Intel Corporation (INTC) Earnings Call Transcript & Summary

Good morning, everyone. We're going to get started with the second day here at the DB Technology Conference. We're very, very pleased to have Intel's CEO, Pat Gelsinger, joining me up on stage today. Before we get started, I'm going to read the safe harbor statement. You'll see the long version up there if you have incredibly good eyes. So before we begin, please note that today's discussion may contain forward-looking statements that are subject to various risks and uncertainties and may reference non-GAAP financial measures. Please refer to Intel's most recent earnings release and annual report on Form 10-K and other filings with the SEC for more information on the risk factors that could cause actual results to differ materially and additional information on our non-GAAP financial measures, including reconciliations where appropriate to the corresponding GAAP financial measures. So with that fun stuff out of the way, we -- Pat's going to have a brief opening comment with the slide, and then we'll dive into Q&A. So Pat, thank you very much, and I'll hand it over to you.

Hey, thanks, Ross. Great to spend some time with you last night with dinner and some of the analysts. And thanks for the opportunity to provide our shareholders, potential shareholders with an update. And following our Q2 earnings, our August 1 earnings announcement, and it's been a difficult few weeks. And with that, we've been working hard to address the issues. And at earnings, we were determined to lay out a clear view, right, of where we were but also some of the next steps that we needed to address for the next phase of our strategy. And obviously, the market didn't respond positively. We understand that, and we're taking seriously what the market is telling us and hearing that clearly. But we're also staying very focused on delivering and executing on the large number of the things that we laid out. We described a set of cost reductions. I can tell you today, most of those are well underway already. Like everybody in the industry, we realize we have to operate efficiently, with nimbleness, with urgency. This is a competitive business and market. And that's part of the reason we took the actions that we did as we build it. And as you think about the turnaround story for Intel, I'd really like to say that we've now begun Phase 2. And Phase 1 of the transformation was this rebuilding of the company and being able to get back to process and product leadership and laying out our IDM 2.0 strategy. And I think about that as the most significant rebuilding of the company since the memory to microprocessor transition and this idea that we had to get back to process leadership. We described that as 5 Nodes In 4 Years journey, and we see the finish line in the sight. And since we released the fifth of those, right, the PDK for Intel 18A, we've seen a market uptick in activities in the industry. We now have about a dozen customers that are actively engaged with us around that PDK. We have 8 product tape-ins that we expect to finish by the middle of next year. And of course, Panther Lake and Clearwater Forest, our first client and server product. And I'm happy to update the audience that we're now -- for this as a production process, we're now below 0.4D0 as defect density. As we touched on last night, Ross, this is now a healthy process that we're looking forward, and we'll start production wafers with Panther Lake before the end of the year. So we're on track to deliver that. We're seeing an uptick in activity from the industry since the PDK 1.0 went out. And as we've also talked about in our advanced -- in our foundry strategy, our advanced packaging remains a huge differentiator. And since the beginning of the year, we have about a tripling of our deal pipeline in advanced packaging. So that continues to be the on-ramp for many of our foundry customers. In addition to that, of course, is rebuilding the product execution. They'll process products, and we're seeing now that we have competitive products. We've gone from many steppings to get to production. We're now delivering an A and B step consistently. And next week, we'll be launching Lunar Lake, the most compelling PC, AIPC product ever. And with that, we'll dispel some of the myths around the battery life of x86 versus ARM in a very declarative way. It simply is a great product. So Phase 1 was process product. Phase 2 has to move into now becoming efficient, implementing cost actions. And I realized when I took the job that we needed to take these actions. However, you couldn't go to the TD and product teams and say, "I need you to fundamentally reduce cost and get back to leadership simultaneously." And we chose get back to leadership first. And now we're focused on get efficient. And the clean sheet exercise that Dave and I began at the beginning of the year as we rolled out the new operating model as a result of the Q2 earnings, we just accelerated what we already had underway. And it's truly a clean sheet effort, best-in-class fab and foundry, best-in-class fabless company. And on the capital side, obviously, having completed 5 Nodes In 4 Years, getting Shell ahead. Now it's getting efficiency of those capital, right, investments and being able to start judging capital investments based on true market signals, where we are in our products, obviously, lower in revenue at the beginning of the year so we're able to make some reductions. But also with that Shell ahead, we're able to now respond to market signals from external foundry customers as well. And on the headcount and operations side, 15,000. These are friends. These are people that I've worked with for many years. And it's challenging to see 30- and 40-year veterans of the company, but they view it as the right thing in this phase. We're well underway, site reduction, portfolio reductions to build an efficient, competitive Intel. And we respect some of the skepticism that we've received from the market. I, my leadership team, the company, we believe we're up for the challenge. We're committed to do it. And as we go from memory to microprocessors to IDM 2.0, fab and fabless businesses, we're well underway to deliver what we said we would. So with that, Ross, let's go to whatever questions you might have.

