All right. Good morning, everybody. Let's get moving to the next fireside chat. I'm Melissa Weathers. I'm one of the lead analysts covering semis here at Deutsche Bank. This morning, we have the pleasure of hosting David Zinsner, Executive Vice President and Chief Financial Officer of Intel Corporation. Thanks for being here, David.
Thank you, Melissa.
I'll just kick us off with a very exciting safe harbor statement. Today's discussion may contain forward-looking statements that are subject to various risks and uncertainties and may reference Non-GAAP financial measures. Please defer 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 for additional information on Intel's Non-GAAP financial measures, including reconciliations where appropriate, to the corresponding GAAP financial measures.
Good job.
I think you guys have had a very busy month this month, so it's great to connect. I think if we could kick it off, can you talk about your historic equity offering that you did earlier this month? I think the final number was about $23 billion in capital raised. Talk about what was the reasoning behind that financing action, where's the capital going towards, and what does this say about how you're feeling about the core business at this point?
Yeah. The capital raise precedes capital investment. We talked about on the earnings call that we would be increasing CapEx significantly next year, and we had already upticked the CapEx in 2026 by a couple of billion dollars from $18 billion- $20 billion. We also talked about the fact that substrates are really constrained and we're going to need to make investments associated with substrates as well next year, given the significant demand. I think what you can take away from this is two things are happening. Number one, from an execution perspective, particularly on our process and also in terms of the advanced packaging, it's going extremely well. Intel 18A yields are progressing, beating the milestones that we've internally set.
Intel 14A, when you look at the defect density, is tracking better than the target curve we had for Intel 14A, and it's also doing better than any of the previous nodes in terms of how quickly we're bringing down the defects. In fact, we haven't seen this performance since 22 nanometer, which is arguably one of the best nodes Intel's ever put out. Things are going very well. So we feel very confident around process. We feel very confident around how we're performing in terms of advanced packaging, in particular EMIB-T. On the other side is also the demand dynamic, and we talked about on the earnings call, we're seeing significant demand in CPUs in data centers. We're seeing this ratio of GPU to CPU move more in the direction of CPUs. That's driving significant demand on our business.
We think that both Intel 18A and Intel 14A will have significant demand in terms of our own wafer usage for products, but also externally, we'll see demand there. We now have confidence in demand. We have confidence in terms of execution, in terms of process. It's the right time to think about increasing capital so that we can make the CapEx investments necessary to take advantage of that growth.
On that process node side, one of the things I found most exciting from your earnings call, Intel 14A, I think the 0.9 PDK comes out in October.
I think you also said on the call that you are now officially committing to high volume manufacturing for that node in 2028. Can you talk about what gave you confidence to specifically ramp up, commit to high volume manufacturing for Intel 14A?
Yeah. We are now within two months of the 0.9 PDK, so you can pretty much now expect us to land in October with that 0.9 PDK. That is pretty significant. Obviously, there is another version. There is a 1.0 PDK, but the step to go from 0.9 PDK- 1 PDK is not as significant as going from 0.5 PDK- 0.9 PDK. So, we have a significant amount of confidence there. I would say also, we have allowed the internal customers, let us call them, to somewhat choose their process, whether they go internal or external, what they do from an internal perspective. We are now seeing demand from our internal customers on Intel 14A. They are actually probably the most cynical bunch out of anybody, and the fact that they are now designing products on Intel 14A was a good confidence boost for us as well.
The engagements with customers externally, from a foundry perspective, has significantly increased. Lip-Bu and the team are now meeting on a weekly basis with customers. They are moving away from just looking at data to thinking about, "Well, how much capacity can I get? What does that supply look like?" We are now at a point where we have conviction around customers on 14A externally as well. Those things just have us now going, "Okay, we are going to be at the 0.9 PDK in October. We are going to now need to start doing risk production in 2027." We got to put the capital in place to be able to do that. We also, if we are going to ramp into high volume production in 2028 on 14A, given the lead times of products, we got to start putting out orders.
That effectively was one of the biggest drivers of the capital raise was putting all those things in place requires us to make some commitments to suppliers, and I need to know that we have got the cash to be able to make those commitments.
I want to get into some of those CapEx and budgeting discussions. Back to the foundry side, on the packaging business, this is a business that you guys have always had a pretty strong competitive positioning in. Can you talk about your competitive positioning at this point? How are those engagements going, things like EMIB-T? Can those engagements on the packaging side, can that kind of get your foot in the door on the front end?
