Intel Corporation (INTC) Earnings Call Transcript & Summary

All right. Thank you. Thank you, and welcome back, everybody. It's a pleasure to have Michelle Johnston, who heads up all of sales and marketing at Intel; as well as Tony Balow from the IR side. So Michelle, thank you.

My pleasure.

I thought before we get started, I have not met you before, and it will be great for the audience to get a sense for your background, what have you done at Intel and what do you do in the current role.

Okay. Great. First off, thanks for having me. So I've been at Intel a long time, but I've been at Intel for just over 24 years. And I grew up on the product side, really managing all of our distribution products. So when you think about box motherboards, box CPUs, our net products, that's where I started, and that's where I grew up. And about 7 years ago, moved into sales and have remained in sales. And now I manage our global sales team, our worldwide marketing team and our communications.

Well so it sounds like you've had a pretty international role as well, right, if you've managed motherboard business?

Yes. Absolutely.

Yes. Great.

Yes. Hi to my friends in Taiwan.

Awesome. So -- and I was thinking before we really get into the Q&A, if you could please remind us what you said about the current quarter? And how have things progressed?

Yes. So obviously, I can't say a lot about the current quarter, but maybe I'll just start with where we were at the beginning of the year. So we came into 2020 already thinking that we kind of had an atypical year, with a stronger first half than second half, and then COVID hit. And first half turned out to be much, much stronger than anyone would have expected. Obviously, the large trends on our kids are all learning from home, we're all working from home. But we've remained cautious in the second half based on the forecasted GDP compression. In the cloud business, specifically, we have seen approximately 5 quarters of strong growth. And we believe in the second half, at some point, we're going to see the digestion cycle within our cloud service providers really start to kick in. We also had an amazing Q2 in networking, and we think that, that normalizes in the second half. For enterprise and government, we think that those stay weaker based on the GDP compression and the continued work-from-home and learn-from-home trends. And then over the last 18 months, we've seen a pretty big explosion in growth around PC, really based on the Windows 10 refresh. And we think that, that starts to mature as we see a large shift from desktop to notebook with the learn-from-home and work-from-home trends. And we do expect to gain share in that category in the second half.

Right. Actually, if we just stay with the share side on the client side, and this was something that was brought up in the earnings call as well. Can you give us a few more bounds around that? Is it mostly on the low end because you have more supply coming online? Or is it across the platform?

So maybe I'll just start within data-centric. We've had a very strong position in that category. And as we look at how we've done through the first half of the year, we've actually done better than we were projecting and believe that, that will continue into the second half. When we think about the client market specifically, and we think about some of the supply shortages we had, we really prioritized our builds to what our customers needed, which was more of our core products. And so as we made that shift, we really left the low end of the market underserved by Intel. And that's absolutely where we have seen our market segment share losses. And we look into the second half, we're going to build a much richer mix of small core product for our customers. And we believe that, that is a more transactional segment, Ambrish. And therefore, we have an opportunity to be able to go and win a lot of those tenders and deals and have seen so. And so as that supply comes to fruition, we believe that is a great opportunity for us to win market segment share.

So if I just think about the PC market for the second half, it sounds like, at least based on the data that we've been getting out of Taiwan on the ODM side, is that the low end, especially Chromebooks, continues to be pretty strong, also supporting learn from home. So that seems to be -- is that consistent with how you're thinking about the PC market?

Yes. That's very consistent. We see a very strong consumer second half, and we see kind of an insatiable appetite for people to get those Chromebooks or notebooks that they can put into the hands of their children as they learn from home. We are seeing, I think, a shift when we think about just the PC market as a whole. We used to think about maybe having 1 or 2 PCs at home was enough. And now that we're all at home, we're learning from home and working from home, everybody wants their own. So we are seeing a density within each household as -- for people to get through the day, each student needs their own device to be able to interact with their classes, and we continue to see that getting stronger into the second half.

