Good afternoon, everyone. Thanks for joining us. For those that don't know me, I'm Vivek Arya. I'm from Bank of America Merrill Lynch, not from AMD. I'm delighted to be part of AMD's 50th year anniversary and really glad that I'm having the opportunity to question the management team and have a discussion about some of the key milestones the company has had over the last many years, but then really focus on what's ahead for the next handful of years. Lisa doesn't need any introduction. Mark Papermaster, Devinder and Forrest are on. What I'll try to do is I'll start with a few questions and then maybe invite one or two from the audience as well. 50 years, right?
Yes.
I guess time flies when you're having fun. AMD has certainly had a lot of fun over the last 50 years. Maybe Lisa, if you could just give us a quick historical journey of what has happened over the last number of decades, but then really focus on the last five years since you took over. What have been the key areas that you have emphasized and you've focused on to help set the company for the next decade?
Very good. Thank you, Vivek, and thank you all for joining us here in New York and for those who are joining us on the webcast. We're really happy to have you here to help us celebrate our 50th anniversary. Look, 50 years is a long time. I think it's a long time for any company, but especially for a technology company in the markets that we're in. I think what's really special about AMD is really our history and our heritage of always pushing the envelope. I think that's what this company's about. When you talk to our engineers, that's what gets them up in the morning, that's what gets them to work multiple hours, nights and weekends, because we love pushing the envelope, especially on high-performance computing. I think you've seen over certainly the last couple of decades that we've done that.
I think over the last five years, what I've been trying to do and this leadership team that's here with me have been trying to do is take the best of AMD's heritage and combine that with really a strong business and execution team. If you think about it, we love putting out products. We get up every day. I tell people we get up every day to put out great products. Now, great products also requires a lot of predictability, good business management, good operations teams. All those are part of it. We look at this as we want to really be the leader in high-performance computing. That's our goal. We think that's our sweet spot. That's where we do the best in when we're pushing the envelope on computing and that's our goal.
Excellent. Lisa, I will be remiss if I don't just ask you about the overall semiconductor industry because you're chair of the Global Semiconductor Alliance, and it's kind of the best of times and the worst of times for the industry because you have an industry that's finding a lot of new applications. There's 5G, there's AI, there's cloud computing, autonomous cars. At the same time, Moore's Law is slowing down, so I imagine that presents its challenges. You also have a situation where people who used to be customers could potentially be competitors. China's ambitions in semiconductors is growing. You have a number of cloud customers who want to insource a lot. What are the things that you focus on, again, for the next three to five years in terms of what you can control to make sure AMD has a sustainably differentiated future?
Yeah. I think the semiconductor industry is a great industry. I think we're all huge fans of the applications that we power. Whether you're talking about the data center or you're talking about gaming or you're talking about PCs or you're talking about mobile, we're enabling a tremendous amount of innovation and actually opportunity for the world. I think the industry is good. I think, look, every industry has its ups and downs, and certainly, in the last few quarters, there have been discussions about certain pockets of softness, and certain pockets of inventory and those kinds of things. We all watch those, obviously, on a quarter-by-quarter basis.
If you look at it from more of a medium term or long term, there has never been a time where we are generating more data and we need more compute to analyze all that data, and there are more applications. I think it's a great time to be in the industry. For us specifically, our focus is on differentiating applications. We want to do things that other people can't do. When we think about our CPU roadmap and our GPU roadmap and the systems that we're trying to enable, we're trying to really push the envelope on what you can do with computing. Our goal is to think about it quite in a different fashion. Yes, Moore's Law is slowing. Moore's Law is still very important. Process technology is still very important, but it's an enabler.
That's some of the reasons that we've adopted our chiplet strategy that allows us to really take the best of both worlds in terms of technology. It's the reason that we've invested in hardware and software around CPUs and GPUs, because we think we can create very differentiated products.
The one interesting aspect of AMD is that you have expertise in CPUs, also expertise in GPUs. As we were discussing at the lunch table, you were mentioning opportunities that perhaps could come over time, where there are interesting ways of combining them, which I think AMD can do uniquely versus your competitors. Can you tell us a little bit more about that?
Yeah. Look, we're really excited about our CPU roadmap, whether you're talking about Ryzen and EPYC or our GPU roadmap, talking about Radeon. I think that continues to be sort of our foundation. When you look over the next two, three, four years, I think what is even more exciting is the new opportunities when we start putting these components together. Certainly, you see in some of our PCs and our semi-custom products that we're doing APUs or a combination of CPUs and GPUs on a single chip. When you look forward, I think as we look into the data center applications, there are many more opportunities for us to take our CPU technology with our GPU technology with much of the IP that Mark's team has created to really put together some very high-performance system solutions.
