Intel Corporation (INTC) Earnings Call Transcript & Summary

All right. Welcome back, everybody. I'm Joe Moore, Morgan Stanley semiconductor analyst. Very happy to have with us today the CFO of Intel, Dave Zinsner, under somewhat better circumstances than last year. I upgraded the stock this time right before the conference, so that was good.

You learned your lesson.

Exactly. I just want to read our quick hedge, and I think Dave does as well, and then we'll go right into it. For important disclosures, please see the Morgan Stanley research disclosure website at morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative.

And then I'll read mine or -- okay, which is not there. He gets to read mine, and I have to do it by heart. So we may make forward-looking statements. And to the extent we do, please refer to the SEC filings, most recent 10-Q and 10-K for risk factors that relate to those disclosures.

Right. So with that out of the way, I guess I'm going to kind of go -- there's a lot to go through here. I want to talk about the recent dividend decision, near-term results, your cost-cutting strategy and then kind of come back to the general growth initiatives. In terms of dividend, the decision to reduce the dividend by 60%, I know that's not something you took lightly. Can you just kind of walk me through the decision-making process, both in terms of cutting it but also in terms of maintaining the portion of that's remaining now.

Yes. I mean we talked a lot to the Board about this over the course of a few months. Obviously, we did not expect the macro to be as weak as it turned out to be in the first quarter. And we -- while we have a lot of confidence around our road map, our strategy, how we're progressing in terms of our milestones, we -- I got to manage the balance sheet in a bit of a more conservative fashion. And when you looked at the amount of investment we needed to make to drive the transformation and you stress tested that and looked at what that did in terms of leverage ratios, it just got to numbers that I didn't feel real comfortable with. So I talked to the Board about it, and then there was a debate as to what made the most sense. I think the Board generally felt like a competitive dividend was important. I mean we certainly have the financial wherewithal to pay a dividend. It's just does it makes sense to pay the level of dividend we had been paying when we have some significant investments to make to drive the transformation. So they ultimately got to this number that was a reasonable reduction for us to manage the balance sheet but was still competitive as it relates to the peer group within the semiconductor space.

Yes. I mean you have a long-term plan that, if you can execute to it, the dividend is a very small component of the value that you'll have, so it kind of makes sense to me. Can you talk about how you're thinking about cash flow and balance sheet objectives? What are the boundaries of gross cash, credit ratings and things like that, that you're -- as you go through this investment phase of your transition that you're thinking about?

Yes. I think from a -- let's start maybe start from a ratings perspective. It's super important for us to maintain an A rating kind of category. We've been able to do that. And that's -- again, that's partially what went into the decisions around the dividend was to drive this A rating kind of characteristic of the balance sheet. We're in a period, I think, through '23, '24 and '25 -- or sorry, '22, '23, '24 where we expect to make meaningful investments to drive the transformation, so obviously, that is going to affect our cash flow. So we really see these next -- this last year plus these next 2 years as really more of an investment as opposed to generating free cash flow. After that, though, as we get back to leadership on process technology, that drives a lot of goodness in terms of what we can command in terms of pricing, what we get in terms of cost. We're kind of through the knothole of having to invest a lot to drive these 5 nodes in 4 years. We pulled back and drive to a more normalized cadence. As you mentioned, I'm driving a lot of cost reduction initiatives in the company. We'll drive about $3 billion of cost reductions, spending reductions this year, on the way towards $8 billion to $10 billion as we exit '25. So I think all of those things mean that as we get out of this, what we call the investment period, we should be back into a place where we're generating very good cash flow. Our model is to be in this kind of 20% cash flow generation as a percentage of revenue.

Great. And on that topic, can you talk about Smart Capital and your ability to recapture some of the CapEx both through some of the investment initiatives that you've done with partnerships as well as through government actions?

