Intel Corporation (INTC) Earnings Call Transcript & Summary

Greetings, and welcome to the first day of JPMorgan's 49th Annual Technology, Media and Communications Conference. My name is Harlan Sur, I'm the semiconductor and semiconductor capital equipment analyst here at JPMorgan. Very pleased to have Pat Gelsinger, Chief Executive Officer of Intel, here with us. Recall that Pat joined Intel in mid-February, hit the ground running, has outlined an ambitious strategy for the Intel team. Prior to joining Intel this year, Pat had a successful 30-year career at Intel from 1979 to 2009. He joined EMC as President and Chief Operating Officer of EMC's Information Infrastructure Products business. And then in 2012, became CEO of VMware. I've asked Pat to start us off with some opening comments, and then we'll kick off the Q&A. But before we start, let me just first read the 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. Pat, thanks so much for joining us today. And let me go ahead and turn it over to you.

Thanks, Harlan, and great to be at the JPMC conference, just a thrill to be able to talk to you all today. And just a few kickoff thoughts. If we think of this period that we're in, it's just a spectacular one for technology and for semiconductors. And I call it the digitization of everything, and that was on a strong path, and then COVID kicked it into high gear. I've also referred to it as the era of the superpowers: cloud, mobility, AI and edge, right? And they reinforce each other. And behind all of them is more semiconductors. In March, we rolled out our IDM 2.0 strategy. With that was a doubling down on manufacturing. There is some period, are you in, are you out? What are you going to do? We said we are in a big way for manufacturing. We've more than doubled our capacity over the last few years. We are on an aggressive path to build out that manufacturing network. We use it in our own, in our use of fabs from foundry partners to deliver this broad capability of our product line, but now we're also augmenting that with IFS, Intel Foundry Services. And as I say, we've opened the doors wide for the Intel factories, for the Intel packaging, assembly test and fabs. We've put an exclamation point on that announcement with 2 new fabs in Arizona at $10 billion each. We also announced the major new packaging capabilities with $3.5 billion investment in our New Mexico packaging facility. We've also seen -- and I've described it as the world needs a more balanced supply chain. And there's strong interest in U.S. and Europe, right? Obviously, some of this is accentuated by the semiconductor shortages, auto lines being stopped because of this. And we're very pleased with the strong bipartisan support for the CHIPS Act and seeing that built out. Also seeing strong interest in EU. We've also laid out a accelerated cadence of our innovation in the packaging, the process, the product, software. All of these were just saying we are on a path to have leadership across all of them. The well-publicized 7-nanometer assembly, that's back on track. We have embraced the EUV. And I've set out a yearly cadence of process improvements. Also, we've restored our tick-tock product and core cadence. We see this cadence of innovation will have us delivering leadership products in every category that we participate in. We do see that Q1 was approximately the bottom for the data center business, and we're now on a solid path of recovery. We're seeing enterprise and government return. We've seen the cloud returning to growth for us with Ice Lake and Sapphire Rapids. We have strong product improvements in Ice Lake with customers now and Sapphire Rapids ramping strong in the first part of next year. We'd also say the client market is on fire, right? PCs, 1 million new to the day, right, that kind of market is back, and we see that in a sustained way. This isn't a COVID blip -- it's driven the acceleration of it, and with Tiger Lake, Alder Lake, strong products. And the adjacent markets are becoming stronger for us, our IOTG, our PSG business. Mobileye is stunning in their progression. To the broad investors, you're going to hear more from us, right? Obviously, March was a big date. We're here today with you. We have our innovation conference in October. Financial Analyst Day is set for November 18. We're going to have a cadence of updates that you'll be hearing from us as we go throughout the year, so that you can rebuild your confidence, understanding transparency of the business. And I'd simply say, the best days for Intel are in front of us.

Perfect. No. I appreciate the insights. Maybe at a high level, you helped architect solutions at the silicon level for PC and server compute at Intel. You successfully integrated those silicon solutions at the system level at EMC and VMware, with complex software and applications across on-prem, hybrid, cloud hardware architectures. You returned to Intel with the customers' perspective, with the customers' view on their compute requirements, data center road maps in areas where they will be spending their compute dollars. So as you look at Intel's future road maps and portfolio of products, I mean where do you see the largest opportunities for the Intel team to profitably grow the business?