EMIB-T is, as Moore's Law has somewhat struggled a bit on just moving down on a node basis, really advanced packaging is kind of a new angle of continuing Moore's Law progression. Our particular technology, Intel has made a lot of not only development investment, but research investment in advanced packaging. There is a ton of IP there. We have developed capabilities to do things that really just are not available elsewhere. We can increase the reticle size significantly, and that is an important component of what AI needs today. We think we have a very differentiated solution. The development of it has gone extremely well. We are already eating our own dog food with EMIB. EMIB-T being the extension of that, now we can take the interposer out, which makes it even more compelling as a product. You are right.
I would say just in terms of how that business probably evolves, I think we will likely see revenue start to ramp in the back half of 2027. Then it will start to become more of a, we will call it more of a run rate business in 2028 and be definitely hitting its stride in 2029. We should see for sure this business is in the billions. I would say by customer, it will be multiple U.S. billion dollars per customer per year of business. In its own right, it is going to be a great business. We think the margins, and I think at times people probably thought of advanced packaging as not being great in terms of margins, but I think these margins will be in the 40%.
They will be 40% gross, 30% operating margin business. It is not very capital-intensive relative to the front end, so the ROIC is phenomenal in that business. By the way, I would say maybe one of the better signals of the fact that it is going to be successful is Lip-Bu hired Seokhee, who was previous Chief Executive Officer of SK hynix, to come over. You do not get a Chief Executive Officer caliber person to lead one of these businesses unless that particular individual sees the growth opportunity in front of us. Anyway, I think it is going to be a great business on its own.
That said, yes, it is a great on-ramp vehicle to show how we can perform not only from an innovation perspective, but also just the blocking and tackling of operationally, how we provide the parts, when we provide them, what our yields look like in high volume. All of those things get tested in advanced packaging, and we win customers in advanced packaging, and I think there is a great opportunity to cross-sell them o n front end as well. Quite honestly, we've already seen that show up even now. As we do better, I think it only gets better for us.
Great. Let's switch over to the demand side of things. Clearly, the data center spending is off the charts. You talked about CPU density rising within the data center, which I think we're all very happy to see.
Yeah
The CPUs come back in vogue. You talked about strong double-digit unit growth in servers, I think this year and next. Can you flesh out a little bit more about what trends are you guys seeing? How big do you think this market could be? How supply constrained are you? Just how do we kind of frame that magnitude for us?
Yeah. There's a number of things going on on the data center space. As we move from training to inference and inference to agentic, the requirements for more CPUs goes up significantly. I think in a regular training data center, if you compare that to an agentic-like activity, it's like a four to six increase in CPU requirements in those dynamics. Obviously, that's driving a significant demand cycle for CPUs. That's somewhat of a core count unit dynamic. Before on the data center front, what we had is CPUs on a unit basis kind of declining, like low single digits, let's call it. But the core counts were still going up pretty significantly. So in general, things were moving up in terms of the growth rate of that business, but it was certainly held back by the unit growth. Now we're seeing unit growth now growing in this double-digit fashion and core count growing at multiples of that in terms of demand. On top of that, you have this overlay of the ASP per core. So ASP per core was declining.
Some years it was almost 20% ASP per core declines. So you were still seeing ASP increases because core count increases were growing faster. But it was holding back on the revenue. Now we've seen ASP per core at least stabilize. In some cases, I think we're seeing even ASP per core on a like-for-like basis going up. Now all the core count growth you get shows up in ASP growth. Now you have units growing as well. So this is going to be a phenomenal business. As you point out, the challenge is the supply side. In fact, as we look at this year, next year, probably even into 2028, this is going to be less about competitive performance of CPUs and more about can you get the CPUs to the customer.
Our advantage is that we own a lot of the manufacturing. We don't own all of it, and we have to go out and get substrates. In certain cases, we need other components. Memory is a companion, so we've got to worry about the memory in some cases. But we do own most of the manufacturing, so we can stress the manufacturing to try to drive better supply and meet the demand. I think share is going to be a function of how well everyone can do in terms of getting that supply. If you can do it, I think you're going to do well in terms of your share. If you can't, you're going to struggle. So what's important for us, most important, is Intel 3, it's Granite Rapids. Granite Rapids is our flagship product in data center. It's manufactured in Ireland on our Intel 3 process.