Okay. I'm going to now bring in the big elephant into the room. And it would be unfair of me to ask you about the process technology side and on the manufacturing side. But given your background, I think it will be helpful for us to understand -- because you're dealing with customers day in and day out with your key accounts. So with the slip that Intel has had on the 7-nanometer or the announced slip on the 7 side, following up on the 10-nanometer side, what has been the customer reaction to the news in terms of the conversations you're having with your big accounts? And how worried are your major customers that this puts Intel at a structural disadvantage against AMD?

Yes, I'm glad you asked this because it is, you could say, the elephant in the room. And maybe I'll just start with, no one's happy when schedules move. And we've been having a lot of very open and transparent conversations with our customers because our relationships are fundamentally built on trust. And that means going back and forth, understanding what their concerns are as well as how that maps to our road map so that they can solve their customer problems. I will say the feedback that we have gotten is they very much appreciate that we kind of put them front and center in how we kind of build our strategy moving forward. Really talking about, hey, our focus from here forward is the cadence of leadership products that we're going to deliver every year, kind of a model that they can rely on. And I do want to remind you that we have an incredible lineup of products in '20, '21 and '22. And those are really going to be our volume runners. And those aren't impacted by the announcements that we made at earnings. So we've got Tiger Lake coming out, launches next week. I think it's going to be an amazing product that is, in my opinion, perfectly defined and built for the environment that we're in today. In client, we also have Alder Lake coming out in the second half of next year, which we haven't talked much about, but it's also an incredible product. And on the server side, we've got Ice Lake in the back half of this year and Sapphire Rapids in 2021. And as I had conversations with customers, both those that integrate and those that consume down channel, their very first question was, are all of those products on track? Am I still going to be able to get those products? And they are. And every one of our customers has active designs on all those products now. So that's kind of at the forefront of their thinking. And they really wanted to understand if those -- if there were any lineup changes to those products because they haven't started designing for their 7-nanometer products now. Obviously, culture at Intel runs really, really the deep and Andy Grove -- and it's painted all over our walls, but only the paranoid survive. So I am constantly worried about competition and having conversations with our customers about what they're seeing and what they need from us. But also what I have heard from our customers is we bring a lot to the table beyond product. We have great support teams, great customer engineering teams that really help them build their platforms, optimize those platforms, ensure they get to market and that they ramp. And so when they think about Intel, they think about all of those people coming together to help them be successful. And so I'll be the first one to admit that no one's happy with the timing. But we're absolutely laser-focused on executing the products that we have in the short term. I think we've given our customers a reason to believe in us going forward and in our opportunity. We've got a great lineup of products over the next couple of years. And we've also told them that we're going to be very flexible in the way that we build, so we can deliver the right product for them at the right time.

And Michelle, if I could just stay with the topic. You are part of the senior management team and the announcement came when nobody was expecting it. So can I just -- help us understand the thinking behind the timing of coming out with it at the time that you did.

Well we have additional responsibility to come out with it when we know it. And when -- we found a problem on our Ponte Vecchio that we talked about in our earnings announcement. And once we figured that we had that problem, we figured out the fix to it, and we saw that it would move products out. We felt like not only our investors and our partners, but our customers needed to know immediately that we had that, that we had a fix for the problem that we had found, that we had fixed it and that we were ready to move forward, but that it was going to adjust schedules. So truth and transparency is one of our core values. And we have learned from the past that when we learn these things, we need to go out and we need to tell people so customers can plan on their products accordingly.

Yes. Yes, that make sense really. So it sounds like you had been doing a lot of background work and then you reached a point where you felt that, a, you had -- you knew it was going to be pushed out, but then you also had solutions to address that, and that's the timing that you came out and you announced it to us?.

Yes.

Okay. Okay. Let me switch to an area where obviously you're -- coming back to the customer engagement side, where your day-to-day dealings are with them. You talked about a few products, Tiger Lake, Sapphire Rapids, Ice Lake. Could you tie it together with -- from your vantage point and what your customers heard? You had a pretty detailed Intel Architecture Day. And there were quite a few announcements there. And I thought you rather walked us through many, many areas. So what did you come away with and said, all right, customers are really looking forward to this versus this, and boy, I'm really excited and I can go to customers? And also, if you could throw in -- this is -- Intel has been talking about this that we tend to get very focused on the node. And one of the things you guys came out and said, well, look, within intra-node, we have a 20% improvement. So if you could please just kind of walk us through what your takeaway, your customers' takeaway was from the Intel Architecture Day. And then also address the intra-node performance improvements that you're...