Got it. Glad you mentioned Mark. A few years ago, when you used that phrase heterogeneous computing.
we were not completely aware of what that means. Can you describe for the audience what exactly that means and what has been the impact of that for AMD?
Well, heterogeneous computing is just recognizing that not every workload needs the same type of engine. That you can really optimize in a heterogeneous environment. As Lisa mentioned, APU gave AMD a big advantage because as we brought the assets, the Radeon assets, and the strong CPU heritage of AMD together, AMD had to learn heterogeneous computing before anyone else, as you say, Vivek, even had the term.
Right.
The other thing that we recognized a few years back as we restarted our CPU roadmap around the Zen family was how high performance played into heterogeneous computing. We looked at where the workloads were coming, what did analytics need, the emerging areas around machine learning, and it not only needed high performance across CPU and GPUs, but how do they play together and how do they play in an open ecosystem so you could even pair FPGAs or specialized ASICs and have a software ecosystem that allows them to play together. We had an early lead on heterogeneous computing. We called the ball that the industry workloads were coming our way to both high-performance computing and heterogeneous, and we're really pleased to really be enhancing that product set behind it.
Got it. You are just in front of the first time when AMD has ever sort of had a lead on the manufacturing side with your foundry partners. Maybe Mark, from a technology perspective, what does that mean? Then maybe Lisa, from a product and competitive perspective, what is the impact on AMD?
Well, from a product standpoint, it's a play that you have to call well in advance on the semiconductor node. We placed a big bet on 7 nm several years back, and you have to manage the risks. We saw that 7 nm could bring a huge difference because it doubled the transistor density, and it allowed the power to go down basically in half. What does that mean? Take our server socket with Rome that we announced in our Next Horizon event in the fall. Doubled the number of CPUs in the same power envelope. I mean, that's a huge impact. We called the ball on 7 nm in terms of a technology standpoint, and we worked very hard with our foundry partner at TSMC, with the CAD and EDA industries, because it was a heavy lift.
We knew 7nm would make a big difference for our products, and we really got ahead of it.
Yeah. I would just add to that, Mark, these decisions are made many years advance. I like to say, whatever decision that we're making today, we'll find out in three or four or five years whether it was the right decision. It was the same thing for us in 7nm. We spent quite a bit of time debating what is the right thing to do. From our standpoint, the key thing was to have product differentiation. I think process technology is an enabler. It's not necessarily the determinant of the winner or the loser. However, the process technology choice in 7nm together with the architectural choices that we have made, I think puts us in a very unique competitive position. I mean, it's a very exciting place to be.
It's also a lot for us to do, our thought process is how do we combine the best that the industry has to offer in terms of process technology, with some very differentiated architectural choices.
The other thing is we added in a smart way. You look at the chiplet approach we've been working on for years. We've architected with our Infinity Fabric that opportunity to leverage chiplets. A great case in point, when you look at that CPU implementation in 7nm, very small, easy to manufacture, and form factor.
Got it. I think that's the interesting part, that the cost structure, the yields, even though you are in an advanced manufacturing node, that it could be much better than just having larger and larger monolithic dies. Do you give up anything if you're going to a chiplet structure?
You don't give it up if you really architected that approach. Again, you can't do that at the last minute. You have to think about it well in advance. What we did across the AMD family, was we architected a protocol, a way in which to connect all of our IPs together in a consistent way. What it allows is actually modularity. As we looked at how we could deploy chiplets to make those seven-nanometer CPUs easy, in fact, to manufacture.
It played right into our architecture that was enabled by that Infinity Fabric, and we designed that in several years back. As Lisa said, it's a long lead time in making these bets.
Got it. You launched Ryzen, Lisa, a few years ago. What aspects of Ryzen do you think have most resonated with customers? Just along with that, how has the traction been with your commercial or enterprise customers also?
Yeah. Look, we are very happy with how Ryzen has, I like to say, grown up. We thought about Ryzen with the idea of the PC industry had not had a lot of change. Frankly, there were very incremental improvements that were happening in the PC industry. We thought with Ryzen, first with the Zen core family that we developed, it was the idea of we thought about it as multiple generations, where each generation we would keep making improvements. With Ryzen, the first generation, we started with our desktop, especially the enthusiast space. The thing that I love about the enthusiast space is those users know what good technology is, and they look at every single spec, and they decide what's the best processor. We were able to offer just a tremendous amount of performance per dollar, as well as just overall performance.