One of the reasons we are spending on CapEx next year is to ramp the capacity of Ireland, where we want to more than double the output there next year. I think we are positioned well to be able to supply. I think we will still be under-supplying. We will under-supply for sure this year. We will probably under-supply next year. But we will do our best to catch up as much as possible, and hopefully by 2028, we are in a good position.
Since you brought up supply.
Yeah.
Let us just dig in on that. Clean room space, that is the big hot topic in semis. Can you walk us through where do you have clean room available? What shells do you have? As we think about your different geographies, I know you have got something going on in Oregon, maybe some high-volume manufacturing there, Arizona, Ohio. Can you just walk us through the fab footprint?
Yeah. Well, I already talked about Ireland. Ireland, we have our shell. Now it is about equipping, and we are in the midst of ramping that output as we speak. We have Fab 52 in Arizona, which is running Intel 18A. There is another fab, Fab 62, that is almost ready, that we have to do some, and again, part of our investments next year are going to be about accelerating that space to get that equipped.
So we are in pretty good shape in the fact that we have that shell ready to go, and now it is just about equipping it. One of the things we will do, Oregon has always been more of a kind of get a node ready, get it up to a modest level of wafer starts, port it over to one of the other fabs as a high-volume fab, go back in Oregon, do it for the next process.
I think we are thinking about it a little bit more differently. Yes, we want a pilot line in Oregon. It will be the beginnings of every process for us. But once we are through that initial phase, we want to get things more into one or two centralized locations at high volume to drive the scale to get the cost structure to be better. So what we are doing with Oregon is we are going to try to push Intel 18A to be in Arizona as quickly as possible so we can free up all that space to make Intel 14A really in Oregon to start. So yes, it will be the pilot line for sure, but also we will run volume Intel 14A in Oregon as well. And so that is where we are making the investments on Intel 14A. And then, as you mentioned, we have Ohio.
We are working on the shells as we speak. There is the ability to have eight mods, or there is two mods per fab, so four fabs in Ohio. Mod one is the one obviously we are working on right now. If we could make it go faster, we would. We are going as fast as we can to get that one ready. So I feel like in general, we are in an okay shape in terms of space. We feel like we are in a good position with the vendors, where I think we are giving them good lead time line of sight into when we need equipment, and I think they are reacting really well to those requests. But there is also a limit for ours. There is only so much fab space we have. It is at a premium right now.
We will continue to make investments around space to make sure that we are positioned as we progress through not only this decade but into the next decade to be able to ramp supply with demand.
And a quick follow-up on Ireland. Those are internal-only nodes. It is impossible to do Intel 3 or Intel 4 with an external?
Let us say it this way. Intel 3 would not be a great logic process for external, but we use it for base die, and it can be used for base die for other companies. And the fact that we have advanced packaging and all these things kind of all under one roof, there could potentially be some opportunities there.
Interesting. Back to the demand side of things. I want to talk a little bit about client. As you are allocating all of this supply, memory prices are going up. It is tough for PC makers. Can you talk about how do you allocate these wafers between your data center customers and clients?
Yeah. We made a conscious decision almost at the beginning of the year, really. We kind of saw this coming. We knew that pricing and memory would get to a point where it was going to start to destroy some of the demand on the client side. We are not surprised by this at all. It was pretty much expected. We started to make shifts around, okay, we are going to focus our CPU capacity for client towards the big core type products. We would yield a lot of the small core stuff to others and try to shift as much of our supply over to data center as we could possibly get away with because we were getting good line of sight because of the long-term agreements with customers that we were going to have this big demand cycle on the data center side.
We've been doing this positioning to optimize as much as possible. I would say it's not a perfect puzzle that we can put all that stuff together. We talked about data center in Intel 3. Intel 3 doesn't have really much of client going on there. Meteor Lake was the only product that was going on in Ireland, and it's rolling off its life cycle as we ramp up Panther Lake. The opportunity to flex that really much is not really there. What we've got to do to get more output for Granite is just build more capacity and do better in terms of throughput and yields, and that's what we're focused on to get more supply. But we have, as much as possible, shifted out of Intel 7 on the client side into Intel 18A as we ramp Intel 18A. Intel 18A's obviously doing phenomenal at this point.