Yes. So obviously, I can't get into too much detail on the Tiger Lake specifications. That product launch is right around the corner. And I think as you see that product launch, you see the benchmarks, you're all going to be very excited about what it offers. But one of the things that customers were excited about and I think is so important in this process and node technology conversation is SuperFin, which is the biggest internode performance increase in the history of Intel, and you talked about the 20%. Those types of technological breakthroughs that we have internode make customers really excited and they make them excited because not only that can we offer them better products but it does show that we still have technology and hold leadership and that we can continue to evolve within a node. You look at what we did on 14-nanometer, and we're on track to do the same with 10-nanometer. We also had an announcement around Willow Cove, which gives better frequency performance over Sunny Cove over a much wider range of voltages, which gives customers more scale and more opportunity. Raja talked about our new graphics products, which we've started to sample our DG1 product for clients. And that's more relevant than ever before if you look at both the workloads that we see in our data-centric businesses as well as our client businesses. And so he really talked about all of those. And then he kind of wrapped it in as we think about engineering and we think about delivering world-class products, we talked about the 6 pillars of engineering and how important it is to be able to deliver more than just process and packaging. So whether that's software, security, interconnects, memory, XPU, it's really about the portfolio of products that Intel can bring to market that gets our customers excited. Customers like to work with Intel. I hear this time and time again that we're good at supporting customers, that our technical resources are top-notch. And so when we can offer them a wider range of products and opportunities to partner with us, it makes it easier for them because they know what to expect, they know the support they're going to get. And we know that we're going to -- they know that Intel is going to help them get across the finish line and get their products to market. And not just get that product to market, but we're going to help downstream ensure that there's product pull-through so that they can sell those products.

Yes. And I'm sorry, I went straight into the Architecture Day...

Not a problem.

Because I thought that was pretty interesting. But I wanted to zoom back up onto competition. Intel has always correctly said that when asked that, hey, we don't take competition lightly. We always -- it's not like there's -- AMD has been around for a long time. And Intel always said that. But we have not seen this in over a decade, a resurgent product portfolio from AMD that has taken share on the client side and also on the server side, they have taken share. So from your perspective, what is the competitive response? And what do you hear from your customers? The baggage from the past, where they had their own manufacturing and the ability to supply, some of that doesn't exist anymore. And now there's TSMC, which is a very credible supplier worldwide for years. So that -- and I'm not saying that you go competitively and say, well, look, they can't supply. Obviously, I'm not implying that. But just talking from my perspective, when we think about all the baggage, I've been following the group for a long time. So what is it that you go to customers with on a -- is it -- it can't just be pricing, it's performance pricing. So how has the competitive dynamics changed? How has that change evolve how you approach the market?

