Intel Corporation (INTC) Earnings Call Transcript & Summary

Good morning, everybody. I'm Ross Seymore, semiconductor analyst here at Deutsche Bank in the U.S. And I'm very, very pleased to have George Davis, the Chief Financial Officer from Intel, joining us for this next fireside chat. So good morning, George. Before we get started, Tony asked me to read your safe harbor statement. So today's discussion includes forward-looking statements, which are subject to risks and uncertainties. Please refer to Intel's SEC filings available at intc.com for more information on the risk factors that could cause actual results to differ materially. So hopefully, every aspect of this will be more exciting than what I just had to read.

We'll do our best.

Yes, exactly.

So why don't I start off with some high-level, more macro questions that I'm asking almost all companies first, and then we'll get into some Intel-specific things. And obviously, there's a going on at Intel right now, maybe more than I've seen in the last decade or 2. So lots to talk about. But first, let's just check the box on some of the macro dynamics. Shortages is the first one. So it's clearly an important topic this year. These have been industry-wide issues. You guys have talked about the substrate side of things. But just talk about where you are, especially considering the fact that you guys actually over the last couple of years, being an IDM have added significant capacity, but you -- yet you still have some shortages in the back half of this year. So just level set us on where we are today, when you think it's going to get better and kind of what the issues are between now and then.

Yes. I think the shortages come from 2 very different dynamics. The first is the dynamics that we have some -- what we're calling our superpower or the forces that are driving the industry, which is ubiquitous compute, pervasive connectivity, AI and then just the cloud-to-edge build-out. So if you look at that, these are forces that are driving increased -- significant increases in demand for semiconductors. Then you can pile -- you pile on that with the impact of the pandemic and all of the elements that -- both on the data center side and on the PC side, as the work-from-home, learn-from-home dynamics are playing out. So a big demand shift and the forces I talked about, we expect to continue for some time. But you have to couple that with the fact that we've been talking to you about shortages now for the last 3 to 4 years, where we were already struggling to keep up with the PC demand and at times, the cloud and data center demand that was in the marketplace. And so we're not only dealing with the demand dimensions, but also we've been short. And It's one of the reasons why we're talking about being in an investment period, and we can go into that a little bit more. But when you think about the CapEx requirements, not only are we seeing higher TAMs across the board but we're seeing the need to play catch up at the same time. So you're seeing a higher-than-normal capital that's required to address that. And we're seeing shortages continue, Ross. If you look at what our customers are telling us, their shortages are really about components that they need to complete their devices. And so they -- we've been able to meet the -- what I would call the suppressed demand because there's a pretty good inventory of CPUs in the marketplace. But other components are really gating what our customers can do right now. And that was a big factor in Q2. We talked a lot about Q3 being impacted by that, and we've continued to see it.

So when do you think the kind of the cyclical aspect, we'll get into your views and past views on PCs and that in my next question, but the cyclical aspect of it to the extent you can control it. Do you see that getting better as you ramp some of the substrate shortages internally and help some of your customers there so it will be better in the first half of next year from that perspective? Or what duration are you guys planning for?

We think it will continue to get better next year. I think we're right in the middle of it still this year. But maybe 1.5 years in total, as you think about how long does it really take to unwind most of these issues and to add the necessary capacity in the system. So I don't think it's over quickly, but everybody -- you can see everybody adding capacity. And as you know, if you're looking at it from a cyclical standpoint, the question is, will we overshoot? Or will we have just what's needed to normalize in 2023?

So last question on the supply side before we move a little bit to the demand side. A couple of years ago, the supply issues were unfortunately a little more Intel-specific. And then like I said earlier, you added significant supply last year and even the plan a year ago is to continue to add to it. But you actually lost some share in my view, at least partly because of just simply a lack of supply. What are your customers telling you now about Intel's ability to satisfy their demand relative to the competition? Is it now viewed to be something where your IDM stat is actually -- gives you a superior access to supply? Or are you just viewed as kind of being an equally problematic supplier to everyone else in the industry?