Yes. When we look at the big businesses for us, I mean, obviously, as we said, client business is sort of at the front of that list. It's big. The market is on fire. We won back share in Q1. It's a supply limited market. And hey, we think we're going to be better at managing the supply network than anyone else. Data center, Q1, as we said, was more challenged. But we believe cloud is now building up again. E&G is building up. And with Ice Lake, Sapphire Rapids and Granite Rapids, we see ourselves getting stronger in a critical growth business for us. So those are clearly the 2 big drivers over the next couple of years. Now next to that, we just got a lot of other good stuff in the pipeline, right? Our IOTG business, now that we're past COVID, we're seeing strong return of that business. Our PSG business, the FPGA business, hey, we've got the best product line. We think our competitor in that space is going to be preoccupied with acquisitions and mergers. Does -- as I said, the Mobileye business, we had a record Q1, over 40% growth. And we just announced the Toyota win, right? We now have 14 of 15 of the major auto manufacturers using Mobileye, right? So it's like, this is feeling like an Intel business all of a sudden. And then our Graphics business is getting very competitive. So we expect to be share gainers in that space. So the 2 cores, but we're now putting these other businesses next to it. And then finally, of course, we announced the Foundry Services business. That's $100 billion TAM by 2025. We see there's only 3 companies that can be leading-edge providers in that space, and we're getting into it in a pretty aggressive way.

So one of the biggest hurdles at Intel has not been with the technology. It's not been with the architecture, the road maps or strategy. I mean the team has always had a strong leadership position in client and data center, compute, architecture. Challenge has been simply executing to your road maps, right, whether it's manufacturing road maps, product road maps or even your prior foundry strategy, right? So what is it that, Pat, that you and the team are doing differently? Changing internal processes, more accountability, organizational structure? Help us understand that. And how is the team responding to all these changes?

Yes. And it's -- to some degree, Harlan, it has to be all of the above, right, on that. So we have made leadership changes like in the process TV space, very happy with Ann Kelleher. We've seen people like Sanjay Natarajan come back, a super next-generation leader for us there. And we're seeing the results. We got past some of the stumbles at 10 and now 7. And the daily updates that we're getting on wafers coming out of fab, the full embrace of EUV, we are very confident that we have that back on track. In fact, right now, we're taking out the compute tile, the Meteor Lake compute tile is finishing, taped in as we speak. So we feel very good that all of those things are behind us. We also have the rich pipeline. Even though we were stumbling in getting it into manufacturing, our components research pipeline, and now augmented with our IBM partnership, is rich. But as we go to this year with cadence, I got lots of things to pick from. And we put the capital in place and the wafers in place and the teams in place to hit this yearly cadence of innovation. So everybody is highly focused. And as I'd say, we're passionate and a little bit pissed off about what's happened to get it back. But we're doing that across the design teams as well, getting our tick-tock discipline and cadence back. And with Sunil coming back to the team, he was the original architect of that, hey, the machine is executing again. And for instance, our Ponte Vecchio, we have the first power on of the full blade of sapphires and Ponte Vecchios for the big supercomputers, but also as a key component of our GPU and graphics strategy. So a predictable cadence of innovation, right, in the process, building on that and the package, doing that at the product level. And then of course, we continue to leverage the leadership that we've always had at package and its software. And we're seeing that we're able to hold more share, execute more effectively with our customers. And I'll simply say, hey, the team's fired up. We're committed. And that momentum is building every day inside the company.

And even it sounds like near term, the team is executing quite well, right? I think you said in your opening commentary that you're on track to launching your second-gen 10-nanometer products for clients, second half of this year, with the Sapphire Rapids beginning of next year. And then on top of that, you just said that team is getting ready to tape out their first 7-nanometer compute tile.

Yes. Lots of good things are happening. And I do feel that -- hey, not everything -- years of bad decisions don't get fixed because you got some new exciting CEO coming in. It takes discipline, team rebuilding, culture rebuilding, et cetera. But every day is getting better every day. And I'd like to describe it, Harlan, it's sort of like it's been a little bit desert, right? We've been beat up a little bit. But as the rains come, flowers start popping up all over the place. And that excitement, that deep technical, and as I call it, Grovian culture is coming back.

