Advanced Micro Devices, Inc. (AMD) Earnings Call Transcript & Summary

Good morning. Why don't we go ahead and get started? My name is John Pitzer, and it's my pleasure this morning to introduce someone who probably doesn't need any introduction. We have, as our first semiconductor speaker at this year's conference, Dr. Lisa Su, the Chief Executive Officer of Advanced Micro. Really appreciate Lisa's time this morning. We've got about 30 minutes in this session to go through some fireside chats.

And Lisa, maybe I can just kick things off. You're coming up on your fifth anniversary as CEO of AMD. And I was going through some numbers last week. When you first became CEO, you guys were doing about $4 billion in revenue, and your market cap was about $1.5 billion. Today, you're going to get close to $10 billion in rev. And amazingly, your market cap has now eclipsed $100 billion. So to say that you've been busy over the last 5 years is probably an understatement. But I was hoping maybe you could kind of help set the table with this first question and talk a little bit about some of the investments and bets that you made 5 years ago that are bearing fruit today and kind of why you made those bets.

Yes. Great. Well, first of all, John, it's great to be here with you guys. Appreciate the time. Look, it has been a very interesting 5 or 6 years. I would say that when I took over as CEO, our focus was really on -- really figuring out what we're very good at. And our focus was on high performance. And when you are in the semiconductor business for a long time, you realize that it does take a while to really build the strength of that technology road map. And so our focus has been on the products and on the technology road map. We very much made some deliberate decisions in our CPU road map with our Zen products as well as our GPU road maps. And that has really come to fruition over the last couple of generations. So this year is a big year. 2020 is a big year for AMD. We are just launching our Zen 3 product stack. So you've seen the desktop version launch, just a little while ago, and we'll see notebooks and servers shortly. And what we've also done is launched our new -- our next-generation graphics road map as well. So it's really been about focusing on the technology, building upon every generation and ensuring that we also build trust and relationships with our most important customers.

Lisa, that's a good segue into my next set of questions. I really want to try to organize the discussion around kind of your end-market focus. And so my first set of questions is really on the client business. As you talked about, you just introduced kind of the Zen 3 platform for your client business. Wondering if you could talk a little bit about the performance schemes we should expect from Zen 2 to Zen 3. And I guess one of the questions, I get often is, now that Intel's finally in the market with their Tiger Lake 10-nanometer client part, how should we think about kind of your market share aspirations over the next several quarters? Is this going to be a more muted period of market share gains? Do we wait until Zen 4 and your 5-nanometer technology before a reacceleration? How should we think about that?

Well, yes. Look, we're very excited. First of all, the PC market has been a very robust market this year. I mean if you think about it, it's over $30 billion of TAM. Lots and lots of needs around desktops and notebooks, particularly this year. Our focus has been on continually improving our road map over time. So if you look at Zen 3, it's a great, great architecture. Our focus there was on improving single-threaded performance. So we saw 19% improvements in IPC. But we also did a lot of other things to make the products generally just better in applications. Whether you're talking about gaming applications or you're talking about some of the important server workloads, we really focused on sort of the workload benefits. So if you're talking about the PC market overall, we've nicely gained share over the last couple of years, very consistently, in both desktop and notebook. And we see Zen 3 adding to that. And frankly, yes, the product is better, but I think what you also see is that our customers are much, much deeper in collaboration with us, putting out better platforms. And so you see better notebook platforms. We have the best set of notebook platforms we've had in 2020, and it will get better in 2021 with Zen 3 as well as a number of desktop platforms. And our focus is not just on the consumer market, but also expanding into the very important commercial market that we've talked about before.

Well, Lisa, maybe you can expand on the commercial side of the business. Because as you pointed out, since the introduction of Ryzen, you gained about 10 points of share in the desktop market, you gained about 13% of share in the notebook market, but that's been mainly, I believe, more consumer-focused. Where are you in getting a better foothold in the enterprise? And what really drives that incremental share gain from here?

Yes. Particularly over the last several quarters, what we've seen is -- first of all, you're right, that a lot of that share gain has been in consumer. But if you look, particularly over the last several quarters, we started to make very nice inroads in some of the other segments. So for example, education is a segment that has done very well for us. Gaming notebooks is a segment that's done very well for us as well as commercial notebooks as well. So again, we have our strongest portfolio ever of platforms from our OEM partners. And we continue to build out our commercial sort of business development approaches, which allow us to win more of the Fortune 1000 and some of those key accounts. So I think overall, I think we've built the portfolio very nicely, and that's why we still see a very good opportunity for us to continue to gain share in the coming quarters.

Well, Lisa, you pointed out that this year has been a particularly strong year for the PC market. There is a concern from investors that the work-from-home, school-from-home phenomenon pulled forward some demand from future periods into this year. Kind of what are your thoughts on that? And I was having this conversation with an investor last week, and I -- first thing I did when I started working from home was buy a new PC. I haven't opened my notebook since, I think, March of this year. But God knows, as soon as I get back on the road, the first thing I'm likely to do is to upgrade my notebook. So how do you think kind of next year plays out as we get into a world where we hopefully have a widely distributed vaccine?