Well, I do think our IDM capabilities have given us some opportunities, for instance, in the substrate area, where we can take on some of the substrate processing at the end of the line, free up capacity to get more substrates overall. That's something that we can do because we have manufacturing capabilities, both in the front end and the back end. But I think everybody has been impacted in terms of share that they could achieve by the simple fact that we're still chasing to catch up with demand. Nobody could have foreseen the 2020 demand picture, and we were still chasing to catch up with the demand that had started to rise in '18 and '19. So I think we -- we're clearly focused on resolving that. We've talked a lot about the build out of Arizona. We've talked about the fact that we're going to aggressively reposition ourselves so that we're in a much stronger supply situation, more in line with demand. And we're going to be creating an environment that we used to have, which is we're going to build ahead on shell, at least the shell ahead of the demand picture that we see today. So we have much more optionality. Because if you have to start from greenfield to solve a problem, that adds a number of years. But having a shell that you can populate with equipment closer to the market is going to be an important opportunity for us, and that's part of our investment plan.

So before getting into IDM 2.0 and all the interesting dynamics there, the last somewhat more macro question I wanted to ask was on that PC demand side of things. Obviously, nobody could have forecasted what was going to happen last year from any number of vectors, but there's an incredibly active debate right now on just how sustainable that demand is. I know I had to buy PCs, laptops for kids at home. I was at home. Everybody did the same thing. But yet Intel, and Pat specifically, believes that there's a structural step-up in demand rather than just kind of a cyclical pop that could create a pause thereafter. What gives Intel the confidence in that structural acceleration and sustainable increase in demand in PCs?

Sure. No, it's a good question. A lot of people ask that question because it's been such a dynamic PC environment over the last year. First of all, just using your case, the density of PCs and the use case for PCs, we believe, has had a step up. And this is the hybrid environment that we're all operating in. You and I are both in our office today, but we probably spend more time in our home environment as do many other people. And even with schools reopening, the need for PCs for the learning environment has just gone up. So the PC density overall inside the house is up. Two, there's going to be a refresh cycle coming based on Windows 11 that we haven't started to see the effects of. And in particular, we think the refresh will be important on the enterprise side because you have an aging fleet as it were for existing PCs. And then, of course, there's new markets. And the new markets are -- the emerging markets are having the same experience in dealing with the COVID pandemic, which means that they're having to increase their density and the penetration of PCs in those markets. We think all of those things add up to a structural shift upwards in demand for PCs. And it's a little hard to see some of that as we're supply constrained right now. And so there is a backlog of demand. And I think the concern that people have is, is there going to be some sort of economic condition coming out of Delta and whatnot where you see a pullback? We're just not seeing that at this point. But we obviously will learn more over time. But our basic view is we have a lot of positive dynamics and tailwinds to support the step-up in demand for PCs over time.

And I assume, given the supply shortages, just the simple dynamic of even if PC demand at the end market slowed on the consumer side, one, enterprise probably picks up a little. But even if that didn't happen, the channel inventory is still exceedingly low. So the tail might be a bit longer on the stepped-up demand level anyway.

No, I think that's right.

So why don't we shift over to the new Intel under Pat's leadership? Talk a little bit about Pat rejoining earlier this year. What's been the biggest change you've seen strategically or culturally within the company since he took the helm? And I'll pivot right afterwards to IDM 2.0. So I know that's going to be part of your answer, but it doesn't -- I don't mean to front-run it with just that.

Okay. I'll save comments on that for that. Well, first off, Pat is an amazing addition, and he's been a great add as CEO. He knows the company incredibly well. He has immediately seen some of the key actions that we needed to take with respect to getting the product and process work done in a more predictable and accelerated way. Pat approaches everything with a mindset of torrid pace. And so we have adapted to support that. For instance, for Ann and her team doing the TV work, for Sunil and his team doing the products work, very strong expectations for where we need to see improvement and increase cycle time. And we back that up with increased funding. So we can take off the, I don't have the resources concern, and really drive towards product and process leadership again. That -- I think that is Pat's #1 focus, and he refers to it as execution. But his natural energy level is really invigorating for the engineering teams who've been a little frustrated that they haven't been able to have the same results that they've had for so many years. So -- and that has attracted really strong people from the outside to come in and upgrade our leadership capability. We have such a depth of engineering talent. Having the right leadership to take full advantage of that and get back on track is kind of the hallmark of what we've seen in the -- it's just been 6 months, it's -- with Pat, 6 months is 12 months. And so -- but that's exactly what the company needed and employees are -- you can tell are energized by that.