Yes. So Smart Cap is going to be a key component of this because, obviously, we've got to build out -- we've got to get through these 5 nodes in 4 years. We've got to get -- we had underinvested in clean room space, and you never want to be caught flat-footed in terms of not having enough clean room space, so we have to make a significant amount of investment to catch up to stage clean rooms. And obviously, that drives a lot of CapEx intensity. So we have kind of followed this approach of Smart Capital, which is actually several different strategies kind of bundled in one. One is to get -- to secure government incentives, government offsets. CHIPS Act was a big a big part of our strategy. Happy to see that get approved, and now it looks like some of kind of the mechanics of how to apply for that is now starting to become clearer. It also included the investment tax credit. 25% of our CapEx invested in the U.S. Our expectation is we get that back in terms of refundable tax credit. There's a European version of the CHIPS Act, and we're already in discussions with our European partners on offsets there to help contribute to the capital investments. And then we have things like Brookfield that was a key part of the strategy that we introduced this year. We'll spend about $30 billion building out the next Arizona fab. About half of that will come from Intel. About half of that will come from Brookfield. It will obviously come with an expectation for Brookfield that they get a return, but the return is actually pretty modest with respect to our cost of equity and cost of capital. And so that's another component of how we will kind of offset that. So when I talked about it at Investor Day, we thought that we'd have a relatively modest year this year '23 in terms of capital offsets measured in the single billion kind of level. I think at this point now, we're thinking that the capital offsets for this year should be in the 20% to 30% of gross capital, which is what enables us to not only hit the kind of mid-30s percent CapEx intensity, which is where we thought we would be during the capital-intensive period of '22, '23 and '24, but even perhaps see it go a bit lower than that. And so our outlook now is kind of in the low 30s. And that's all despite the fact that revenue obviously is at a different level than we expected it to be when we were talking about it at Investor Day last year.

Yes, that makes sense. And then in terms of just thinking about the CapEx, obviously, the 5 nodes in 4 years is something that you're going to do, and I have a lot of faith in your guys' ability to execute to that. How much of the CapEx is just going to be required to get that done versus building out capacity of those nodes? You mentioned having shells. Do you have flexibility to flex that number up or down depending on your progress with foundry and other initiatives?

Well, I've kind of -- I would bucket in like 3 components. There is the investments necessary to drive the 5 nodes in 4 years that absolutely we don't want to touch. That's important for a number of reasons, and it really drives the entire strategy. So 5 nodes in 4 years has to happen. We will make the capital investments necessary to go make that happen. Then the next bucket of CapEx is the clean room space. And that also, we have a heavy bias towards wanting to invest there because it's always cheaper to build the clean room than it is to equip it. And there's a really long lead time to actually building a fab. So you're talking about a 4- or 5-year period to have construction done. And it's hard to say what will happen in 4 to 5 years. So having a few empty shells ready to be equipped when we need it, our bias will be to drive that investment. The rest of the capital, which is equipping those clean rooms, is really where we flex. And we're essentially looking at wafer demand out over 5-, 10-year periods node by node and looking at what percentage of the share we expect to get of those nodes, both from an internal perspective in our own products and from an external perspective in terms of foundry customers. And we kind of plot that out. And of course, we have some scenarios. And that's what we use to determine and flex our CapEx, what I would call our capacity level CapEx. And that's what we have constrained in the near term. Obviously, demand has changed. So we have constrained that a bit more. That's enabled us to stay within the envelope of the CapEx intensity number that we promised.

Great. So shifting to the near-term environment, I think when you announced the dividend change, you reiterated Q1 guidance. But I know this is kind of not where you want to be from a revenue standpoint. I guess as you look at the challenge you're seeing now, it seems clear to me that you're under-shipping in demand and that you're burning inventory. I kind of thought that in Q4, but I guess, I guess maybe not in every market at least. Can you just talk to that? What's your line of sight to how much inventory there is and when it might be burned off?

Yes. I mean it's different by market, for sure. And we have varying levels of visibility into the inventory levels of customers, depending on the market. Where we have the best view is in the kind of PC OEM space because we get somewhat detailed views of what the inventory looks like. And it has been burning. And actually, in the second quarter of last year, it actually -- we had a meaningful burn. And then the third and fourth quarter were more modest but definitely burning. This quarter, we expect a pretty meaningful burn. And likelihood is that it continues to burn into the second quarter. I think at that point, we're probably through inventory burns. I mean, a lot has been somewhat dependent on what we thought the end demand looked like. And obviously, we weren't expecting kind of a client unit volume to come down like it has this year. And so what looked like a good level of inventory suddenly becomes not so good when demand has softened. But I think demand is generally stabilized. Based on talking to OEMs, I think we feel pretty confident that this will be a very big quarter in terms of burn and then maybe a bit more modest in the second quarter. And then I think we will be through it. There's probably were a little bit -- there's probably a little bit more to go in the data center space versus the client space, but I think largely through that in the first half of the year as well.