Yes. So I know we're all looking forward to that day, John, for sure. What I would say is when the demand signals first started, let's call it, in the second quarter, I think all of us who have exposure to the PC market were wondering, "Hey, is this a short-term type of pull-in phenomena?" And the truth is, as we've now gone through the last 6 or 7 months, we spent a lot of time talking to our customers and sort of their customers' customers. And I think what you'll hear is that, really, we think this is more of a longer-lasting situation where the PC has just become essential. And so the idea that a household, you could do 1 PC per household. Perhaps if you're only using it an hour in a day, makes sense. But now it's really about 1 PC per person, and you're much more -- you pay a lot more attention to the feature, function, capability when you're doing as much as you're doing as a productivity tool. So I think we are optimistic about the PC market, certainly in the near term as well as in the medium term. I think we see 2021 as a growth year for PCs, which perhaps we haven't heard in a while. And more importantly, I think the product portfolio that we have is sort of a very focused on some of these higher-end feature sets as well that we think will play well for our long-term share gain prospects.

And Lisa, just given sort of your share aspirations in the client market, do you feel that even if the PC market goes through a period of digestion in '21, that you guys should still be able to show growth?

We certainly do feel that we still have potential in terms of the representation of our market share in the PC market. We don't feel like there's any artificial ceiling or anything like that. So certainly, from the strength of the products, from the strength of the customers and the platforms, we believe, even if the PC market were go through some digestion, as you say, that we would still have an opportunity to grow. And that's certainly our focus.

Perfect. Switching gears to the server market, which is, I think, the market that probably gets the most attention on Wall Street. You've had a great run with your EPYC family, first introduced in 2017. You're now on your third-generation of EPYC. Similar question to my client question, how do we think about market share gains for Milan over the coming, call it, sort of year or so, especially in light of the fact that as we get into next year, Intel will start shipping their first 10-nanometer parts? Do we see another digestion period of share growth for you before Genoa? Or do you think Milan's going to be a pretty strong product in and of itself?

Yes. So as you said, the data center market is very, very strategic to us, and we've always planned it as a market that would take several years to really gain share and gain the reputation that you need with the customer set. I will say Rome has done extremely well and has continued to ramp. It's even continuing to ramp now. So we still see platforms coming on, new design wins and so on and so forth. I think Milan is a very, very strong product. Again, it benefits from the Zen 3 core. So it has all of those benefits of higher instructions per cycle as well as other improvements that we've made that were workload-specific. And I think beyond that, again, customers are now familiar with us. So in the data center market in particular, it's about building a reputation and building a brand and building a track record. And so as we see with Milan, we actually see that the Milan cycle should be larger than the Rome cycle in terms of just how it ramps. And we see new customers who are very excited about Milan as well as our current customers who are already engaged with Rome planning to migrate to Milan. So overall, I would say we're very optimistic about how Milan will do. And again, I would say that the market, in general, has been good in the data center market. And we feel within that market, our products are extremely competitive and will be more competitive as we go forward.

That's helpful. I mean, Lisa, oftentimes, investors kind of lump the server market together. But quite frankly, it's become almost as segmented as the client market. And to date, a lot of your success has come within sort of the cloud, hyperscale and the HPC market. I think that's probably an underappreciated story, especially with your heterogeneous CPU/GPU. I'm wondering if you could just spend a few minutes kind of talking about your positioning in each of the markets. And importantly, how do we think about growth in share, in the enterprise and what I'll call kind of the edge server market, the networking edge side of the market?

Yes. So as you said, the data center market does have subsegments that are important to understand. We have done very well in cloud and very well in HPC. And I think that's because these customers tend to be very, very technical in the way they evaluate products and technologies. And they also are able to sort of make decisions based on those technology attributes pretty quickly. And so we've seen a nice ramp, certainly as we went from our first-generation EPYC to second-generation with Rome. And we see that broadening as we go into Milan. So as I said, there will be new customers that have chosen Milan as their entry point to the AMD road map, and we're extremely excited with that. As well as customers ramping the number of workloads that they did with Rome, they're expanding the number of workloads that they're using for Milan and beyond. Now as you asked about enterprise, look, I would say enterprise is very similar to commercial PCs in the sense that it takes several generations for people to really build a reputation and a brand there. I think we've made strong progress. I think we have great partners with HPE and Dell and Lenovo and a number of others, who are now offering our products more broadly. You'll see a nice slew of announcements with Milan as the enterprise guys really get up and running in the first quarter of next year. And so overall, I think we're making nice, steady progress. I expect that we'll continue to make progress with Milan from a product positioning standpoint. And then, John, as I said, in this market, it's very, very important to have a strong road map. And so as we go forward, we're launching Milan, but then we're also very far down the road with our 5-nanometer Genoa product as well. And so that builds out sort of the long term as well as the short term.

Well, Lisa, I think as you were reestablishing kind of your presence in the server market, I think it was really important for you to put out milestones, both to sort of the investment community, but also to your customers around kind of share aspirations. Now that you're sort of in that double-digit share range, I think you guys have backed down a little bit for giving additional share targets, which is reasonable. But I'm just kind of curious, how should we think about the share possibility over a 3- to 5-year period? I always harken back to Opteron and that cycle where you got up to about 26% market share. And I guess as you address the question, one of our big views has been, yes, market share is important. But the other thing that's going on is as the world moves from just creating data to actually analyzing data, the compute TAM is just accelerating, and everyone's going to benefit. I'm wondering if you could address both kind of the share aspirations and kind of your view of the longer-term growth of just compute.