Well I think we've always been very focused on competition. And when I took this job 3 years ago, customer obsession was one of kind of the key values that I wanted to bring to our sales, marketing and communication organization. I didn't feel like we've spent enough time listening to the problems that customers were trying to solve, understanding those problems and making sure that we were coming up with solutions for those. So that is forefront at how we approach every customer conversation. And obviously, we have learning and improvement along the way, but that's where we want to start. And so price is always one lever that we can use. But for us, it's never the primary one. It has been a more competitive environment, even over the last 6 to 9 months. And if you look at our ASP mix, you look at our core mix, we've actually been selling up in the stack, both in data center and client, to all of our OEMs and CSPs. So we do see a strong desire for our performance products. In earnings, you probably heard George talk a little bit about the fact that we always have provisions for competition. We always have ways that we can respond, and that hasn't changed since the beginning of the year. But if anything, we feel a little better about our position than where we expected to be even 180 days ago as we've responded to the market and had conversations with customers. And one of the things I hear a lot from customers is there are so many things that we do for them beyond price. And I've mentioned a few of them, but maybe some examples are we have over 15,000 engineers inside of Intel that just help optimize software stacks for our customers. So think about that, whether that's your driver stack, whether that's optimizing a workload, whether that's doing BIOS help, we have all of those positioned around the globe to handle our global customer base. We also have a very special set of dedicated engineers for each customer that helps them get through the transitions, helps them debug problems real-time and gives them better opportunity to be time to market because of those resources. And then we have a pretty amazing and broad sales team that not only is focused on ensuring that we win the socket design, maybe at an OEM or CSP, but we have sellers that go all the way through all the channels, whether it be in distribution, whether it be in retail, whether it be in finance, that help pull the products through, that work with end users who are looking to deploy technology and then we'll partner with our upstream customers that are building that to ensure flow-through in sales. And that's really important and something I hear time and time again that customers appreciate. And in cases where we really want to win back a socket, price may be part of that equation. And we have an amazing portfolio of assets that we can lean on. So I never think it really comes down to one benchmark. It's really about delivering a complete set of performance experiences for our customers, which we absolutely know how to deliver at the right price. And then I would be remiss if I didn't say on top of all of that, we have an amazing top-15 brand that really does serve as an accelerant. It's something that customers continually come in and ask for. We also have leading partner programs, both with our channel and OEMs, that they all continue to say as a differentiator versus other competitors. And we continue to lean into those, evolve and fund them based on customer feedback. And so it's kind of this cycle, Ambrish, where we want to make sure we're absolutely listening. We're going to have to pivot and respond when we need to, but it's a series of things that Intel can respond with that differentiates us from other people in the marketplace.

Okay. Fair enough. Another topic that you probably have a lot more insights than at least me is heterogeneous compute and AI and Intel's approach. So how are you seeing this deployment rollout from -- by your customers? And of course, Altera was one piece of the puzzle that you acquired a few years ago. So can you help us understand Intel's approach: a, the end market, where do you see it being deployed; b, what is Intel's approach? And I know you acquired Nervana and Habana Labs was acquired also. You addressed that. But maybe focus in on what you're seeing on the deployment side and how you are bringing Intel solution to address that.

Okay. So first off, I would say AI is really in the [ incumbency ] phase, and I see it being deployed in so many different ways. Some customers are running workloads that have -- need very light inferences, while we see others that need high performance, low latency and they kind of have this insatiable appetite and are challenging us to figure out how we're going to solve some of the problems that they see AI fixing and helping with. And our approach has been to be able to offer our customers a variety of choice based on the workload and how they want to specifically deploy. So when we think about Xeon as our family of products there, we're the only processor out there that has an embedded AI engine with DL Boost. So that gives an opportunity to deploy accelerators for customers who don't want to buy something like an FPGA or do something more specific like a Habana. And then on the other spectrum, we have customers where they want low-latency, high-performance Habana solutions for inference and training, and they want it to be able to run their specific workload and to have it be optimized. And then in between, obviously with our Altera acquisition, you have our FPGAs and then in the future, our GPU products. And we're really starting to understand that all AI workloads, like training, employ GPUs, which is one of the reasons that we are moving into that because we want to be able to offer our customers a full portfolio of products. And like I said, I really do believe we're just at the beginning of seeing where this is going to be. And all around the world, we're seeing customers use AI in things like health care, universities conducting research, life sciences. And maybe just one example, I think this is a feel-good example. As we think about COVID-19 and we think about all the learn-from-home situations, obviously when you think about deploying that across the world, there are some countries that are better set up from a learn-from-home scenario than others. So we worked with a company called Career Launcher in India, along with AWS, for something called Project Aspiration. And Career Launcher was an online portal to teach adults new skills and trades. And we wanted to take that portal, and we wanted to double its bandwidth along with the cloud, and we wanted to be able, in Delhi, to be able to take this platform and deliver it for K-12 users so that 150,000 students could start to learn online, be it from a PC or from their parent's phone, any time, 24 hours of a day. So we wanted it to be live or consumable based on when students could get access to the platform. And that required partnership between Intel, AWS and Career Launcher to build a scalable platform based on Intel Xeon, based on AWS capabilities and Intel Launcher's (sic) [ Career Launcher's ] platform. And that not only required all the AI to make that work, to partner with the government, but also go out and train everybody to be able to facilitate an implementation like that almost overnight. And that's just one example. And we see so many in health care, as you see AI look at the strains of the viruses and how you can learn from it. We've seen customers in the U.K. where you -- I believe it's the U.K., where you can cough into your phone and you can actually take an audio recording of your cough -- excuse me, this is in Brazil, an audio recording of your cough, and you can send that in. And based on AI, they can do an evaluation of your cough, and they have a high rate of being able to tell whether or not you're infected with COVID or not. And in an environment where testing may not be available, it's just one of the amazing ways that AI is being deployed around the globe and we're part of. And Ambrish, this is just the beginning. There's going to be so many opportunities.