That frees up some capacity that we can use on Intel 7 to supply the data center market in addition to what we're doing with Granite. It's a balancing act, but that's how we're managing the different pieces to make it all work. In reality, the client business weakening was probably the best thing we could have, given the data center demand, because we needed as much to be freed up as possible to meet what's being required.
On the share side, is there anything to call out within client? I guess you're just trying to ship whatever you can, but
Yeah. Obviously, while doing all this, it is important to maintain strong share, particularly with the lead customers in that space. Like I said, I think Panther Lake is a killer product. So we've done very well with the customers in that space, and I think doing fine. Where we, I think, could do better is in the desktop arena, the high-end part of the desktop space. Arrow Lake was an okay product, and we did a refresh, but it hasn't addressed every concern that customers have. The good news is Nova Lake is going to be great. It's broader in terms of what it's addressing as we look at it relative to where we think competitors will be. We love it. We look really good there.
We'll see on the share side, but I feel really confident around how things are going to go for us next year on the product side. Then, of course, we'll see how demand goes. Clearly, this memory thing is probably going to be a hangover into next year. Again, probably not the worst thing because we're going to need it for the data center side of the business. Eventually, I think it'll snap back as we progress through 2027.
CCPG is the new name of the segment?
Why we should have a rule never to have four letters to an acronym.
Yeah
Because it's impossible. I was doing it on the earnings call, and I could not get CCPG out to save my life, but anyway, it is the new acronym.
Yeah. Somewhere in there, I think there's an edge business.
Yeah.
Somewhere in between the CC and the G.
Yeah.
It's something you called out on your call is edge and some of the trends that you're seeing. Obviously, it didn't get as much attention as some of the other parts of the business, but can you talk a little bit more about what you're seeing there?
Well, because it is small right now, b ut we have had a long-term business with a lot of customers in that space, and it was very profitable because we took products that we were developing for the client space, used them in Edge, so very little R&D associated with it. The margins are very good in there, and so profitability is very good. But it was not really growing that much. It was a relatively stable business. But now with AI, as AI becomes more distributed and heterogeneous across a lot of different applications, the edge is going to start to be an area that is going to be very interesting. Of course, everybody knows humanoid robots will be one of them. But the humanoid robot is going to be a small portion of the overall robotic requirements.
I mean, a lot of the robotics is an arm or it is like or some kind of other, just one single motion that needs to be more intelligent, quite honestly, for businesses to take better advantage of the physical AI space. So we think there is a tremendous opportunity there. We also think there is going to be a lot of opportunity in AI, just in compute, in industrial applications, agentic, and so forth. So we think this market is going to grow significantly, and I think we said it on the call, we could see this market being the same size as the client space, quite honestly, over a period that is not that long. So here we have this tremendous opportunity.
We have to take the capabilities that we have in client and bring them to the edge and physical AI. And we also have to think in a more system way about how to bring those devices. It needs more than just the component. You need to think about the software stack and things of that nature. That, obviously, was not a skill set necessarily we had internally. So Lip-Bu made the decision to bring in Alex, who does have a lot of that experience to think about how we can build out this capability so that we are better positioned to take advantage of it. We already have all the customer contacts. We are engaged with them in a way that I do not think any other player in the marketplace is.
It is about making sure that we have got the right applications, the right products, the right software stack, the right system-level thinking that those customers can take advantage of. And that is what Alex is working on. We have been having some of the conversations in the last few weeks, and I am super excited about his strategy and the way he is thinking about it, and how he is going to take a lot of the existing capabilities that we have within the company and kind of remold them into solutions that will, I think, be pretty compelling in that space.
I think, since he became Chief Financial Officer, Lip-Bu seems like he has injected a lot of creative thinking and a lot of these new initiatives. I guess, before we get to the financials. As he is thinking about his product lines, where is most of his energy being spent at this point in your?
I would say in the near term, just as a side joke, is now we're in a transitionary period where our new sales leader, Dean hasn't come on yet, so we're without a sales leader. Lip-Bu right now is actually spending a lot of his time on sales because he is the interim sales leader right now. I'm sure he's anxious to get Dean on board, so he doesn't have to think about that as much. What Lip-Bu spent a lot of his early months doing was just kind of getting the culture to where he wanted it to be. I think if you look back at Intel and the challenges we've had over the last decade plus, a lot of it can be boiled down to culture. Lip-Bu, I think, understood that.