So you -- in your commentary, and obviously, one of the first things that you did as CEO was you talked about the IBM 2.0 strategy, enabling the team to catch up to its manufacturing competitor in a couple of years. And in order to achieve this, it will require the team to maintain an aggressive cadence on technology, node migrations beyond 7, back in March, on the earnings call as well. And today, you talked about a yearly cadence. And there has been some confusion as what you meant by that, right? Because the team does do internode optimizations, right, 7, 7-plus, 7-plus-plus, on an annual cadence. But I assume you mean yearly cadence, you're talking more than just internodal migration. So is it a combination of internode optimizations, combined with partial node shrinks, maybe transistor enhancements like you did with SuperFin? Maybe you can just help us clarify what you mean by yearly cadence.

Yes. And there will be -- I'll say, we're laying out, Barry and -- a lot of work is going behind this to deliver yearly improvements in the process technology. And beginning not minor steps, but yearly significant steps. Now right, with that, you have to be very thoughtful about how you manage the risk associated with it, whether you're doing a major step on the front end of the process, the back end of the process, when and how you make the lithographic step for it. I'll just say at this point, with the full embrace of EUV, we see that we're able to do multiple generations major improvements with the EUV. You'll see major process improvements at the transistor level as well. So we see that there are new transistor architectures that we're able to execute upon and then other back-end process improvements as well. And the combination, as we pull all of that together, will be a yearly improvement in the process technology to restore parity and to leap ahead, once again, into sustained leadership. But as you said, that's going to take us a couple of years. And later this year, we're going to give a much more thorough update of exactly what that road map will look like. But we're now engaging customers through IFS, right, to start laying out that exact process as well. But as you well know, it's more than just the process technology. It's the packaging technology as well. And this is an area that we never lost leadership. So we're building on our packaging technologies with Foveros, EMIB, Co-EMIB, things that will lay out, again, later this year of how we're sustaining the package leadership. And then we got to line up the core, so we'll have both big and little cores. And they will have a yearly cadence as well, right? So we're going to be able -- so we see that we have unquestioned power leadership with our Atom Core. We have unquestioned performance and power performance leadership with our big cores. They will have a yearly cadence associated with it, right? And lining that up with the team so that we're delivering on products. And as you said, hey, products will set on 10, on 7, on 5, on 3, on 2, on 1, and going as sub 1 on a yearly beat rate. And then making sure we have all the software architecture and infrastructure so that the software interfaces are stable and the yearly updates of hardware allow it to be consumed and fully embraced in the industry. And this -- I'll say this disciplined top-down view of execution machine. We have the resource. We have the teams. You've got to organize them correctly. And I would feel very good that, hey, '22, '23 sort of the road map is what it is. We have to go execute against it. '24, '25, unquestioned leadership is on the horizon.

Maybe more kind of near term. Some of your competition reported strong data center numbers in Q1, while the Intel team was down slightly quarter-over-quarter, down year-over-year. What are you seeing out there in the market? And how are you thinking about data center competition going forward? And how might this impact -- this is a question that we often get from investors, but how might this impact the pricing strategy for Intel?

Yes. Yes. And first, we'd say, hey, we think Q1 was approximately the bottom for our data center business in '21. And with it, there were some cloud digestion we talked about. We do see that there's now increased momentum in the enterprise and government sector. Q1 was a more competitive environment. And even as we exceeded our plan a little bit in Q1, it was not acceptable. And going forward, we expect cloud improvements sequentially, we expect E&G to return. And obviously, with Ice Lake and Sapphire rapids on the horizon, we're just getting better products. I'd say data center is about better products. Now we do think that in this competitive environment, we're going to be more competitive. And with that, we have a lot of tools in our toolbox. Some of those are going to be system benefits. We're on the front end of system technologies like CXL, and those have big benefits. So we're going to be on the front end of memory hierarchy improvements and using the Optane technology. We're in the front end of that. We have strong software assets that allow us to optimize. And hey, if I can get 20% or 30% out of the software stack when it's running on an Intel processor, that's a big deal, right, for competitive advantage, tools our competitors don't have. And of course, sometimes, that will mean we're going to have to be more price aggressive. And we're okay with that. But it's only one of the tools in our toolbox. I'd also emphasize that it's a strong market. So in a strong market, a strong road map, good execution, limited supply capacity, we expect that there could be some pricing pressure, but not that significant relative to these other tools that we have in our toolbox to go fight, to hold and to win back share.