Perfect. Thank you very much for that. So what I really want to try to do is talk a little bit about the near term and connect that to the IDM 2.0 transformation. So I think what investors are truly concerned about is the market will go up and down in the near term, and that's fine. But in Intel's case in the middle of the transformation, the concern is that it's a reflection of something more existential at the company. So I want to be able to kind of bridge the gap between those 2. So why don't we just start off in the near term, for the first half of this year roughly, you expected super seasonality in the back half of the year, and now it's sub-seasonality. What changed? And how much of that is a market-wide change versus Intel-specific?

Yes. And as we came through and started to look at the second half of the year, what we saw was markets were a little bit weak, some of our cyclical businesses. Auto has been reported, FPGA, some of those other businesses didn't come back as quickly as we thought they would. But we also saw in the PC industry and some indicated that was share loss. We believe that wasn't the case at all. Simply sell-through wasn't as strong as we thought on the PC. And I think some of the more recent signals says, yes, we probably got it right as we saw a little bit of inventory build with OEMs and channels. So we took our Q3 guidance to address that inventory a bit more aggressively. So we believe that our PC position remains strong for that. Servers, we're still in the environment where the AI build has continued to depress the overall server market. Now we are seeing some more green shoots associated with head nodes, with some refresh cycles that are getting more compelling. But that portion of the business hasn't been as robust as we would expect. And then we also have some unique issues with -- like China in the geopolitical sense, which has also been a bit less aggressive in its recovery than we would have expected at the beginning of the year. All of that taken together, though, we don't see there's not any like surprises inside of that, very understandable. And I think in some cases, like the inventory positions in the industry, we were the first ones to really understand what was going on given our large market share and our focus on sell-out, not just sell into the channel.

Got you. So on the revenue side, it sounds like it was more of a market dynamic and nothing specific to Intel in share loss or anything like that?

Yes. And I mean, we've been expecting to see a stronger server cycle on this. And now I view it some way so now that I have competitive products, again, we're ready to go, but the energy is in the AI server area. So that area has been a bit disappointing that we haven't seen more uptick there. But again, with good products, we're now starting to compete much more effectively.

And then the second part of the near term to the longer term connecting the dots I want to do is on the gross margin side. That seemed to be admittedly with lower revenues, of course, you're going to have lower gross margins, but it was a bit worse than that. So this seems like it's a bit more Intel-specific. Talk a little bit about why the gross margin, whether it's second quarter or second half, unfortunately, has a 3 handle again.

Yes. Yes, very unacceptable, I'd say. And pieces of that, obviously, we talked about, Meteor Lake, the new fabs in Ireland. And it really was, I'll say, we sort of had this confluence of moving out of a TD fab, which is higher cost, into a manufacturing fab, which will be lower cost once it comes down the learning curve, right? It's sort of like, oh, man, right? This is the right thing to accelerate Meteor Lake, and we're still aiming for the 40 million units this year, which looks good. So good momentum there. But early life of a big fab, you still have a lot of those depreciation costs at the low volume on the front end. So that was surprising. And then the fab well overachieved, right, its total wafer output. So we have more of those high-cost wafers sitting on the front end, which, again, is good news, right? It's going to move down the learning curve more rapidly in the cost curve, but it did markedly depress the margins. And as we're separating the company into foundry and fabless, some of the systems issues, right, weren't properly reflecting that in our forecast. So that was also a disappointment. But that was only a piece of the margin story. We also had some areas that -- where we had onetime effects in businesses that we were exiting, right, which impacted the margin plans that we had, time write-offs, et cetera, which were also affects that, further weaken the margin structure and then some other areas of new products, et cetera, that also contributed to it. But overall, unacceptable margins. And internally, we're putting just a lot of focus that, hey, we need to get back into the 4s very quickly and then back into the 5s. And obviously, some of our product lines need to be in the 6s.