That has then gone into more of the OEM customers. Now you see many more notebooks that are being launched with AMD. If you look at the number of platforms that we have, in 2019, we will have over 50% of the platforms that we had in 2018 in the notebook space. That also goes over to the commercial customers, because the commercial customers, again, it's a longer sales cycle, but what they want is good value for the product. I think Ryzen has done very well. Last year, we gained four points of share in the PC market. We believe the PC market continues to be a good market for us because of the size of the market, as well as our technology fits very well into it. We are going to keep pushing the envelope with Ryzen.
We're looking forward to the launch of our third generation Ryzen. You'll hear a lot more about it in the coming weeks, lots of excitement around Ryzen.
Got it. Just one other thing on the PC side. There's a lot of talk of component shortages and whatnot. Do you think that has distorted the demand signals in any way that once, let's say, your competitor's supply situation is resolved, that you could see a different market share scenario in the second half?
Yeah, look, I think shortages are never good for any industry. It's difficult for our customers and for the supply chain. From our perspective, I don't think shortages are a significant part of our business drivers. Most of the shortages are at the lower end of the market, and those are low ASP products. I believe that as we look at the PC market, I think it's a multi-year growth engine for AMD.
I see.
It's one of those areas where I believe that every year our product gets stronger, more competitive. People recognize the brand, recognize what we're trying to do with the technology. We view it as a multi-year growth driver.
Got it. One question about the full-year outlook. I want to bring in Devinder into the conversation as well. Frankly, going into the earnings, after what we heard from Intel, I think investors were somewhat apprehensive because the spending environment is not that great. It was quite a positive surprise to see you kind of maintain Q2 and then obviously full-year expectations of high single-digit sales growth. Can you maybe start, Lisa, with what are the kind of the challenges and the opportunities as you look at this year on a whole? What could make it better than that number? What could be a challenge in the second half?
Yeah, absolutely. I'll start and maybe Devinder can add. Look, as we think about 2019 is an incredible product year for us, probably our busiest product year in recent memory. We're really launching new products across the entire portfolio, so across our PC portfolio, across our gaming and graphics portfolio, as well as our data center portfolio. That is really the crux of our story. Our story is we have great technology. We think we are underrepresented in these markets. As our new products come out, we have the opportunity for revenue growth. Now, there is noise in the system. We all hear the noise in the system. Some of the comments around the data center and inventory in data center customers, I think is valid.
Yeah.
We've heard that from customers as well. Our view is most of the customers are saying it's a first half phenomena, and things get better in the second half. Our year was always second half weighted anyways. We view it as, look, on the things that we control, we are on track to where we thought we would be at this point in time. Our products are on track. The performance looks really good. The customer interest, sort of the customer pull for what we're trying to do is very good and we need to really execute, and we'll see how things develop. Maybe Devinder, do you want to add to that?
Sure. The only thing I'll add is a lot of questions from the callbacks in particular, as Lisa has covered the revenue and the new products, is the margin side of the equation, right? It feels good to be. We laid out, actually right here in this room many years ago, a long-term target model of margin between 40% and 44%. Q4, we ended at 41. Q1, we finished at 41, and we are guiding to 41 in Q2, and then with the mix of the product, obviously, will determine what the margin is for the rest of the year, but it feels good to be within that range as we start a new year here.
Got it. The one interesting question and debate that came up with some investors this morning was, are you overspending or underspending in Q2, right? I know there's never the perfect level of spending, but you have a great opportunity in the second half because of the pipeline, because of the opportunity to gain share. Q2 OPEX levels were very high. I don't know who gets the credit and who gets the blame.
We're not going to let Mark answer that question. Mark's a spender.
That's right.
Look, we believe that we are investing appropriately in the business. It's always a balance between how much we invest versus how much drops to the bottom line. Our guidance for this year is approximately 29% expense to revenue ratio. That's where I believe we will be. I think the expectation is that we're investing at the right level to drive the type of growth that we are trying to achieve. I don't know if, Devinder, you want to-
I was asked the same question a couple of times at the lunch table is, it's just very targeted, right? We've heard many times over the years, do you have enough money to spend to compete with our competitors? I would say it's mostly the target and the focus with which we do what we do. There are many other things that Lisa and Mark will decide, "Hey, we're not going to do that because now is not the time, or we're not going to do it from a targeted standpoint." The products that we have today, as you have heard us say many times, are chasing a TAM of $75 billion. We are sitting at about $6-plus billion of revenue, we can invest and make these products multi-generation and better and more competitive.