On the subject of -- there's a lot -- obviously, there's -- with the chip shortages, a lot of focus on diversification of manufacturing, obviously, a lot of government initiatives to try to bring back manufacturing capability and technology back to the U.S. As a part of President Biden's proposed infrastructure spending bill, the amount of the support for domestic semiconductor manufacturers and R&D spending was upped from $35 billion to $37 billion under the CHIPS Act proposal to $50 billion, right? Of which, I think, we think about half or more, I think, could be used to subsidize U.S. domestic manufacturing program. So call it, $25 billion to $30 billion for manufacturing subsidies. How important are these subsidy dollars as it relates to Intel's ability to execute on its IDM 2.0 strategy, especially Intel Foundry Services?

Yes. It's a great question, Harlan, and one I spent a fair amount of time on since coming back to the company. And first, let's put a little bit of context on this. In 1990, Europe was 44% and the U.S. was 37% of worldwide semiconductor manufacturing. Today, Europe is 9% and the U.S. is 12%, right? A dramatic decline. And I -- we still are 75% of semiconductors are consumed in U.S. and Europe. Great. And maybe just say, COVID, the auto shortages, other industry shortages are acute reminders that we need a more resilient and a more globally balanced supply chain at that level. That is the picture. And we need more of that on U.S. soil, more of that with U.S. IP. And with that, we just said, we've been very active in saying that U.S. needs more semiconductors, right? To go back to my opening comments, the world is becoming more digital, all of that runs on semiconductors. So the world needs more semiconductors. And the U.S., right, after essentially 30 years of decline, needs to reverse that decline. This is more important than oil. This is more important than a lot of other things as the world becomes more digital. So with that, we said, hey, the U.S. should step up, and we've seen strong bipartisan support for that. We're excited about Biden's Jobs Plan. We're excited about the Enduring Frontiers Act, the CHIP Act, the respective programs around manufacturing build-out, R&D, packaging, jobs creation. All of these are good. This is a very good first step in the U.S. As I said, when we made our announcement around the Arizona fabs, we are putting our chips on the table. I put my $20 billion on the table for Arizona, the $3.5 billion for New Mexico, the investments for Israel and Ireland. Those are our investments. And I want to go bigger and faster with additional help from the governments, right? So those are -- our business support some, we're making those investments, we need to make them. But we believe they should be bigger, faster, broader. And that's necessary to fully support the foundry capabilities, also securing the U.S. supply chain, also for U.S. security and defense interest as well. Similarly, we would also say the same thing if we were sitting at Brussels right now. And in fact, when I was there a few weeks ago, very strong resonance to the same issue. And their decline was even more significant than the U.S. industry decline at that point. And the moonshot picture that I've laid out would be that the U.S. becomes 30%, Europe becomes 20% over the next decade or so. And I think if we were sitting here and we said, hey, 50% is western, 30% and 20% and 50% was in Asia, that would be an extraordinary outcome, right, for us to reverse this decline and secure our long-term supply chain. So overall, we're very happy to see the U.S. go -- move forward. We're doing our part, but we can't do it alone. I'm pretty excited to see the government stepping in as we see it both in U.S. and in Europe.

Certainly, as we talk to a lot of the companies that we cover, we cover over 20 semiconductor companies, they definitely are thinking about diversification of their foundry partner base as well. When we speak with investors though, the feedback that we get is that Intel did enter the foundry business somewhat selectively back in 2010-2011 time frame. And outside of a handful of customers, the team never really drove much scale in the business. So fast forward to today, what makes the team confident that they can be successful, especially when you are 1 or 2 technology nodes behind TSMC and when many of your large potential foundry customers probably consider you somewhat as a competitor?