And the data center, particularly in the hyperscale space, that's a matter of budget cuts or inventory or both?

I mean I think the budget cuts have actually certainly muted the growth for this year, but it hasn't like -- that hasn't been coming down. Like, the clients -- I think it depends on the number you quote, but it was roughly, call it, 300 million-unit market last year. And it's certainly -- we've given a range of 270 million to 290 million with a heavy bias towards 270 million. It's definitely biased towards the 270 million. So that's been a totally different dynamic than the data center space, which is more of a, what I call, kind of tap the brakes kind of market.

So in that context, I mean, I think it makes sense that you didn't guide to full year revenue. But I mean, it seems like there's a general shape of the first half is where the difficulty is kind of concentrated and then...

Yes, I think that's fair to say. I mean, our expectation would be first half is likely to be tough as we guided in the first quarter. But our expectation is that the back half of the year, if the macro kind of hangs in there, we know that some of our markets were more exposed to China, were more exposed to enterprise and the data center space. Those 2 markets got more impacted than some of the other parts of the cloud space. As those markets come back, as we start to normalize back to what end demand looks like, the expectation would be we'd have a better back half of the year than the front half of the year.

Okay. That makes sense. So again, in terms of the cost cuts, now that you guys have talked about for this year and then long term the headline has been $3 billion of cost cuts this year and $8 billion to $10 billion exiting 2025, understanding there's a lot of moving pieces now with the depreciation terms and things like that, what's the benchmark of that savings? I mean, am I looking at here was your cost in 2022, and we're going to reduce that by $3 billion. Is that where -- versus the trajectory that you were on before. Just how should we think about measuring that?

And to be clear, the depreciation change does not, in any way, count towards our savings. Our savings has to be really as real...

Well, the $3 billion was net of a depreciation increase, and now that's going.

Yes, exactly. So we do better in that regard. Yes, I would say the way we kind of looked at it is we looked at the first half of '22 kind of how we were running both in terms of cost of sales and in terms of OpEx. And we said, okay, from that point, we want to see a $3 billion savings. So we actually did start to see -- make some progress on the savings even in the back half of '22. But the lion's share of it is really coming in '23.

Okay. And that's -- but that's $3 billion of net reduction in kind of if I take cost of sales plus R&D?

Yes. It's roughly about $1 billion in cost of sales and about $2 billion in OpEx is roughly the way I would look at it. On the cost of sales, I would say, it's a lot of kind of blocking and tackling, just tightening the screws on various areas of the spend. And the team there has looked at opportunities to kind of reduce -- at least on a temporary basis, reduce the labor count, which is, I think, driven some of the more meaningful parts of the savings. On the OpEx side, we did -- we looked at all of our overhead functions, and we set a target to reduce them by about 25% year over -- from the baseline of the first half of '22. And then we've gone really program through program and looked at areas where we felt like we -- the investment was not justified by what we thought the return looked like, and we've kind of jettisoned those. So in some cases, that's been shutting it down. In some cases, it's been selling it. I think as many people know, there's some news out last week around a more streamlined graphics road map. So that was part of it. We went through the graphics in excruciating detail and looked for what were the areas that made sense and what were the areas that didn't and kind of streamlined that savings. So -- and we're not done with it to get to the full $3 billion. We still have more work to do, but we're, I think, well on our way to getting that. And now my attention is really kind of shifting to, okay, got good line of sight on the $3 billion, but that's only a slight -- a down payment on the $8 billion to $10 billion. And in the $8 billion to $10 billion -- the rest of, call it, the $8 billion to $10 billion, so roughly the $5 billion to $7 billion, that is more biased towards cost of sales than OpEx. And that's important because, obviously, and I think important for investors because it improves the gross margins, but not as easy to come by. So this is going to have to be really where we break the company reporting into 2 pieces, one that really reports a P&L based on kind of a foundry model; and the rest that reports all the business units by product because we're going to really drive efficiency in this foundry that is both supporting our internal needs and supporting our external customers in foundry.