There's a long way to go before we ever think of other opportunities that we need to go pursue and dilute ourselves from a viewpoint of the focus and attention.
Got it. Devinder, you brought up gross margins at 40%-44% targets.
Yes.
You are about 41% right now. If you think about just the pricing environment and what impact does that have on margins, how comfortable, how confident do you feel about that target? Do you think there's opportunities to do even better over time?
It's going to depend on our product mix. I think we are sitting here in Q2 of 2019, where we have, as Lisa said earlier, across all businesses, whether it's PC, graphics, or the server piece of it, we're introducing new products on a new technology node. The mix of that product and the ramp of those products relative to the rest of the products will determine the revenue. The new products we're introducing, including the seven nanometer products we just talked about, in aggregate, are delivering greater than 50% gross margin. As we introduce more and we ramp higher, it could be better. Otherwise, we have said that greater than 41% gross margin is our goal for 2019, having delivered the 39% that we did in 2018.
Got it. Let me bring in Forrest into the conversation because your products will be a very important part of meeting a lot of these targets, so no pressure.
Very important.
When you look at the initial response to EPYC, how has it been in terms of what the cloud customers are using it for, what kind of traction you are seeing with enterprise? How has the response been so far? What attributes of the product do you think are resonating most with customers?
We've been very pleased with the reaction of the market so far to our re-entry into the data center CPU side. We designed the first generation of EPYC really targeting where we thought the most important markets were and where the most important markets were going to be. We took that Zen core that Mark's team had been working on for many years, and we developed a multi-chip part that was really designed to excel in virtualization, in cloud, and in HPC. In those market segments, we've gotten very good traction with EPYC. Very pleased to be adopted by some of the leaders in cloud computing, obviously Amazon AWS, and they've continued to launch additional cloud instances even in this past quarter with us. Likewise with Microsoft Azure, with Tencent, with Oracle, and with others, very good traction on the cloud side.
Probably better traction even than we thought on the HPC side. We've seen a lot of interest in adoption of EPYC in high-performance computing, and we think that'll continue to grow. On the enterprise side, both for virtualization as well as for big data and things like Hadoop, we've seen excellent uptick. We've had some real marquee names adopt that first generation of EPYC. Really across the board with first generation, we think it's gone very well. That's just the beginning. The 2nd-generation product was always designed to dramatically increase the performance and increase the competitiveness of the product, not only continue a winning strategy for those segments and further open them up, but also be able to attack and do so in a highly competitive way in a large number of additional workloads.
We feel really good about what the team has done in terms of delivering Rome, that 2nd-generation EPYC. We'll be shipping that in production this quarter ahead of the launch in Q3. We think we're right on track. We think the adoption with that product and maybe most importantly, the very predictable execution that we've established now with Rome and that our customers can see, even beyond that for the long-term roadmap, puts us, we think, in a great position to continue to grow our share in all of those segments.
Right. Forrest, you used to work at a major hardware OEM before. I'm just trying to picture myself in your shoes that you have a supplier who used to be very strong many years ago, comes back to the market, with a very different value proposition. How do you react? What things need to stand out for you to make a decision, or for AMD versus the incumbent who has a lot of resources, has a very big product portfolio. What's that decision-making process like?
Well, there's a couple of aspects to it. Obviously, first and foremost is can I build differentiated products for which there's end customer demand and end customer preference? I think, check with both EPYC first generation, with those throughputed performance advantages that are articulated in those segments, as well as now with Rome, I think that we've achieved that. Very important, both for the end customer as well as the OEMs, they've got to be able to trust us as a partner, as a supplier of high-performance technology. Actually, both Mark and I, both sat in the chair at a major systems OEM running large server businesses. The thing that we know is of paramount importance is of having compelling products with boring execution. Rock solid predictability that our customers can depend on.
Both of us have really emphasized that as we've built out the execution plan over the last few years. We're very confident that we can be that predictable, high-quality supplier of high-performance components for the long term.
All right. Maybe one last question, then I'll invite a question from the audience. Lisa, from a supply chain and an operations perspective, how well is AMD positioned to execute? The opportunity is clear. Customers do need an alternative. How is AMD positioned to make sure you can execute this time?