Yes, sounds pretty fascinating. You mentioned graphics, and I wanted to bring you back to that. Can you tell us a little bit more about Xe graphics? And then folks like me, who have been following Intel for a very long time, always think about Larrabee. So the questions are why is Xe any different than Larrabee. And then more importantly, this is a very well-served market by NVIDIA and AMD, not on the data-center side, but at least on the client side, and CUDA, which has been around, I forget the generation NVIDIA's on. So where and how does Intel see -- first, where do you see the wedge point that you can make inroads in? I'm assuming data center has to be a big one. And then you're probably the best one to answer this because you also talk to the customers. How do you -- a, how do you see where is the unmet need? And then how do you address that? So sorry, starting out with why will Xe be different than Larrabee.

Okay. That's an excellent question. I absolutely probably remember Larrabee much like you. Maybe I think the biggest difference is when we just think about the installed base of integrated graphics that Intel has today and how big that is, I believe our customers already see us as a graphics provider. So I think it's very different than we were in Larrabee, where we have amazing market segment share. We don't share specifics on integrated graphics. And our customers are very used to working with us. And we've continued to making an investment and progression in those integrated products. You saw it in Ice Lake and then next week when you see Tiger Lake, you can tell me if you think that investment has continued. We've hired experts like Raja to come in and help us really understand the importance of that unified developer framework that you asked about and how you can bring those together. And so we've launched oneAPI, which is our approach to a developer framework, where you can write code and have it optimized across the breadth of our portfolio. So imagine writing code and it working on our integrated graphics solution, and that same code being codable to our graphics cards. It really -- for our customers, if you think about managing their portfolio, it means they don't have to manage and have the same resources to manage all these multiple software stacks. And we always hear from our customers that the more we can do to simplify, simplify, simplify, the better. And so oneAPI is launched in beta, and you're going to see more on that later this year. Our products really span a lot of products or a lot of segments from LP all the way to HPC. I see data center as an amazing opportunity. If you just look at the growth in data center over the next 5 years, $200 billion TAM, 6% to 8% growth, it's an amazing opportunity. But I think one of the things that you may or may not have heard before is, Ambrish, is our customers are really excited about us getting into this market. Customers love choice. They love optionality. And what we're hearing from them is the more choice they have, the better they can solve their customer problems. So as I engage with them, most of the time it's how can I get it sooner and faster. So I do think we have the brand and the reputation to be taken serious in this category, especially based on our integrated graphics performance to date. And we have customers that are pulling for us to absolutely get into this marketplace.

That's an interesting point. I didn't even think about it that the integrated graphics is a very large installed base versus 10 years ago. And I think it was 10 years ago when Larrabee was supposed to be launched.

Yes. 2012, yes.

Right. So it's fair to assume that -- and I don't remember off the top of my head, if you could just please remind us what is the cadence of Xe as it comes out. The client comes out first, right? It's been sampled and all that?

Yes. The client comes out in -- Tony, help me with the dates -- late 2021, early 2022, is what I remember.

Yes, right. Right. Okay. All right. I'm going to pause here and take questions from the line. So folks, again, same method. I think everybody is familiar by now. I'm just going to refresh my screen here. Okay. So here's a question from the line, Michelle, this 7-nanometer push. How has this -- I think you addressed some of it earlier. How has this influenced your discussions with the customers? And do they want you -- and this is a new one, do they want you to outsource 7-nanometer or keep it in-house? Is there a customer input into the make-or-buy decision, I guess, is the question.