He understood that at the board level, and as we were talking about him coming on, I think he knew right away he had to fix that. Part of the way he fixed that was management change. He brought in a lot of people that he trusted, that thought differently, more aggressive, kind of a take-no-prisoners mentality, but also knew how to operate in a leaner environment, be more successful the first time out, those kind of things. I think that was one kind of thrust he was working on. The second was really just eliminating a lot of bureaucracy. He did it in two ways, which was he collapsed the number of layers of the organization. He took it from 12 down to six. In doing so, kind of pulled a lot of middle management out, which bogged down a lot of the decision-making.
There was a lot of veto power across that organization, so a lot of things got slowed down, and it just led to suboptimal outcomes where we'd do product spins four or five times before a product was ready, instead of getting it out the first time. There are lots of startups that can get products out the start. Lip-Bu founded or invested in a lot of them, and we just weren't seeing that in a company that should be executing at way better pace than anybody that's got 50 people in a shop, and yet we weren't. He did that, and then the second thing he did was he eliminated a lot of the VP layer. I think we had probably at our peak, 450 VPs, and we are down to, I do not know, maybe 200. It is a massive change.
You see already some of the improvements in that. Like I said, we were not getting products out in the first, what they call A-stepping, like the first approach to the product. We now have had multiple products out in A-stepping. The execution around the process we have talked about, way better since he has been here. That was a key component. The third thing he did that I thought was brilliant was he really elevated the need for transparency. It seems like a very simple thing, but amazingly, Intel fell down a lot on transparency. A lot of people at one level knew what was going on, and by the time the PowerPoints got modified to where it was getting presented to the Chief Executive Officer, it was an entirely different story.
I think they thought they were doing the right thing, "Hey, I am not going to admit defeat," or whatever, but man, did it lead to a lot of bad decisions. It had the management team thinking we were going one way when we were actually going another way. Fixing all of that, and I can go into all kinds of anecdotes as how to fix the transparency, some of which was, I think, pretty painful for the people that were not being transparent at the time. It really has made a lot of difference in the organization. I think fixing culture was his number one job, and of course, that is an ongoing thing. You are never quite done with it. But I think he has really moved the needle on that as a company. Now it is about, okay, so he has got that done.
The second thing was, "Hey, I got to get these processes right." He went out and focused on that, and now we are on a good path there. Now he is in the third phase, in my view, which is on the product side. How can I bring winning products to the marketplace that are super compelling to customers, make a big difference in their own business applications, what have you? He has these thrusts. Obviously, data center, we are not where we want to be, but he is now identified Clearwater Forest as the product that he wants to put his fingerprints on to really make a difference. He is building this ASIC business to make customized silicon for customers, listening to customers figure out that he is working on that.
Some of that, by the way, is just cobbling together all the IP capabilities we have to bring them to market. Some of it is going out and finding other IP blocks that we need. One of the great things about Lip-Bu is he is so plugged into every small startup company out there that he knows where everything is getting developed, how it is getting developed, which ones are compelling, and thinking about how to. Some could be acquisitions, but it could be partnerships, it could be joint go-to-market type approaches. He is working a lot on that stuff to really make the ASIC business super compelling. Within Intel. Then, of course, I already talked about all the Edge and physical AI and so forth that he is working on. So I would say a lot of what he is doing is now geared towards getting the products right.
He gave us a little nugget at the end of the call last quarter about some interest in memory. I assume you are not going to become a memory maker, but any more comments you can make on any aspirations on the memory side?
I would put it this way. I escaped the memory space, so I was hoping he wouldn't say we're getting back into the DRAM space. I think his view is that memory is going to play an important role in AI workloads across all workloads, quite honestly. You're feeling it a lot, obviously, in the data center, in the hyperscaler space, but it's a challenge across the whole space. There are ways we can approach our products that can help customers in terms of the memory bottlenecks and the cost associated with memory. I think he views us as being a key player in making that all work. That includes working with the three major memory players out there, and he has great relationships with all three of them.
He is in regular contact, and we're in regular conversations with them on things we can do together. Then there's just ways you can develop product, architect the products we have to be better in memory. I think that's the way he thinks about it. There probably will be compelling products that really address this in a way, and I think that's when he said stay tuned at the end, if I'm not mistaken.
Yeah.
He's got some thoughts around how we can go out there and be differentiated and really help customers, and so stay tuned.