Yeah. I think a lot of our focus has been on this execution engine, the execution engine has many different parts. On the design side, I think we have a very strong execution engine. You can see it with our delivery of Zen, now our delivery of Zen 2, we're far into the development of Zen 3 and beyond. On the supply chain and operations side, actually our supply chain leaders here, Keivan Keshvari, I think what we have is also a very strong execution engine. Our partnership with TSMC is very core to that. I think being a leader in their seven nanometer technology, we wanted to make sure all of the yields and the capacity were there, everything that we're seeing is we're on track for that.
The other piece of the execution is on the customer ramp side, ensuring that we help our customers get ready in their systems and platforms. We have large customer engineering organizations across the product lines to make that happen. I would say there's a lot of execution to do, but I'm very clear that we have the right execution engine to get that done.
Right. With that, if Ruth, you wouldn't mind helping me with pulling questions from the audience.
Lisa, question for you, just to carry on from your last point. Clear design market share gains across the business on the demand side. On the supply side, is supply unlimited at TSMC, or how should we think about that for our models?
I'm sorry. The mic is not perfect. Was the question around how confident are we that we'll have the capacity to supply the demand?
Yes. How should we think about on the supply side capacity from TSMC at seven nanometers? Is that unlimited?
Look, I think the way we've designed our roadmap has been very thoughtful. I think as it relates to capacity at TSMC to satisfy the type of growth that we're talking about, I think we feel very comfortable that we'll have access to that capacity. As it relates to just how we've put things together, for example, on our Rome chip, we have both seven nanometer and 14 nanometer components. GlobalFoundries is also a piece of that for the 14 nanometer I/O die. Like I said, I think everything that we see is we have the capacity available, and we have the execution available to drive the types of revenue growth that we're talking about.
Can we just go back to margins? I hate to go back to the mundane. Your roadmap is 40%-44% for the gross margin. Is that a roadmap like going from here across the country, or is that like going from here to Cleveland and by the time you get to Los Angeles or San Francisco or San Jose, it will be much higher or somewhat higher?
I like the way you answer the question. Maybe I'll start and Devinder can add.
Using aeronautical terms.
Yes. The geography of the country. Look, I think the 40%-44% margin target, as we did with all of our long-term targets, by the way, were set at a time, actually several years ago, and was the idea of, look, we have multiple product lines. We're not going to count on every single thing going perfectly because life is, every single thing doesn't usually go perfectly. We believe that 40%-44% is a very good range, and at the time, we were sitting in the low to mid-30s. I think our goal is that if we're putting out great products, we should be increasing our ASPs and our mixes. I think 40%-44% is the right target right now. We will see how things go as more of our product portfolio is transitioned to the newer products.
I don't know if you want to add to that, Devinder?
No, I think that's well said. The other thing I would add is we do have a mix in the business. You hear about our semi-custom business, which is a captive business that we have with certain customers. Typically with that business, because the model is different, where we get NRE funding upfront to customize the chip, the gross margin when the product is sold is lower than corporate average, and that does have an impact overall on the gross margin.
Do we have more questions in the room?
Just a quick one because it hasn't come up too much is your, the combination of the CPU and the GPU in the future into one architecture. Where would we see that add the most value and why? Is there any sort of timeframe to where this APU, whatever nomenclature you guys are tying it to, where it would show up and become a real product?
Yeah, maybe I'll let Mark and Forrest address that.
Yeah. I'll start right off tangibly. Of course, we understand, as I mentioned earlier, with our heritage of APU, how CPUs and GPUs can play best together. Already in any of our PC applications or workstation applications APU, you're already seeing workloads that can traverse across CPU and GPU as sharing that same memory. The contents of the memory don't have to move. It's very efficient. Translate that knowledge of how workloads can take advantage of that to supercomputer class applications. The first thing you need to do is have very efficient data movement across the CPU and GPU. We are a leader at AMD at adopting PCIe Gen4. We're rolling that out across our products in 2019. That doubles the I/O bandwidth between the CPU and GPU. What you'll see going forward is we can even further tighten that connectivity.
There'll be future applications today. Software can't really take advantage of coherency and big compute across CPU and GPU, but that's coming. We have all of that capability in our arsenal at AMD, and it's about matching the capability of our hardware as we push up CPU and GPUs working closely together, into high performance with that software enablement. Let me turn over to Forrest for that.