Yes. It's an interesting question. So obviously, we've had lots of conversations recently with our customers. And frankly, I don't believe customers really say, "Oh, I want you to outsource that product, or I want Intel to continue to build it." What customers tell us is, "I want you to use the process that is going to allow you to be time to market and deliver the best product performance at the right price point for me to win in the marketplace." I do think that customers are happy that we are open to external manufacturing in the sense it gives us optionality and it gives us more opportunities to be able to deliver them a wider range of products. But I have yet to have a customer say, I want my product built at A, B, C.

Yes. Yes. Makes sense. I mean I think for them, what's important is the timing and also the performance that you're promising them. Maybe we could just stay with that a little bit longer. And just refresh, for everybody's benefit, what have you said in terms of -- publicly, and maybe, Tony, you can help us here, what has Intel said publicly about what are the points that are being -- within the solution, what is being outsourced versus all that is going to be internal?

Tony, I'm going to let you take that one.

Yes. No, absolutely. So...

It's definitely outside of my wheelhouse.

I'm sorry. I put you on the spot there but just wanted to make sure that we're all on the same page.

That's all right. You didn't put me on the spot. If our investors want to know, we want to answer.

So yes, no problem. Let me try to tune that up a bit for you. So basically, if you look at what we talked about back in earnings, we talked about, first of all, Ponte Vecchio, which was our data-center-focused GPGPU. What we basically said there is that, that product, and there was some further disclosures in Architecture Day as well, is obviously built on a disaggregated format. And coming into earnings, the plan had always been to take some of those parts external and some internal. It had previously, the compute tile on Ponte Vecchio, been slated to be our first 7-nanometer product. And what we said at earnings is that we would be moving that tile externally to a foundry. And then therefore, our first 7-nanometer would now be the client CPU in late '22 or early '23. But what you're seeing is, as we move to a disaggregated format, we haven't specifically disclosed which tiles on which products are internal. Leading edge, internal; trailing edge are external going forward. But that is part of the plan. Obviously, as you move to disaggregated format, that's really enabled by our leading packaging capability at both high bandwidth, low latency. It allows you to make those decisions in a way that's basically using Moore's law in an optimal way. So if a piece of IP doesn't need leading edge, you can use trailing-edge capacity or you can even use external-foundry capacity in a way to let you optimize your CapEx and really optimize your cost for those pieces that don't need the leading edge. They aren't differentiated, right? You're not getting paid for it. You just have table stakes in terms of the product design. So we haven't said a ton about exactly which tiles are going internal and external other than Ponte Vecchio, which we said is basically there's a number of tiles, including compute tile, that will be moving to foundry.

Okay. That is helpful. Thank you for chiming in, and sorry for putting you on the spot there, Tony, but I figured you had come prepared. So back to you, Michelle, coming back to the product launches a little bit more nearer term. You talked about the products that you're expecting, that your customer is excited about. I actually wanted to zoom back a little bit -- zoom up a little bit and talk about some of the longer term. What changes had you brought on in your -- in the way you're handling the role versus your predecessor? So how has your organization changed over the years? And what is the evolution that is -- that we should be looking forward to?