That's a great summary. Why don't we move over to IDM 2.0, which we could spend hours on this topic by itself. But just as a refresher for everybody on the line, it's kind of a 3-step process. First, the commitment to the IDM and internal manufacturing; second, use foundries where it's applicable for products; and then third, becoming a foundry. So why don't we dive into those in order? The commitment to being an IDM in the manufacturing side, at the Intel Accelerated event, you guys laid out the most detailed road map. It's a little bit of name change is that, forgive me if I use the old nomenclature, but the big message there was you were drawing a multitude of lines in the sands about exactly where you needed to be. While that's great to have that clarity, what gives you the confidence in delivering all this in such a short period of time after, unfortunately, many years of missteps in 10-nanometer most acute, but evenly, depending on how you measure it, the 6- to 12-month delay in the 7-nanometer soon to be known as Intel 4.

Yes. I mean part of what gives us confidence is seeing the results that we've been able to watch and laying out this road map with a very clear set of time lines is part of getting investors and others more comfortable that we're going to be able to execute to the things we say we're going to execute to, which has been a criticism of the company. So -- but let me jump back to the first part of your question, which is IDM. A year ago, there was a conversation -- a lot of conversation, is Intel going to have to consider going fabless? And I think what we saw -- first off, we felt that would have been a terrible mistake, and that foundry -- being an IDM is a very important part of our overall development capability and reinforces not only product excellence but process excellence. So we didn't believe that was the right answer. And quite frankly, the large majority of our investors didn't believe that was the right answer, which is why it didn't get any traction. It is -- Pat is thoroughly committed to greatness as an IDM, and that will feed into the -- how to think about foundry. But at the same time, when you're behind and you have the ability through the architecture of your chips with a disaggregated approach, we want to be pragmatic about using the -- whatever it takes to put out the best products for our customers. We're going to be open to working with external foundries if that, and in many cases, were on a device where you could see something from a TSMC and something from an Intel on the same device. But the combination makes for a better performance product with respect to the needs of our specific customers. So we're going to be very pragmatic and there -- in our use of external [ foundries ]. I'd say graphics is currently the most obvious example of that. But we'll see selective use of foundries as we go through the next several years. And so the -- both the internal and the foundry are just going to be a way of life going forward. And we've always had external foundries, but it's been on -- 99% is lagging nodes. Now you're starting to see more use of near leading-edge nodes with the foundries because we have the ability to tile that in and improve the product offerings over the next few years. So then where does foundry come in after that? And I would say, foundry is in the very early days. And obviously, we have a tremendous amount of interest from customers. And the interest is not only on the potential Intel represents on the leading edge, but also the lagging edge and packaging capabilities, which are much more attuned with the marketplace and marketplace capability. So I would say over the next few years, we'll exploit more of the lagging edge and the packaging opportunities for customers. But we've also announced that we're working with Qualcomm for instance on more advanced nodes. And the key there is as we close and we come back into parity and leadership in the '24, '25 time frame, you're going to see tremendous interest because everybody wants another capable foundry. And everybody wants more diversity of choice so we can offer a pretty wide range of technologies through our foundry. And as we close any gap on performance and density, it's attractive for people to have an alternative to what is largely today an Asian-based supply chain, and we would represent a very attractive Western component. But as you know, people want a leading-edge capability. And so the timing, we're going to be very smart about our investments in foundry to make sure that customers and the company are lined up on the technology and the capabilities that are needed at the time that they're needed. And I believe that's going to be a great long-term opportunity for the company.

So if I go back to the core to all of this succeeding, and IDM 2.0 is getting your own execution engine fixed from a manufacturing perspective. I think that's obvious to everyone. Like I said earlier, you laid out your road map. But what would you say are the benchmarks external investors can monitor to see that you're progressing? It was unfortunate earlier this year that Sapphire Rapids got slightly pushed out. And I don't mean to be airing your dirty laundry on that, but every one of those makes people consider Intel a show-me story. So what are you going to show us to give us the confidence that you are progressing towards fixing the manufacturing?