Yes. Yes. Great, great question. And I'd say, first, if you look back on that, I think Intel was sort of tiptoeing in the business and wasn't demonstrating intentionality. They're like, hey, you want to send me a few wafers? Right? A lot has changed since then, right? And I'll just say, one is, hey, there's only 3 leading-edge companies in the world: Samsung, TSMC and Intel. Right? As we've already talked about, Harlan, we are on a path to parity and leadership over the next few years. And by the way, the design starts are 2 to 3 years in advance. So essentially, I am designing today with -- on leading edge that's as good as TSMC or Samsung and offering that to customers today. That's as good as anybody will have because the 24 designs are starting now, right, as they start engaging with their customers as well. So we're presenting them that PDKs now for a leading-edge process technology that we have under development today. That was a -- we're taking that off the table. We also said, hey, you only have a few choices and you want supply chain flexibility and resilience as well, right? You don't want to hold your egg in the basket off a timeline fab, right, because of the geopolitical situations. Just supply chain resilience management post COVID, right? You want to look at your second supplier alternatives as well. Third, we're setting it up as a separate business, reporting directly to me to be able to manage the concerns of, okay, will I really get supported? Right? Are you really going to give me wafer capacity vis-a-vis your own products? And it's like, yes, right? I just announced 2 fabs in Arizona. We're building capacity, putting that into our long-range plans. We're going to put the terms and conditions next to that to meet their expectations as well. It's a competitive business, so we're going to have to be competitive. We're also building out the full IP libraries. 10 years ago, the design on Intel was a herculean task because it was highly proprietary to Intel. We've now opened that up with the industry IP blocks, industry design tools, et cetera. So we're lowering the hurdle for people to look at us vis-a-vis their other foundry partners. And finally, we're making all of our technologies available, the packaging technology. I have customers today that are saying, can I take my TSMC wafer and run it through your world-leading package assembly and test? And the answer is, yes. We're also making all of our x86 and other core IP available: Intel Graphics, Intel Networking, Intel Big Core, Intel Little Core, things that were never made available before to foundry customers. Those are being made available for our foundries. And we're seeing particular interest for what I'll call these hybrid designs from the large-scale cloud guys. They say, boy, I love your standard Xeon, but I don't use these transistors because at my cloud scale, I do things differently. But I have some of my transistors that I'd really like. And now we're saying, yes, let's co-mingle your designs with our designs. And we also -- when we rolled this out, we had an extraordinary support from leading companies: Amazon, Microsoft, Cisco, IBM, Qualcomm and others, all were, yes, this is good, we're excited about you stepping into this business. And obviously, we have to go build out the right talent, the right services mindset. Building a world-class business is more than just process technology and IP. But we're getting great response from the industry. And as I said on our last earnings call, almost 100 customers are now part of that pipeline.

Perfect. Perfect. And then maybe just kind of back to the 2021 outlook here. Good news is, as you mentioned several times during our conversation today, core data center business, cloud spending is picking back up after a 3-quarter digestion phase, enterprise spending is coming back. Telco service provider has continued to be just a great business for the team, all good for your data center business. And it sounds like business troughed in Q1 and maybe remains good throughout the remainder of this year. PC demand, as you mentioned, also continues to remain relatively healthy as well. But your outlook for 2021 assumes a relatively flattish second half over first half. So help us unpack the muted second half outlook by the team.

To some degree, and we were explicit about this on the Q1 earnings call, was it's a supply-constrained outlook. We are supply constrained. We have substrate constraints. Also, our customers are supply constrained. We're now wrestling through the issues that they say, boy, hey, I don't have enough power controllers, right, to have a mix -- a matched set. I don't know enough easement devices to my cloud to be built out. So everybody, both, -- we are working our supply chain issues and our customers are working their supply chain for complete matched sets to ship our products. So -- and my supply chain head who's been in the business forever, he said in his 25 years, he's never seen the supply chains more disruptive than they are today. So I'll say, at the highest level, we are wrestling through that, and that has muted our outlook for the second half somewhat. If I could build more, my revenue outlook could be higher. But I'll tell you, we are wrestling through this every single day, keep pushing those numbers up from what we committed to on the earnings call. As you noted, PC demand is, I'll say, just extraordinary, right? We're continuing to see, right, education being built out. We're -- we still have a major refresh cycle in front of us. Everybody's PCs in the business out of date, more demand in this environment. So we are considerably more optimistic for sustained growth here than we're seeing from some of the analyst forecast. And just seeing no backlog in the supply chain whatsoever, right, as customers have minimum turn cycles in their history. As we said, on the data center business, we do expect to see second half returning to year-on-year growth. And there are some unique issues with entity lists in China that we're fighting through. But overall, we expect that both enterprise and government as well as cloud will enable us to be resuming to growth in that business. And then a lot of the other areas like we touched on, IOTG, PSG, Mobileye, these will be long-term growth drivers for the company, as we're essentially small, a big company who has a small business entering good growth potential in those areas like 5G cycle. Hey, I think we've got a good 5, 8 years in front of us in that particular area, not just for 5G, but also private 5G as this starts entering the space. So overall, this is like the best TAM environment. And right now, it really is supply chain limited overall, and boy, semiconductors are hot. They're going to be hot for the next decade.