Yes, that's a great question. So over the last 2 or 3 years, I really focused on how do we build a robust global sales and marketing communication team that's 100% focused on customers and partners. And so I think I said before, I thought we could be much better listeners. I think the key skill of a great sales marketer and communicator is listening, understanding the problems and then figuring out how, with your set of tools and capabilities, you can help customers solve problems. So that from day 1 has been job 1. Listening, and I've done things like -- so we have a whole robust set of customer surveys. We interview our key customers across 100 accounts by interviewing them for 45 minutes to an hour, 2 times a year to get their feedback. My team is paid based on that performance and whether or not we're meeting our customers' expectations. Every time we have an executive meeting, we talk about the customer feedback, what we're doing about it as well as continuing to change and mold our sales force, right? So if you think about 3 years ago versus today, we have a much stronger CSP sales force, much higher-educated sales force from a technical-depth perspective. We continue to hear from a lot of our cloud service providers, your sellers need to know what it's like to be in the middle of a cloud, what it's like to deploy, what it's like to do an upgrade. And so we've rolled out a university to be able to train all our sellers around the globe as well as go out and acquire a lot of talent for new areas. If you think about where Intel was 5 years ago on 5G versus where we are today, kind of being the lead silicon provider in the base station, we had to go and acquire a lot of talent that knew how to sell to those communication companies, that knew the marketplace, knew how to compete and help all of us not only build better products but really approach customers differently. And I'm trying to do that across all of our growth areas, whether it be CSPs, whether it be communications, whether it be edge. And then I think the last thing is really how do we show up in the marketplace and partner with our customers. We have been a steward of the ecosystem for so many years. But one of the opportunities that we have is how do we make our customers more successful because they chose to build with Intel. And we've changed a lot of our programs to be more oriented what customers were asking for. We've merged a lot of them because we used to have programs kind of for every segment. And we found customers were using and buying from all of Intel segments, and it was confusing. So we focused on a variety of things, Ambrish. And I hope through that customer survey process, we continue to evolve, right? You can't be stale because if you're comfortable where you are today, somebody is outpacing you. And we have a high level of trust based on those surveys from our customers, and they do not hold back in giving us really tough feedback. But feedback is a gift, and we'll continue to take that feedback, evolve and try to be more competitive each and every day.

Awesome. Yes. My clients don't hold back either, instant e-mail. You have calls, "Terrible, Ambrish." I have another question from the line. And this is actually an interesting one. How does the Intel view the possibility of PCs and ARM-based CPUs coming into the PC market? You've had an ARM license forever. Would Intel look into developing ARM-based CPUs as well?

It's interesting. ARM has been in the marketplace for a very long time, and I won't comment on what we're thinking from a road-map perspective, but we do believe x86 is a superior architecture and that our customers are asking for what that delivers. I will also say that we're always listening to customers and understanding what they need. And if you look at like our Lakefield products and the work that we did there to be able to give lower power, but higher performance, I think we're continuing to evolve that we -- the way we build products, Ambrish. And if ARM were something that customers needed in the future and it made sense for Intel, we would absolutely have that conversation with the customer. It has not been an ask that I've had to date.

Okay. Okay. Fair enough, fair enough. Before we sign off, I'd like to leave it -- for you to leave us with some closing remarks if you had any, just kind of whether on Intel, as an investment pitch on Intel or about what you're trying to do, either is fine.

Yes. Well, first off, I'd thank you for having me. And I would just say, I believe Intel's future is bright. Our customers believe in us, they've leaned in a lot over the last 6 to 8 weeks, asking how they can help, they're energized about our products. We have an amazing portfolio in 2021 and '22 that is going to be the volume of our shipments over the next couple of years. We have optionality now in the way that we manufacture. So when you think about, is it going to be manufactured by Intel or by a third party, we have more optionality now than we've ever had before, which gives our customers choice as well as comfort that they know we can make that optionality and that we can deliver those products. And we have really strong values across Intel, and customer obsession sits at the top. And we're going to continue every day to come in and do our best to deliver on behalf of our customers and help them solve problems because when they solve problems, their businesses grow and when they flourish, we flourish.

Okay. And I apologize that question did come in and we have a couple of minutes. The question was, was Apple's recent decision to in-source its chips something that Intel -- and I guess it is referring to the ARM announcement that came out at the forum, is it something that Intel expected, or was it a total surprise to you?

Well, we don't generally comment on specific customer details. So I'd say you probably need to ask Apple about that because those are confidential conversations between customer and supplier. But we understand their decision. Again, that's going to be a 2-year transition for them to move to their new products. And I believe that we'll continue to build products that Apple will want to have within their machines, and it's our opportunity to win those sockets.

Great. With that, I'm going to say thank you so much. I know we're all thinking about something else as well, given where we live. So good luck on that front.

Thank you. You too.

I appreciate your sharing your views. Tony, thank you very much. Thanks, folks. We'll be back in -- after a bit of a pause. Thank you.

Thank you, Ambrish.

Take care.