two potential large customers?
That's still the case, yeah.
So was this one of those two, or could there be?
Yeah, this is one of them.
And then should we expect something on the other customer by the end of the year? Do you think this helps your case, certainly with the other customer?
Yeah, maybe. I mean, look, you know, it's nothing's done until it's done, you know? So we're working with both customers to try to, you know, make sure they're comfortable with our process and our manufacturing and what the plan is, and get them comfortable with the terms, you know, that we can both live under. And, you know, if, you know, it's never over until, you know, we get all that stuff sorted out and then a document signed. I'm optimistic that both of them could be signed. I think we are absolutely committed to try to get one signed this year. And I just point out that in addition to that, we're likely to, I think, have some announcements on the packaging side as well.
You know, some of those might be overlaps, some of those might be other customers that we're engaged with.
Yeah. Certainly, we would think packaging would ramp a lot quicker than foundry.
Sure. Yeah.
Yeah. And what would be the margin profile in the packaging business?
You know, remarkably, you know, because this is advanced packaging, pretty get difficult to do and does drive performance advantages to customers, the margins are actually fairly good. Now, you know, all of the-
Corporate average, good or good close to corporate average, or-
In the foundry space, the margins are lower than the corporate average, you know? So obviously, that's, you know, you gotta take that into account. But, I mean, this will be accretive certainly to the foundry margins.
Got it. Yeah, so maybe just spend a minute on, you know, when we could expect meaningful revenue from... I mean, you already have, what, $1 billion for foundry or something like that. But in terms of the, the leading edge or the, the, the bigger customers, when we could expect meaningful revenues from that business, and then what's, what's sort of the business model, you know, when you guys are maybe not fully loaded, but let's say five years out, when we're looking at the foundry business, are we thinking, you know, 10% growth per year? And are we thinking, you know, 50% growth and 30% op margin, and twenty... But give us a sense of what the sort of the ideal business model would be in, say, five years.
So, you know, it will take time to ramp this thing. You know, it's gonna be relic- you know, somewhere in that kind of range for several years. But I do think, you know, by the time you start running into maybe 2026, 2027, you'll start to see some real real revenue that moves the needle. That's probably what to expect. You know, obviously, we've got to get the customers to sign up. But assuming we can do that, I think that's really where you'll start to see meaningful revenue. I'll get in trouble from Pitzer if I announce the model before he wants to announce the model. I would just tell you that these margins can be quite good.
You know, the gross margins are likely to be lower than the corporate average, but the OpEx intensity is likely to be much better than the corporate average. And so operating margins, margins actually can be pretty comparable-
Yep
To the overall model. You know, maybe not quite the 60/40 model, but certainly, you know, at a healthy level.
Sure. Before we hit manufacturing, we might as well talk about the long-term model. And by the way, this is intended to be interactive, so if anybody has any questions from the audience, it's hard to see, but feel free to raise your hand, and we'll come grab you with a mic, and you know, hopefully somebody will get my attention. You guys recently came out, not recently, but came out with sort of this growth model with you know, double-digit revenue growth is the goal. Maybe just run down how you guys got to that, and then what would you expect to drive that growth? What would be the higher growing areas and then some of the lower growing areas, and then-
Yeah, I mean, you know, obviously, you know, the model's heavily dependent on the TAM growth and, you know, how TAM grows over time. But I would say, you know, client, it's this kind of low- to mid-single-digit growth market. You know, data center with AI and so forth is certainly in the double digits. NEX, you know, last I looked, was kind of in the high single digits, maybe getting approaching double digits. Then we have what I call the wild card, which is, you know, the foundry business, which may be immeasurable, you know, given that it's. If it comes, it'll be ginormous, you know, in terms of its growth. And so that gives us a lot of opportunity to drive revenue.
But I have to say, you know, the way we're planning the business is more modest than that. You know, we're thinking, "Okay, let's do all those things necessary to drive that growth, but let's manage the business from a capital investment perspective and operating perspective, to be conservative around that." And so, a lot of that stuff then becomes upside if it shows up, but we haven't gotten ahead of our skis in terms of spending for the business. So that's been my approach to how to manage.
Okay.
On the margin side, you know, I've known you for 20 years in four different companies. So Intersil's margins went up 10 points, ADI's margins went up 10 points, Micron's margins went up 40 points or something like that. You know, the Intel gross margin range was, I think, you know, 55%-65%, and then it was 55%-60%. What should we think of as sort of the long-term gross margins for Intel? And then what will be the drivers to, you know, take it to the high end of the range or higher, or, you know, God forbid, what could, you know, disappoint and bring it to the lower end of the range?
Yeah, I mean, the trick of this is to drive a margin that is really good, but also, you know, takes into account the size of our business and the share we wanna have and so forth, and the fact that we wanna be in the foundry business, which, you know, doesn't generally operate at 60%+ gross margins, with one possible exception. And so, you know, I think the goal would be to drive a 60%, and Pat's talked about this model. It's driving 60% gross margins. That's probably the optimal spot for us.
50-60 kind of thing?
What's that?
50-60.