So the last near-term question I'll have is kind of bringing those 2 things together. So part of the gross margin pain is too much goodness on the AIPC side. Meteor Lake and then Lunar Lake is going to ramp. How is it big enough to be bad for gross margin but not big enough to be good for revenues? How do people reconcile those 2?

Yes. And there is a piece of that, that as we look across the full stack of AIPCs as they come into the marketplace, we -- across the full range of our product line, we've actually seen pretty good ASPs on those high-end AIPCs. But across the mix, it hasn't yet been enough to really pull up the overall picture. And we did see some of the lower price points have pretty good strength. And obviously, we're here to maintain our market share. So when you blend all of that together, clearly wasn't as strong on the overall revenue picture as we would expect to become the case over time. And obviously, as we're now launching Lunar Lake, we have a very powerful product coming in there, good ASPs associated with that. And that is the revenue story that we've indicated that we do expect to see nice revenue expansion over time. And clearly, with a strong launch next week, the expectations of the rest of the year that we have as we start to ramp that and the continued Meteor Lake ramp, we do feel better about both ASPs and revenue growth as we go through Q4 and into next year.

So let's switch gears over to the restructuring side. What is the overall goal from an efficiency level, a profit margin, return metrics, what led to that being 15,000 and $10 billion as opposed to $5 billion more or $5 billion less. How was Intel thinking about it and even potentially splitting between the CapEx side and the OpEx side?

Yes. And what we did was we had a consultant on each side basically take a clean sheet view of best-in-class foundry, best-in-class fabless, right? And we use that as a mirror back to all of our existing businesses to say, okay, right, are you going to get to top quartile on each of these functions? Are you going to get to, right, best-in-class R&D efficiency on front-end design and back-end design? Are we going to get to being world-class in fabless metrics versus TSMC and best-in-class in terms of equipment utilization, of chemicals, right, labor costs, labor hours per wafer, just every metric. And that's what led to the goals that we laid out. Now the 15,000 as we view and some have said is that enough pad, is that not enough? We are committed to get through it quickly, but we also realize that we have portfolio actions that are beyond that. And some of those will continue longer because adjusting portfolios have a bit longer cycle associated with it. As I said, we're already -- the majority of those 15,000 actions are already identified. So we feel good that we'll get those finished. We'll complete some of the near-term portfolio actions and then get to the point that we can rebuild the energy, the momentum and the execution of the vision of the company. Some of the examples inside of this would be we've accumulated sites over our almost 60-year history, right? The sort of -- you just pick these up, okay, we're going to clean up all the sites. As we get to the other side, the 5 Nodes In 4 Years, our CapEx efficiency becomes a top priority. And as we've talked about a year ago, Ross, churning through Intel 7, Intel 4, Intel 3, Intel 20A, Intel 18A, you're not capital efficient when you're going through such a rapid catch-up period. Well, as we see the end of that get us on capital efficiency. We had no shell-ahead capacity. Okay, we now have shell-ahead capacity. So I can scale up with the more expensive equipment portion of it as we see clear product demand and foundry customer demand. So the reductions that we've taken in our CapEx budget are clearly aligned to a smaller Intel products revenue but also to a greater efficiency target that we believe, hey, we got to be best-in-class compared to it. And all of those efforts taken together, we feel like we're taking the actions necessary with haste to get the company where it needs to be. And I also say in some of these areas, these aren't issues of the last year or 2. In a number of cases, these are decades issues. We've been in IDM 1.0. We've had business processes. As we've talked about before, here, we're setting up 2 ERP systems for the company, a company that needs to be entirely manically focused on customer success and operational, right, on the fabless side and a company that needs to be innovative, efficient and capital-light on the fabless side, creating changes to the Intel culture that are very deep and longstanding. And we're well underway with the clean sheet effort to get those completed rapidly as we move into the second phase of the transformation.