And just to make sure we understand the $8 billion to $10 billion exiting '25 is relative to what? Like, $8 billion to $10 billion less cost in the -- then same thing meaning...

Obviously, there's a variable cost component. So when revenue goes up, there's a variable cost that we're not going to obviously see that go away. But -- so it's maybe hard to compare it, but it's really looking at the kind of the nonvariable aspect of our cost of sales and driving efficiency out of that spend.

Because I feel like if I think -- we have a detailed cost of sales model, which is a lot of guesswork, but like there's labor, there's overhead, there's depreciation. All those things are kind of going up, all of those meaningful. Both head count and labor costs going up, overhead going up, raw materials costs going up. So that amount of cost reduction is a...

Yes. I think about it more like on a wafer cost basis. We're trying to drive a certain amount of wafer cost improvement.

Yes. Okay. Okay. Great. Shifting to the growth drivers, maybe just start with market share. I know you don't want to get into a forecast, but in your core businesses, PC and server, how do you guys feel set to defend your market share going forward?

Yes. I think on the client side, we're doing pretty well. Alder Lake was a good product. It's performing really well as is the follow-on product Raptor Lake. And we'll have Meteor Lake out this year on Intel 4. So I think we feel really good about our product position in the client space. And I think we picked up a little bit of share in '22 when you look at it on an end consumption basis. And my expectation is we probably will do at least as good as we're currently at, if not improve it a little bit in '23. So for all intents and purposes, I think we've righted the ship there on our product performance. On the data center side, obviously, it's been a lot more mixed. Sapphire Rapids is out and ramping, which was important. And I recognize that it was not at the time we originally expected it or the second time we revised it was at the time we originally expected it. But something that probably wasn't really felt by the external community was we actually, in our last reset of this thing, kind of reset it to the end of the year, and we pulled it in a quarter. And I thought that was actually somewhat of an important step in the right direction that suddenly we were beating a commit date that we set with external customers and our own internal expectations. And so Sapphire will ramp here. I think it's going to be one of our fastest ramping, if not our fastest ramping, product that we've ever had. And now we look towards the follow-on products. We'll have Emerald out, of course, but the real important products that I think make a meaningful difference is Sierra Forest and Granite Rapids both out in '24, and they make a meaningful difference in terms of our competitive position. So the way Pat characterized it is, he said, "Hey, we'll have a falloff in market share in the first quarter. We're expecting that to stabilize through the year." I would not call that victory. I would call that we're slowing down the share loss and taking advantage, I think, of some of our markets being very depressed and seeing a recovery through the year. And so a mix with -- mix will help us through the year. And the net result of that will probably keep us, in an aggregate sense, relatively stable. But really, it won't be till Sierra and Granite that I think we have real products that really kind of stabilize it. And for a couple of reasons, one, the performance of Granite relative to what customers need and what the competitors will have is going to be fantastic. But also, we have -- we will have our first E-core line, a really efficient core that answers the needs of customers, that really have kind of a power crunch issue that really need a more efficient core. And so the combination of those 2 things, I think, will be really important. And we're actually going to have a webinar, I'm going to give them now. Hopefully, you won't mind I'd give an advertising, but we're going to have a webinar on March 29 on...

Pleasure of advertising competitors' conferences.

No, this is for us. So it's not -- yes. So data -- on data center, Sandra and John will go through the entire road map and give you, I think, a lot of good breadcrumbs to understand what the position is and how we think we will perform over the course of the next couple of years.

And can you talk to just the reception of Sapphire Rapids? I mean it seems like right now, when everyone is in kind of a budget cut environment sitting on excess inventory, there's a lot of focus on the current generation, I guess. And cost increases at a platform level for both you and your competitor seem to be kind of slowing down some of those transitions. Is that fair, do you think?

Yes. But this -- I mean I think the reception has been quite good on Sapphire Rapids. I think in certain workloads, it actually performs extremely well. You look at AI, security, these are workloads where -- are smaller but growing. And this -- and Sapphire Rapids is particularly well suited for that. So I think we're pretty pleased with how things are going.

Great. So I'll ask one more data center question, then we can open it to the audience if there are questions. You mentioned AI, and I think you mentioned on the call generative AI and some of the enthusiasm around that. How do you see that affecting Intel? Is that primarily some of the Gaudi assets and some of the specialty AI stuff that you have? But what -- could there be impact on the core server business? Do you see opportunities that, as people sort of say, hey, we're going to reorient our budgets towards these new workloads, how does that affect Intel?