Well, another thing we are staying tuned for, I just want to make sure we ask this before we get into financials.
Yeah.
Analyst Day.
Yeah.
You have not had one in a long time. A lot of changes are happening. Any plans to announce one?
Yeah, that is a good question. Definitely, we have plans to do one. The question is when to do it. I am somewhat scarred from the first Analyst Day I had at Intel, because if you remember, I think I started midway through January, and I was doing an Analyst Day in February. That is not a great dynamic. You need time to get a sense for the business and come up with your thesis and your strategy. He is in the process of pulling in all these people, that to have any of them a month in, go and present to investors, I just do not think is fair.
So I think he is putting his finishing touches here with Dean coming in in September. So we have a couple of people that are relatively new in the company that are in leadership positions that we would want to put in front of investors. Alex would be another one. I want them to all get their sea legs, make sure they have the story down as to what they want to do with their particular function or business, and then we'll be ready. I don't think it's this year. For sure, it's not this year. I think it's sometime next year. We just got to, once we feel confident. The other thing is, Lip-Bu's in the middle of driving a transformation, and he's really riding the team hard to
go make those changes. It's fairly time-consuming to do an Investor Day. I just want to make sure we're past some of that, the team is all settled, and then Pitzer and I, John Pitzer and I will work out a date that makes sense, and we'll come out. We're not hiding it. It's something that we want to do. We just want to make sure it's the right time. Maybe one other thing I'd add on the Investor Day is, because I think investors are correlating this in some way with wins on the foundry side. It's not. In fact, even when we have the Analyst Day, you can rest assured Lip-Bu is not going to announce a customer because he's been very clear that he does not want to be announcing customers on the foundry side.
That's not done in the foundry industry, and we're not going to be an outlier. Obviously, if customers want to make announcements, they can do that, but we're not going to do that. So it's not in any way correlated to that, and you shouldn't expect, even when we have it, that we're going to come out and talk about customers in that space.
Yeah. That's helpful. I think I speak for all of us in the room that we'd rather have a good Analyst Day.
Yeah
Than a rushed one.
Yeah, of course.
And would much rather it be-
So would John. Mostly. Most of all.
So on the financials.
Gross margins, let's start there. A lot of moving pieces, a bunch of different things to track. I think you just printed a 40-ish or above 40.
Would you say you are now comfortably in the 40s? Or how do we think about stability of gross margins from here?
Yeah, that's a good question. Let me just say, when we came into the year of 2026, I think like a lot of companies, there is a plan, and then there is like, "Okay, let's try to push ourselves a little bit to do more." The plan submitted by the businesses was something with a three on it. It was a 30, high 30s gross margins. I have, in my career, always been hyper-focused on margins. I have improved margins in all three other semiconductor companies I have been Chief Financial Officer at. So to say that we were in the 30s was like, gut check.
One of the things I was really pushing on the team to do was to outperform that and to get ourselves into the 40s in terms of gross margin. I am actually pretty pleased with the execution. Some of it is things that were outside of our control. Revenue has obviously done better, and we have had some opportunity to do pricing, and that has obviously helped as well. But most of it was real roll-up-your-sleeve hard work to get gross margins to where we have them. I am really happy that, and I can safely say, I think, that we are now kind of comfortably in the 40s in gross margins. Low 40s, but we are comfortably in the 40s. Now it is about improving from there. We want to get into the mid-40s and the high 40s and ultimately something that starts with a five.
That is absolutely the goal. On the plus side of this thing, the things that we are really driving to improve margins, one of which is we have to get a better cost structure just across the board. On the product side, that is about just better use of silicon, less silicon as much as we can, make the packaging as cost-effective as possible, think about all the components and test times and stuff. We are working all of those things to drive a better cost structure for products and get improvement there. Obviously, as we move from older nodes that are not that great to nodes that are more competitive, have better yield profiles to them, we can do better in terms of throughput. We get margin lift from there.
The things that Foundry is doing to move down the node curve and improve scale are absolutely going to help on gross margins as well. Those two things are, I think, going to be the positive. The one thing that could weigh it a little bit, and why we maybe are a little bit more reticent to throw out, "We are going to be above 50% at X date," is some of our businesses actually do not get 50%+ gross margins, but they are really high-growth businesses.