60. To try to be in the 60 range. Now, we're, you know, 15 percentage points below that, roughly, and so we've got a lot to do to cover the gap. And I think it's a couple of things. One, getting process leadership and getting products back to their rightful place, I think is number one most important thing we can do for the business to generate, you know, good profitability. That kind of is number one thing. Outside of that, there is a ton of efficiencies that we can get within the business.
And so, you know, we've talked about this thing where we're kind of separating out kind of the manufacturing, which is gonna look like a foundry business that provides wafers and back- and packaging services to not only external customers, but also our own internal customers. And so, you know, by doing that, we kind of isolate that business's P&L now and be able to measure that relative to to peers. And, you know, you're not gonna be surprised to hear that it's woefully inadequate. You know, it's not in a good place. But just having that clear view of how it operates will drive a lot of behavioral changes-
Yeah
At the company. You know, better loadings, you know, more intelligence around how we increase our capacity, less hot lots and, you know, unexpected wafer starts, a better dynamic of how we mix our business. You know, right now, they're constantly moving the mix around, which is not a, not an efficient way to run a fab that, you know, now supports multiple customers. You know, less sample activity, only the sample activity that we need. And so there's, like, billions of dollars of savings opportunity there that I think will greatly improve the business.
We'll also, once we get through this five nodes in four years, there's a huge amount of cost in start-up costs, billions of dollars incrementally higher than we normally operate in terms of start-up costs, that go away-
Yeah
And improve. And then, of course, you know, revenue cures a lot, and so as we improve our revenue, the fall through is quite significant. We should see good margin expansion from that as well.
Do you think just regaining process leadership should probably get you to 60, and then anything after that in terms of the costs, et cetera, would be well above that?
I think-
That's what we've written.
That's what you want, did you say?
We've written. What I want, too, yeah. I'm an American, you know, right? I want Intel to succeed.
Yeah.
I'm a semi guy.
Yeah.
No Intel, no daily job, no good.
I think getting leadership will be an important part of the gross margin improvement. You know, I think what we feel committed to is driving this 60% number. You know, if things turn out better than expected, then obviously we'll take it. But, you know, for now, I think you should think about it as the end goal is to get to 60%-
Yeah
As a company.
And then on the cost cuts, I think you're $3 billion into the $8 billion-$10 billion. Is that-
Mm-hmm
Correct?
That's correct.
Where would you say it's come from so far? It sounds like you're pretty confident that you can, you know, snag another $5 billion-$7 billion.
Yeah, you know that and, you know, I kind of listed off a bunch of areas where I thought we could get the $5 billion-$7 billion, where we have line of sight to. The $3 billion was mostly a, "Hey, you know, we're, we're in a down market. We need to spend less." And so it was about, you know, driving kind of overhead efficiency at the company, and also, I would say, you know, just driving austerity and, you know, eliminating pet projects and the like. So that's really where we caught most of the savings. I think now a lot of the savings is gonna be driven off of just driving efficiencies.
You know, just you know, operating more efficient fab, operating a more efficient engineering roadmap, you know, being more efficient in terms of our overhead, in terms of our go-to-market. It's gonna be those kind of driving efficiencies, which, you know, in one respect are, you know, more permanent, but take longer. You know, you kinda iterate your way through those improvements, and those usually take multiple years to show up. So, you know, the goal is to kind of exit 2025, you know, with most of that savings in the bag, and really feel it in 2026.
Okay, great. One other thing I wanted to ask was on product lines. You know, as you start cutting these costs, are you gonna have to jettison any of the product lines, or do you think it can just be all internal efficiencies?
Well, you know, and by the way, I think this is an area where I'd say we, but you know, and Pat in particular, we don't get enough credit for in terms of just you know, enhancing shareholder return is we're being very aggressive in terms of looking at places to unlock value where it may not make sense to fully operate that business within the four walls of Intel, and that you know, unlocking it to allow our owners to get to get a return from that is like top of mind every day. And so Mobileye is a perfect example. Well, the IMS business, the mask writing business, is another example of that.
You know, we're gonna transition the NUC business from Intel, and there will be others, I think, over time, where we'll just, you know, we'll have done a good job building it to a certain level, but, you know, to really unlock the value requires it to be more of an independent company.
Mm-hmm.
And that when... You know, we'll be very pragmatic about how we do that.
Great. Just to make sure we reiterate, so manufacturing is on track as far as the four nodes and five years, everything's fine?
five nodes and four years.
Five nodes and four years, sorry.
Yeah. Yeah, completely on track. Things are going well. Like I said, I would expect a 0.9 PDK here pretty soon on 18A.
Mm-hmm.
I already said Intel 7, done. Intel 4 is pretty much there. You know, we're ramping Meteor Lake in the factories as we speak. You know, we pulled in the timelines for both Sierra and Granite, and they're on Intel 3, so they look good. So, you know, I have to say, you know, this has been, you know, one of the most significant transformations we've made, and, and I think in large part, you know, the credit goes to Pat and the, you know, the engineering teams that literally are meeting on this thing on...
I sometimes attend some. Some of them are just completely over my head, but you know, on a weekly, sometimes daily basis, to get this right, to get products out on time, to meet customers' requirements, and to do it as cost-effectively as possible.
Great. All right, I think we're out of time. Thanks, everyone.
All right, thanks.
Appreciate it.