So it's a good segue to the manufacturing side. You talked a little bit and last night, we talked about some of the defect density and a bunch of kind of nerdy stats and there are some [indiscernible] folks. When can investors expect to have positive proof that 18A is, in fact, on time, delivering what you expected at leadership or ahead of your competition, if you include backside power and all those sorts of things? When is that going to be -- when is the show me story going to be shown?

And clearly, for us, in our product portfolio, it's Panther Lake and Clearwater Forest, right? It's the 2 products that we'll be delivering, the less than 0.4 defect density that I cited early. That's a good yield at this point. And Ross, of course, I always have show and tell. Here's a Clearwater Forest, 288-Core, 18A product. It's a 18A top die with an Intel 3 base die, which is now a mature process as well, right? Intel Labs running, we'll get sampled to customers late this year. So you're going to start to hear some of the chatter coming from customers as they see it. We, of course, not to be outdone -- get to my -- reach to my other pocket. Here's a Panther Lake, powered on, looking healthy, 18A, bringing wafers back from TSMC. And we'll again be sampling -- beginning the sampling process to customers late this year. So our first 2 products are clear statements. And when I go talk to potential foundry customers, I walk into the room saying, "Hey, we're running this at volume," right? Your products wouldn't ramp until late '25, '26, '27. I am ramping my production wafer starts beginning on products like Panther Lake by the end of this year, right? We're well ahead of it. We're derisking it for you. As I also said in my initial comments, we now have a dozen active external designs that are under discussion, and we expect more market confirmation coming from that. As we said before, those customers, hey, they have big relationships with their current foundry providers, so they're reluctant to be public. But again, I expect that market chatter will start to reaffirm some of that as we go forward. And as I said, we now have 8 designs, some of them ourselves, some others that we'll be taking in, in the first half of next year. So I believe that we're now with the PDK 1.0, you're starting to hear EDA vendors, Synopsys, Cadence, reaffirming it, we're seeing -- and by the way, those defect numbers that I gave you, that's transistor and backside power, right, which is something that only Intel provides in the industry. So it's a composite defect density ratio. So we do feel like we're now finishing that 5 Nodes, 4-year journey, right? And again, Intel 7, Intel 4, Intel 3, 20A, 18A in 4 years. Hey, this is a remarkable achievement as we bring it across the line and super proud as we do it. But now we've got to go to the next phase, scale it in our products in the industry and more efficiently.

So that's a perfect segue to the foundry side of the equation. If you're approving with these products internally that 18A is working the foundry side, that should be a great carrot for your potential customers there. Talk about the pace of progress on the foundry side and going from where you are today to somewhere in the low to mid-single-digit billions of dollars and then eventually $15 billion, I think you have, of lifetime value. Talk about how long it's taking to get to where you want to go. And when will you be hitting those bigger numbers?

Yes. And let's start on the negative side, the 2 things that have surprised me in the foundry journey. One is that the hideous memories of COVID supply chains have diminished, right? And I was on the phone with Secretary Raimondo Department of Commerce, right? And she was lamenting that people aren't more concerned about the resilience of their supply chains, right? And I said, me too, right? It's like 3 years ago, oh my gosh, right? Now it's like, oh, we're okay with our Asian supply chain. So that's been surprising that, that hasn't remained higher priority, right, for customers just say, disappointing. The second piece that's been disappointing is just the -- we underestimated, I underestimated the amount of heavy lifting beyond producing good wafers the EDA, the IP ecosystem that needs to get enabled to bring designs on to the foundry. So those have been the 2 areas that in this current environment have been a bit harder than I would have expected. Now on the other side, the thing that's been a piece of super good news is the advanced packaging, right, where that's been the fast ramp into the foundry. I said we've seen a tripling in the amount of active deals that we have on the table this year for, we're at full capacity and ramping more capacity. Obviously, that's less CapEx intensive for our New Mexico facility. So that is ramping quickly. Our technologies, our highly differentiated technologies like EMIB are much more effective than technologies like CoWoS in terms of cost, scale, et cetera. So very differentiated technologies, a deal pipeline that's tripled over the year. We've converted many of those designs. And many of those customers are the active wafer customers that we now have underway. And the PDK 1.0 that we released in July, it was really the flip, right? And now the activity has expanded quite significantly. Since then, numerous designs, about a dozen potential customers in active discussions today, some of those with formal RFQ, some of those earlier in the design pipeline. The EDA partners are now aggressively engaging with their design customers and they look at this, oh, hey, that customer was only using me for TSMC, new revenue opportunities. So there's a lot of motivation on the EDA and the IP customers part. So we've really seen that energy pick up quite a lot since July. Now we have to bring those deals across the line. We have reserved a modest amount of our capacity for those customers. But with our shell ahead strategy, we're well positioned to expand as we see the designs materialize, and I am optimistic that we'll be delivering on the targets that we set.