The Foundry businesses, both from the front end and the back end, advanced packaging, are probably 40% gross margin businesses, at least for now. Maybe Foundry, the front end over time, does better as we become more competitive. But I think we have got to recognize we got to earn our keep here first, and so I think it is right to think about that as more like a 40% gross margin business. Still has great operating margins, still be in the 30s, or 30, but still, I think weighs down the overall corporate gross margins a bit.
The other is the ASIC business. Depending on the products you win, some of them do have a higher margin profile, but a lot of them actually are kind of 40s, 40-ish kind of gross margin type products. Depending on how successful we are there, that also could put some pressure on gross margin. Lip-Bu and I have been thinking about this and, okay, we obviously want the most profit possible. We also want the most growth as possible. How do we drive each business to do the optimal level that we can get out of them? Lip-Bu comes from Cadence, more of a software thinking, and in the software space, they have this Rule of 40 kind of thing.
We said, "Okay, we should have a rule of then for ourselves, so why don't we call it the Rule of 45?" We kind of backed into what should be reasonable to expect over time. That then allows, hey, if you're a low GDP plus type growth business, okay, then you better be delivering really great gross margins, really great operating margins, to get to the Rule of 45, because we're adding the revenue growth plus the operating margin to get to a 45 number. But if you're a business that's growing like a weed, then okay, fine. We can make some investment, both in terms of maybe pricing or in the cost structure to get the product to where we want it to be, or we'll invest more in terms of operating expenses to drive better growth. Fine.
You can yield a better growth rate than you can yield a higher spend level and still get to the Rule of 45. We really told every business unit, "This is the way we want you to think about it.
Is every business out of the gate next year going to be hitting the Rule of 45? Probably not. There's some things that have got to be fixed in certain cases. But every one of them has embraced it. Every one of them has shown me a long-range plan that achieves it. I think we will be driving the right behavior across the businesses and holding them all accountable to this number. Lip, when we talked about this in the capital raise, this was one of his main slides, I think, was the Rule of 45. He's embraced this. He's super supportive of it. Of course, it's like the best thing a Chief Financial Officer can hear is-
Yeah
everybody's going to be tasked with this kind of measure.
The last couple minutes that we have,
Yeah
the timing to Foundry break even.
Yeah
Can you remind us what the
How you guys are thinking about that? Yeah. Okay. We got to hope that none of the foundry guys, because I'm going to give a little bit of a double answer to this. We are driving the foundry business to go breakeven by the end of 2027. Mm-hmm. That's their internal target. That's what we're driving them to. When I talk to Naga and his leadership team, that's all he hears is me telling him this. That said, I do recognize that if you are more successful on foundry, it has a cost to it. The more customers we've got to port over, that requires some investment on our part. Mm-hmm. The more demand we have, the more startup expenses we will likely see. That could elongate the time to breakeven. There's no question about that.
I could see us going into 2028 and not being there yet, and maybe it's the end of 2028 even that we get there. But it will have been for a good reason, because it will be that they're more successful, ergo, they need to make the investments associated with that. That said, I think as you look at operating profit, like in absolute dollar terms, it's been running at like a $2.5 billion loss or so per quarter. And what we're also doing is just, regardless of whether you cross over at the end of 2027 or you cross over sometime in 2028, I want to see relative steady improvement every quarter. Mm-hmm. And I think you will see that it could be lumpy, obviously, not every quarter is going to work out exactly this way.
But in general, I want them kind of just fighting it out to get better profitability every quarter. And I think that also will be helpful and value-creating for investors. The combination of we win a bunch of businesses, maybe that elongates the time to break even, but you see the steady improvement in operating profit, and I think that alone should be helpful.
Last couple seconds that we have that I have to squeeze in.
Okay.
CapEx.
Yeah.
Any guideposts how to think about how you're thinking about CapEx going into-
What's significant is?
Yeah, I mean, I said that's insignificant. I'll be honest with you, we haven't completely locked the number in for next year. It doesn't get done until towards the end of the year. Mm-hmm. And quite honestly, of course they've come out with a number.
They've told me what the number is they think they should spend, and I've told them no. They're going back to rethink how they can be more efficient. I'd rather hold them to that, drive more efficiency, and then hopefully by the end of the year, we'll have figured out exactly how to be as efficient as possible, and I can give a better indication of the number.
Perfect. All right. Thank you so much.
Thanks, Melissa.