Absolutely. Going back to the core technology. As you mentioned, the team just announced a $3.5 billion investment to equip your New Mexico operations with the capability to support your heterogeneous advanced packaging technology like Foveros, your EMIB processes. And I know the team is moving your core client products in the second half. And this year, new architecture, data center products will follow. We're actually hearing that your potential foundry partner is actually very interested in these 2 technologies as well. So what kind of foundry customer interest have you seen? Will the packaging be a part of the foundry strategy as well going forward?

Yes. A lot of the fab designs, they take 2-ish years for people to move a design onto a new. Even if it's a fairly straightforward design, it's a minimum of 18 months. So major new designs are 2-plus years. And as we said, hey, we have a 22-nanometer offering that we're presenting to customers today, 7, 5, 3, 2, 1, right, for advanced, but those take longer. A lot of interest near term in the packaging because those things, we can start to literally start to deliver those to customers into the marketplace in 6 months. And we've seen considerable interest. Foveros, in particular, our 2 and 3D packaging technologies, are getting a lot of immediate intrigue by customers. So I'll say those are some of the early on ramps for people to jump on to Intel Foundry Services. And I love this because it starts to get some of our foundry services muscles in place, where I'm able to get that done much sooner than I might for a new design start. And we've seen that from some of the automotive customers. We see that in particular from some of the cloud customers, have been super motivated to jump on this, the different segments with a positive. Obviously, part of what motivated the acceleration of the New Mexico build-out was exactly this, right? Boy, we have -- everybody thinks fabs at Arizona are sexy, and they are. But packaging is going to be pretty cool. Particularly, and you mentioned this in your -- in the question, as we go to the disaggregated designs, as we call them, where it does become multiple tiles. It's called heterogeneous, right? We'll be launching at [Alder Lake ], but big and little core, as well multi tile, right, being brought together in an integrated packaging technology. This is pretty differentiated, and we're well ahead of competitors in this space. It will become a key part of our competitiveness even beginning late this year.

Perfect. We've got a couple of minutes left. So Pat, why don't I turn it back to you if you have some closing remarks.

Thank you so much. And to everybody that's been able to join us today, I'll just say thank you for your interest in the company. It's been, I'll say, an exhilarating. I think today, in fact, might be my exact 100-day back in the job. So the 100-day journey is well underway. We've gotten a lot done over the first 100 days with our strategic clarity on our manufacturing, how we're going to make that available and engage with the industry in new ways with IBM 2.0, strong geopolitical support from U.S. and Europe to accelerate the strategy. And we're building. And my internal phrase has been execution, execution, execution. right? We need to be in that position that our customers can count on us to deliver their technologies that they can build their businesses around. And they want us to win. That's been hugely encouraging, Harlan, that we've seen some strong support from our customers for us to come back. In a supply-constrained environment, they need us more than ever. We're stepping up into that. We're rebuilding that cadence of execution that they can count upon. And as the world becomes more digital, I believe we have 10 good years of semiconductors in front of us. And Intel plans on stepping into that demand, that need, that opportunity in a huge and effective way. And look forward to future updates later this year to really help build out the rest of that narrative for you and all of our investors.

Absolutely. I mean, we look forward to following the progress and execution of the team this year. And again, thanks, Pat, for joining us today. Really appreciate it.

Very good. Thank you all.

Thank you.