So that was going to be my final question on that. So between the positives and negatives that you just mentioned, does that change anything about the breakeven and I think 27, midpoint of the 25 to 30 range, I think, is what you said earlier this year?

Yes. No change. We still believe that we can achieve those. Obviously, we're driving our teams quite aggressively to accomplish that. The $15 billion of long-term deal value that we have today gives us clear confidence of the $15 billion revenue by the end of the decade. The deal pipeline that we've seen expand on advanced packaging and now on wafers, I think there's plenty of business for us to go win to meet those objectives.

So let's switch gears to the AI side of things, and this might be the longest I've ever gone without talking about AI up on stage here. The PC side, we'll get to, but I think the bigger question is what is Intel's plan on the infrastructure side, whether it be on the CPU side, where I think the plan is a little clearer, or on the accelerator side? So just talk a little bit about how you play in that, especially given some of the GPU-centric folks that are doing billions and billions and billions of revenue.

And if we start on the CPU side, clearly, the evolution of AI, and I sort of view it as the high-end training and the at-scale cloud inferencing versus the enterprise and edge deployment. And if we separate those 2 worlds, clearly, as businesses are now saying, how do I go monetize AI into my business value, they're more on this side of the page. And today, 70% of computing is done in the cloud, 80-plus percent of data remains on-prem or in control of the enterprise. That's a pretty stark contrast when you think about it. So the mission-critical business data is over here, and all of the enthusiasm on AI is over here. And I will argue that, that data in the last 25 years of cloud hasn't moved to the cloud, and I don't think it's going to move to the cloud. And in fact, it's much more about how do you bring AI workloads to the data to those mission-critical use cases in the business. And obviously, that's an area that our strengths are much more substantial where the role of the CPU and even as you go think about some of the evolutions that we're seeing in the architecture of AI systems, Grace Hopper is what, a CPU and a GPU because this is where the data is. This is where the applications are, et cetera. So our position in the CPU, and we're starting to see businesses ask us, where's my TCO from AI? How do I get an ROI from AI? On this side of the page where our strengths become stronger, our Xeons are stronger, and our system platform of CPU plus GPU are stronger. So I'd say at that level, do we see ourselves competing over here? Not nearly as much, right? And I mean, that race, they are so far ahead some of the other challenges that we have. We're just not going to be competing anytime soon for high-end training. And as I view it in the 4-horse race on this side of the page, NVIDIA, build your own, Trainium, Inferentia, TPU, AMD and Intel's number, that's hard. But on this side of the page, we have a lot of strengths. And we are starting to see -- we have very high market share on head nodes, very high market share of people using AI workload on our Xeon. And that's one of those areas that we are seeing a bit of green shoots. Today, an enterprise-focused cloud vendor, IBM, announced their use of Gaudi, right, was announced earlier today as part of their cloud offerings. So remember, they're not your typical at-scale cloud. They're an enterprise cloud player. So I think that resonates with the framing of the market as I describe it. Now we -- as we've refurbished the CPU product line, with Xeon, Granite Rapids, Sierra Forest much more competitive. I just showed off my Clearwater Forest here, the next generation on 18A. So the products are getting better. We're seeing life in that area. And that's where I think the battle that we have the armament building on our x86 franchise to go be competitive.

So in the last 5 minutes, given that this audience tends to care about financials, let's talk about some of the financial implications and targets that you have. A year ago, we set up on stage, and you roughly talked about $100 billion in revenue and a 60-40 gross margin, operating margin 85 out of the 100 would be from products and 15 would be from foundry. Given what you're seeing on the revenue front, given what you've done on the restructuring front, how does that change?

