How AMD's Lisa Su Is Thinking About AI
ADI IGNATIUS: We're going to start with an executive who we think is the perfect start to our day. Her company is working at the cutting edge of the technology that we're all thinking about, artificial intelligence. As chair and CEO of AMD, Lisa Su has, over the past decade, transformed AMD into a leader in high performance and adaptive computing and made it one of the fastest growing semiconductor businesses in the world, with customers ranging from Subaru and Tesla to Microsoft and Google.
An electrical engineer by training, with a bachelor's, master's, and doctorate degrees from MIT, Lisa is also part of the new breed of tech sector leaders who combine deep, technical expertise with strategic and management prowess. She's a proponent of responsible risk taking, and she's probably the most prominent woman leader in tech. She's here to talk about the future of AI, the opportunities, the risks, and how we can best harness its power. Lisa welcome. LISA SU: Thank you.
It's great to be here with you this morning. ADI IGNATIUS: Well, it's amazing to have you here and amazing to have you here kicking off this event. To the audience, as Alison said earlier, you can put your questions, any questions you have for Lisa in the Ask the Speaker chat. We'll get to as many as we can later. So Lisa, let's start with a basic question. For those who don't know you, maybe who don't know AMD beyond that it makes semiconductors, can you tell us a little bit about the company and how it has evolved? LISA SU: Yeah, absolutely.
Well, look, it's really great to be here with you. AMD, we are a semiconductor company here, and really, our focus is on building the bleeding edge of semiconductors, so highest performing, compute capability, as well as adaptive computing. We are in sort of lots of different areas, so if you think about large data centers in the cloud or enterprise, these are areas where we spend a lot of time, as well as in embedded devices. So think automotive, industrial, your test equipment, some of your aerospace capabilities, and then also in end user devices.
So if you think about things like your PCs, your notebooks, your desktops, as well as if you happen to be a gamer, we're in a lot of your gaming consoles and things like that as well. So lots of different areas, and the key underlying all of that is high-performance computing. ADI IGNATIUS: All right, that's super useful, and I'm going to have you be a little bit of a spokesperson for AI today.
As I said before, we're all thinking about it. We're all trying to figure out, is it coming after us? Is it a tool we should use? So you're more of the front end of all this. What excites you about the technology, and how do you see AI? And maybe I'm talking about generative AI evolving in the next phase, let's say. LISA SU: Yeah, absolutely. So I think if we take a step back and look at the whole idea of AI, or artificial intelligence, I mean, it's actually been around for quite some time, but what's really happened over the last-- I don't know, let's call it 15, 16 months or so, with the advent of ChatGPT coming out and generative AI, I think what we've all understood is there's a superpower here.
I mean, AI is really the most important technology that we've seen over the last, I don't know, let's call it 50 years. And when we look at the capability, it really has the capability to augment so many things. So I mean, to your basic point of, do we need to care and how important is it, it is really an opportunity to really accelerate all aspects of our business and frankly, of humanity as well. And I think of it as advancing things like research and health care and all of those important aspects of our lives, as well as our businesses, as well as our personal productivity, as well as things like content creation and all of these things are opportunities for us to use this technology that is now, let's call it bringing it much closer to home, where everyone can really practice AI in their daily lives. ADI IGNATIUS: Are there use cases you've seen that maybe I haven't seen yet that have you excited? Anything you want to share with us? LISA SU: Well, I'll tell you some of the things.
First of all, there are so many, so hopefully, everyone has experienced it in one shape or form. But perhaps the things that excite me the most are opportunities to transform things like health care. I mean, I view this as, when we think about something like that that's actually very personal and being able to think about accelerating things like disease research and diagnoses and how doctors care for us, I think that's super exciting. And each of our businesses, I think there are huge ways of acceleration. I mean, I can talk about some of the things that we're doing at AMD. We're accelerating the way we build chips and for us, everything is about how do we get more high-performance chips to market at higher quality, at better performance, at lower price and cost.
All of those things are things that I find very exciting. And then there's just the personal stuff that we all get to play with and see that it really can change each of our personal productivity as well. ADI IGNATIUS: If you're watching this, if you have questions for Lisa Su, please put them in the chat box, and we will try to get to some of your questions later. So Lisa, I want to drill down a little bit on this. So every conversation about AI that I have eventually evolves into something very dark. Is AI an existential threat in some way to not just our jobs, but to our very existence? I'm assuming you're a relative techno optimist, but help us out.
If AI is going to be a force for good, how does that happen? Will technology save us from the downsides of technology, or do we all, all of us, need to be contributing to the discussion now to make sure we don't get the worst possible outcomes later? LISA SU: Well, as you said, I'm probably a techno optimist, but I'm actually a very, very pragmatic way of thinking about this is, the technology is not perfect. As good as technology is, we're still in the very early stages of the deployment of AI. And we do know that the AIs are not always right. And so part of what we have to do as a set of leaders is figure out how to use the technology for good and also, protect the downsides.