I think anything that is a catalyst for requiring more computing is good news for us, quite honestly. Obviously, we have the Gaudi asset, which obviously is going to be helpful in terms of our ability to address the needs of AI workloads. Sapphire Rapids will support and be, from a CPU perspective, I think an important part of the product offering and for AI. And then we also have our own graphics product road map. And to the extent that it requires that level of compute, the parallel compute, then we'll have offerings as it relates to that. So I think AI, of course, is super helpful to help drive the business, drive the growth rate and create a catalyst for it.

Yes, I feel like there's a real opportunity in the sense of 3 months ago, the conversation of inference cost was kind of a nerdy semiconductor dialogue, and now it's suddenly Google, Microsoft figuring about it. But there ought to be -- and particularly when cost is the focus, there ought to be a lot of opportunity, for sure.

I mean there are obviously things that only parallel computing can handle. There's -- but it's expensive. And they're going to look at the most cost-optimized solution that they can drive. And CPU -- offloading the CPU, I think, is absolutely something that architecturally, they're going to want to go after.

Great. So I do want to get into foundry a little bit if there's time, but can we first see if there's questions from the audience. There's one up in the front here. Somebody have a mic? Over there.

So I guess considering the large amount of capital that's required to do such a really fast transition to some of the nodes, what is the strategic rationale to continue to own Altera and the stake in Mobileye?

Well, let me -- without talking about businesses we haven't discussed whether they make sense to monetize, let me talk about Mobileye. So -- and I'm limited, obviously, to what I can say, but we do view Mobileye as being particularly well positioned in what we think is a pretty exciting market. And so we will take, of course, opportunities to monetize some of that. And you're right, that is helpful in terms of funding the needs of the core business. That said, there's -- we don't want to leave too much on the table either. And so we will make those decisions when we think it maximizes the value creation and then leverage it to fund what is the core strategy of the business. And I would say, as it relates to any noncore asset within the business, we're -- and I don't think Pat probably gets enough credit for this quite honestly. We're super pragmatic around this particular topic. And anything where we think there's a good opportunity to create value for shareholders in excess of what we can create fully under the umbrella of Intel, we will go out and do. And I'll leave it at that for now, but -- and update you as the dynamics change.

And I think all of these are good businesses on their own, including the potential for foundry. But is it distracting to own all of these businesses relative to the goal of let's just make the best CPU on planet Earth.

Yes. I mean I think focus is absolutely a good strategy to have. And so yes, of course, some of these things like -- and I think Mobileye is a perfect example. It's now kind of off running on its own. We obviously pay attention to how they're doing financially because we've got to roll it all up. But they're executing and now they can be less of a distraction for the senior leadership team of Intel. And there'll be -- like I said, likely there'll be other opportunities to get more focused. For sure, our key focus is get 5 nodes out in 4 years, get our product portfolio back to right [ decisions ] and leverage what we think is a real core competency of ours in terms of process technology and manufacturing at scale to deliver to the set of customers in ways that we have in the past. Those 3 things are the main things that we're trying to drive and everything else will kind of be flexible around how we approach that.

There's one more here.

In your mind, what are the key milestones that Intel needs to hit to prove that the strategy of jumping 5 nodes in 4 years is working?

Yes. Well, I mean, it's really getting products out on those process technologies. Intel 7 is out and running. So we're -- we've already checked the box on one of them. We're really close on Meteor Lake will be Intel 4, so we're super close on that one. All things point to that one being in really great shape. And then we're on to Intel 3, which is 2024. And we've -- and you'll -- without stealing too much of the thunder of John and Sandra, I think you'll hear some things about our level of progress on those products would suggest that we're well on our way. On ETA, I think probably the best evidence to investors that we're really on track is announcing an ETA customer on the foundry side, which our expectation and hope will be that, that happens in 2023. If that happens, they are putting us through the paces, I can tell you that right now. And so if we can execute on that, I think it's a good endorsement of the ETA process.

All right. On that note, we're out of time. Wrapping up there. Thank you.

All right. Thanks. Appreciate it.