Yes. And we still see the basic metrics that we've -- our goal, Dave and my goal is 60-40 this decade. We're off to go do that. Breakeven on foundry, 27, and we're off to go do that. Obviously, from a lower revenue base, though, we don't see that we're going to achieve those revenues. And we're doing all of our planning, and I'll call it the investment plan, assuming a lower growth rate on the products revenue, making sure that we've built a financial plan that gives our shareholders greater confidence. These guys are going to hit it. They have an investment plan that's lower on OpEx, lower on CapEx, and they'll increment up when we see true market demand associated with it. So get our cost structures, get our margin structures, get our profitability consistent with those metrics, assuming a more conservative view of the products revenue so that we don't disappoint our shareholders again, and that's our objective. So that's part of this clean sheet effort, right, is to assume a much more modest growth rate, build the investment plans and the only increment on top of those. So that's how the CapEx plan is being set, right, to a lower number, given that we've now built the shell ahead capacity, we can increment upward. We've also as part of those efforts leaning much harder into our supply chain and equipment providers. That's how you run great factories. How do you create long-term capital returns? We can't be peaking processes. We have to create more continuity. The equipment set from Intel 3 to 18A to 14A, 10A, how do we create factories that are much more flexible between those nodes? It's not going to be a factory that you ramp and then scale down moves, changes, ads need to become much more effective. And all of those things are the baseline of this clean sheet effort that we have underway. And we feel good that we can accomplish those broad metrics even with a lower revenue base. As we said, we're taking actions to continue to refine the portfolio. Some of those things will -- you'll see rolling out as we go through the year. For that, since I've been at the company, I've exited 11 businesses since I've been here as we've been taking steps in terms of spinning out Mobileye, Altera, IMS, the NAND business. And we have others that will allow us to become a simpler, more efficient, focused foundry and fabless company.

Why don't we wrap up with talking about gross margin? It's always been a very key metric for Intel and for semiconductor investors in general. Walk us through what the key milestones are to go from roughly 40% to your 60% target?

Yes. And this is a challenging period for that and my decision on Lunar Lake as an example. This is unquestionably the right thing for us to lead the AIPC. I declare the AIPC generation. Everything since then has been the right thing to go do it. We've seen the Qualcomm and other vectors there. So we took the challenging decision, take the Lunar Lake volumes up dramatically even though it was a high-end part that we're bringing into more mainstream price points. That depresses margins. So the best thing that we're going to be able to do to improve margins is move to the next-generation products, bring those wafers home. And that's exactly what Panther Lake is as we bring more of those wafers home. But that doesn't affect next year's margins very much, right? Since the ramp only starts late next year and really affects '26. So '24 and '25 are, I'll say, it's much more of a ground game, right? Every penny, every nickel, every test time, every element just grind out the improvements. And as we get to the 18A product generation that ramp significantly in '26, okay, happy day, bringing wafers home, the cost structure as we've done detailed analysis of 18A cost structure versus competitive nodes, we feel very good about the number of mass layers, the number of EUV layers, days per mass layer, defects and so on. So we feel like as we get through '24, '25, the margin picture has structural improvement, and in the near term, it really is about building more and more cost effectiveness into the business in every aspect how we price the products, how we deliver them through the channel, market segments where we have more pricing power versus more competitive challenges. So all of those together, we're quite committed to seeing that pathway, near-term, ground game, long-term structural improvements to get us where we need to on margins. And finally on this point, I'd also emphasize that as we get to a more normal cadence on our process technologies, all of the elements start working for us. You're able to run the nodes longer. You get better capital efficiency. And this move to get back to a modern EUV process technology, right, with multiple nodes running across that factory with much more flexibility, those all things are set up beautifully for Intel. And as we bring those wafers home, which represent a large portion of those wafers, right, that we're going to be able to bring home, all of that margin drop through to Intel uniquely in the industry. That's the plan that we've laid out as we go through the decade.

Great. Well, I could go on for another half hour, but unfortunately, I'm not allowed to. So we're out of time. Pat, thank you so much. Thank you for coming.

Thank you, Ross. Thank you all.

We appreciate it, and thank you to the great audience as well.

Very good.