Well, we've given you a tremendous number of milestones, not the least of which is when is Intel 4 going to be in production on product? When is Intel 3 going to be in production? When do you see 20A? And so you're going to over -- and these are all happening in the next 1 to 2 years. And then you have the products that are coming on these that we're already talking about. In addition to Sapphire Rapids, you have Alder Lake coming. You have -- we've talked about our graphics capability and the timing for that. We've talked about having a, what they call infrastructure processing unit and the timing for that. So we're trying to lay out more of the road map, both on the product side and on the product side, so we can be held accountable to executing to those milestones. And you saw on Intel Accelerated a lot of specificity on these issues and with the people that are leading those projects. Because we think that's something that was a little bit too black box before, and people need to get more confident. And one of the things I would say that is an important factor in the way we think about can we catch up, can we move more aggressively to get to this node-to-node transition that we talked about? Yes. Number one, Intel for a long time had a very independent path. Its design methodology was Intel and the IPs were all Intel. And the ecosystem moved rapidly on the fabless side to provide very capable IP blocks. And you had Cadence and Synopsys and others that allowed people to plan transitions node-to-node-to-node with much higher confidence. And we weren't taking advantage of that the way others were. And so I think it's the realization that there's a lot in the ecosystem that would simplify our world. And if you don't have to basically update every IP block, every node and all the issues that come with that, when there is a fantastic ecosystem that is designing those capabilities, multiple generations out, you can actually simplify the process that we have to do to move node from node. We weren't doing that, and we're increasingly doing that now. That helps. EUV is a big factor. If you look at the push going -- being able to go to EUV for Intel 4 is going to be a great capability, and it simplifies the process, fewer process steps. We have the benefit of learning what the industry has already learned on EUV. When you're a little behind and catching up, you actually get the benefit of all the learnings, process tools. We used to operate more independently and would use our own thinking on how to optimize the tool, when the industry has optimized those tools for years, getting throughputs improved, getting the most capability out of the tools, this is something that is now becoming a part of the culture. These are all things that will help deal with some of the issues that we're getting in the way of our execution. Pat's -- like I said, Pat's primary focus is on execution to get our product and process capability where it needs to be. And we have some underlying factors that are going to help us there.

So we have about 10 minutes left. I'm going to go into a little bit more of a lightning round mode and get into some financial questions. And I know you have a financial analyst meeting in November so you're -- some of the questions you might not answer. But before I get into the specific financial one, you talked about parity and then leadership on manufacturing 2 to 3 years out kind of '24, '25. Why shouldn't we be concerned that the products between now and then will be inferior and therefore, share loss, pricing pressure, those sorts of negative headwinds won't persist until you get to that window?

Yes. What we've seen is there's been good reception for the products that we have. And CPU capability is incredibly important but it's not the only thing. It's built-for-purpose workloads. It's the -- having a platform capability like we have in our -- in the PC environment. Enterprise, being able to customize workloads and having the knowledge of our customers in a way that you can only have from having been the -- really the sole supplier for a long time for those customers, all of those things help bridge between today and the '23, '24, '25 time frame where you start to see the process and product capability come together in a way that is much more formidable. But the products are strong. Sapphire Rapids, very good reception from customers. So we'll get that into the marketplace in the first quarter. And we've got Alder Lake coming on desktop in the client. It's got a lot of features that people are looking forward to. So we're not -- we don't feel that we're on our heels when it comes to the products over the next couple of years. We feel like we've got the right products with the mix of capabilities that only an Intel can bring coming into the marketplace. But yes, we are excited about competing as well when we're at process parity and beyond.

So IDM 2.0, in a nutshell, accelerating your manufacturing side of things, getting back to parity and/or leadership position. You're also expanding capacity at the same time. So the good news, if you pull it off, that can be transformative for Intel. The bad news is it's expensive to pull off. Do you need government assistance and/or subsidies as part of this? Or is that a nice to have, but not a need to have to successfully pull this strategy off in an economically and kind of margin-acceptable fashion?