No, Mark, I think you had most of the key points. I think that coherency point we think is an important one. It is a different way for developers to think about writing applications, but it's one that offers them tremendous advantages, both in terms of performance, in dispensing with superfluous memory copies. It even lowers power consumption, which again opens up more performance. We think there is a significant shift coming that will open up higher levels of performance for quite a number of data center applications.
David?
Thank you very much. Actually, just to follow up on that point, Forrest, and ask very specifically, are there plans for an APU data center product or to use your chiplet capability to be able to connect GPUs and CPU chips within the same package?
Yeah. I think we don't want to get too far ahead of our skis in terms of announcing the details of future products. I think the macro point of being able to interconnect CPUs and GPUs together, really to have them both acting as peers, and to be able to both share that certain virtual memory address space, really offers some pretty compelling advantages for software developers and for performance. The other thing that I'll say is there is no one right answer in terms of the ratio that you want for a given workload between the CPU and the GPU. We need to take advantage of the modularity and the Infinity Fabric capabilities that Mark's team have created to be able to offer a range of solutions depending on the workload and the characteristics that provide the maximum performance for that workload.
Just one question. I know the GPU topic has come up. Just maybe, Lisa, near term, how's the inventory situation in GPUs? Because the industry suffered a lot over the last six-plus months after. Longer term, you are finding new applications for the GPU. You announced a big game streaming service. If you could just give us kind of a near term what's happening in GPU, and longer term, what are AMD's ambitions with that whole game streaming project?
Yeah, absolutely. The GPU, I presume you're referring to the channel inventory-
Right
That was a result of some of the cryptocurrency build-up last year. I think the channel has improved significantly. As we look in the first quarter, we saw channel inventory levels come down. We also saw accelerated sell-through of our product. In general, I think the channel is in better shape. From our standpoint, when I look sort of at the medium term for GPUs are a very good market. Whether you're talking about gaming or you're talking about gaming at the PC level or you're talking about cloud gaming or you're talking about GPUs in the data center, there are lots and lots of different applications. We see it as a growth business for that. We're very honored to be part of Google's cloud streaming solution with Stadia. We continue to work with several others who view this as a market.
Sort of our vision around graphics and gaming is we want to be the provider of choice. Whether you're talking about PCs or you're talking about consoles or you're talking about the cloud, we believe that the AMD Radeon solution together with what we can do in our CPUs is a very strong solution.
Right.
I'll ask a more long-term question. This being the 50th anniversary, obviously there've been huge changes in the chip business, semiconductor business in the last 50 years. Just asking for speculation, 50 years from now, what will be the biggest change between now and then?
I think that's a CTO question. Doesn't that sound like a CTO question?
Well-
I gave him at least 10 seconds to think of an answer.
Well, I'll tell you what, it's funny you ask. I do think about this. The funniest thing is that we always think that there's some barrier to innovation. Well, I hear this all the time. With Moore's Law slowing, et cetera, doesn't that mean that the pace of innovation will slow down? It really doesn't, because what you've seen time and time again in history, what I see for the next 50 years is innovation does rule. The trends that you're seeing of everything being connected, of everything becoming more and more powerful and shrinking in size, those continue. The biggest way to think about how it affects our lives, both our work and personal lives, is the cognitive capability that's coming in on top of that.
It is the techniques that have been pioneered on artificial intelligence machine learning that allow these devices to become much more useful for us. They will more and more manage all that explosion of data around us and make it more and more useful to us over time. From an AMD standpoint, the next 50 years, it looks really, really bright because this trend means that there will be an insatiable demand for more computing, to really give us these more personal data analysis and services brought to us. Eventually, it'll just get woven in the fabric of the clothes you wear, everything around you. Computing will be a part of every moment of your lives. Okay. Before we close the webcast, Lisa, maybe any final words? What would you like investors to take away from these meetings?
Well, first of all, again, thank you all for joining us. It really is a very proud day for me, the leadership team, and all of the 10,000-plus AMDers. Look, our focus is on being a great company. When you think about what great means, I think that's great from a product standpoint, that's great from a customer standpoint. We want our customers to believe that we are their best partner. Also from a shareholder standpoint. I think our focus is on the long term, although we manage every quarter as it comes. Our mantras are about building great products and having very deep customer relationships and having strong financial returns. Thank you again for joining us and being part of this celebration.
Right. Thank you. That's the end of that.
Before we finalize today's festivities and celebration, if everybody could hold your seats for a moment, we'd like to do a 50th anniversary toast. Thank you.