Yes. Well, first off, I think in terms of long-term value accretion, we're investing heavily to close whatever process and product gap there is, and that is incredibly valuable for our shareholders. Yes, we've had to ramp OpEx significantly for that, but that's for catch-up. And yes, our capital is higher because -- and by the way, over the next couple of years, most of that capital is all about catching up with shortfalls as opposed to some sort of speculative bet. We're going to have a very smart capital approach, which is let's make it so that we can ensure that our commitments to the big dollars are very close to market windows, so we can know that we have demand behind it and we know that we have customers behind it. And so most of what we're talking about and what you're seeing on investment today is really about catch-up and getting the capacity in place that we see for the core business requirements and investing to drive the node progression and the product process uplift to meet all that. That's where our investment is, and that's where Pat's -- is really focusing the company. Now, I'd still come back and when you look at foundry, the question will become, hey, when you're out in '24 and '25 and customers are seeing your process, allowing them to deliver their leading-edge products in line with other potential foundry players. It's the reason why we're seeing interest from customers today. They want to get -- it takes a long time to transition anyway. So let's start to learn together how we can best support them. And then the capital that would come with that is really going to be focused on when we both see the right window to come together, with them committing and us committing. But that will be -- again, it will be driven by, let's get to where we need to get to, and then we can make those decisions and then we'll be able to get the right kinds of returns on that product. There's no reasons why we can't get long-term leading-edge returns more akin to the leading-edge foundries today.

And is government, whatever you want to call it, incentives, subsidies a prerequisite or a nice to have?

I would say for foundry and for certain locations, I think they'll be -- it will be very important to get -- seek government support. And the fact is it's not a hard sell to governments. If you're in Europe or the U.S., you've seen that you've gone from over 80% share of semiconductor manufacturing in 1990, down to something -- I think we're -- let me just double check the numbers, but I think it's 12% in the U.S. today moving towards 9% and a very small percent in Europe. So you're seeing very strong interest for both the U.S. and Europe to seek leading-edge manufacturing capacity in their geographies. And so there's a natural alignment. We'll have to choose where to put facilities both for internal use and for foundry use over time. It's much earlier days for foundry and -- but we're -- we've obviously -- we expect to see the CHIPS Act passed and drive more government subsidy. It's going to be critical, particularly though for foundry, where we have to compete with customers or with foundries that are very highly subsidized by their governments, and that factors into their ability to compete. We have to get to a more level-playing field. And we've made that point to the government. If you want to have this kind of capability, particularly for foundries, then there's going to have to be much more support than historically. And we're seeing very receptive ears to that, both in the U.S. and in Europe, and in the U.S. on a bipartisan basis.

So George, we're running up against the end of our allotted time, but we have time for one more question from me, and I'll try to make it something relatively short for you to answer. Gross margins are obviously going to be a key metric you're going to highlight at your upcoming Analyst Meeting. I'm not going to get into the puts and takes and try to front run that, but what -- there's significant concern about exiting this year in the low to mid-50s and what it will do with all this spending, depreciation, pricing competition, entering a foundry market that's inherently less profitable in people's minds. What do you think is the most misunderstood part about the gross margin at Intel as you look forward, given those investor concerns today?

Yes. And obviously, we will talk a lot about this at the Analyst Day when we can put the whole picture in front of everybody. I think the -- I think people are underestimating how much of the near-term capital investment is really catch-up and the implications of that, which means that it will be supported by demand and the returns on that. But is there -- does increased investment affect depreciation? Absolutely. No question. The question is, are we investing in the right things? And if -- and do you believe that Intel is going to be able to make the process -- progress that we talked about at Intel Acceleration, which is a very exciting road map. You couple that with the products on top of it, as a company, there's a significant discount on valuation based on those elements. That's what we're most focused on addressing. Foundry in the near term has much less impact on anything to do with major capital investment. And so we are truly focused, and we're putting dollars behind it on driving the things that we believe will ultimately significantly increase the value of the company. And we'll make the case for that at Analyst Day.

Great. Well, George, we are officially out of time, but thank you so much for attending virtually today.

Yes.

Lots of interesting stuff going on at Intel, so we look forward to monitoring the progress, and we'll see you in November at your Analyst Meeting. So thanks again for attending.

Great. Thanks. Thanks, Ross.