And look, I think this is a very vibrant conversation. I think all of us are learning in the process. I will say that I've personally learned a ton over the last 12 plus months in terms of how to apply AI even within our own company and also, talking to many of my peers, how things are going. And I think we all recognize that we're in a learning process, but the key is to be very active in that learning.
So my belief-- and I know there's a lot of doomsday theories about how AI is going to take over all of our jobs-- I actually am a subscriber to the belief that what we have to do as leaders of companies is to really learn how to harness the power of AI and also, bring our employees along with that so that we're actually making our employees more productive, and we're able to make our companies more productive knowing that there are some areas where we have to be careful with the use of AI. ADI IGNATIUS: Yeah, that's helpful. There's also another aspect of this, which is just the balance between speed, bringing products out to the market as quickly as possible, now that there is a market, versus caution, and that's reflected-- as you said, you're learning and there are things we don't know yet about the technology.
How do you think about-- from where you are at AMD-- how do you think about this balance between speed and caution? LISA SU: Yeah, I really believe in fast experimentation and implementation, so I don't believe the answer is, let's slow down. I think what we have to do is experiment, where we've spent time. We actually have a responsible AI council. I think all of us as leaders, if you're leading companies or teams, you have to think about how to utilize the technology responsibly. We think about things about intellectual property, how to protect our intellectual property, as well as protecting our customers and our partners' intellectual property. But that being the case, I think the power of AI is finding those use cases that give you very, very significant return on investment.
And we've seen in some of our workflows, like in some of our design workflows, we've seen what used to take weeks and months really come down to days. And when you think about how valuable that is to your enterprise, you have to really push the envelope on using the technology. And there are lots of people who are out there to help in terms of experiences. I know that it's a very active conversation whenever I'm talking to my peer CEOs these days in terms of, what are you learning? Where are the use cases that are most beneficial? What are the things to be careful about? So I think this active dialogue is really helpful. ADI IGNATIUS: And I be interested in your advice for people who-- well, let's say when ChatGPT came on the market, lots of people experimented with it and played around with it.
And now, I don't know what wave we're on, but now it's, OK, but how do I actually use it? How do I actually apply it in my company? You mentioned health care, and people often mention health care as a clear use case, but that's very specialized. For the general audience here, what would your advice be? How do people figure out-- I guess there are two things, how to protect themselves against being disrupted by AI solutions, but then maybe more pertinently, how do I use AI to improve my business, whether it's efficiency or something else? What's your advice for people who are even just trying to think through that problem? LISA SU: Yeah, I would say, again, look across the use cases and the workflows in your business. The places where it's obvious. Very near-term successes can be in things called co-pilots or where AI is actually a helper to someone, to your employees. And I think about these types of copilot exercises, whether it's on the engineering side, we're using co-pilots to help us design code and really write code and to help us look at test cases and use cases, improve our quality, those kinds of things. When I look at things that are more sort of business oriented, we're looking at how we use AI in our marketing and our communications and our content creation.
Again, these co-pilots will allow you to, let's call it get close to the answer, and then of course, the final touches are being done by your expert employees. There are many, many cases like that, through every enterprise, where you can think about workflows, where you can accelerate your time to get an answer. The places where, of course, you have to be a little bit more careful are places that you would rely more on the AI itself to come up with the answer. And there, you have to do a lot of testing to make sure that you get the right answers.
But again, my advice is, lots of pilots, experimentation, and then figuring out where it has the most value. We've certainly seen, as we've deployed AI across our business, that there are some places where very high, value very low barrier of entry, and then there are others where, frankly, the tasks are you have to put sort of a lot more work into making sure that the models and the AI are more adapted to your particular use case. So lots of experimentation and really looking at where you can get the most bang for the buck in the near term. ADI IGNATIUS: Yeah, thank you for that. So here's an audience question. This is Melissa Quillen.
Not sure where Melissa is, but the question is, when it comes to AI, how are you anticipating returning real time data, data mining so that you can pivot your business almost immediately to current trends or to resolve issues that pop up? LISA SU: Yeah, absolutely. There is quite a bit of work, and we've also done work ourselves on looking at things like being more predictive in sales cycles and looking at some of the data that comes in to those trends. I would say that requires a bit of training on your business, because every business is different, and there does need to be a bit of training on your specific data. But I do think that you can get some very nice patterns and trends that come give you insights of where to dive to the next level of detail going forward.
ADI IGNATIUS: Yeah. So I want to ask you about your run at AMD. You've been CEO now for about 10 years. You said early on that one of your goals was to bring focus to the company.
How do you determine which business to prioritize, and how do you get buy in? How do you think about focus in that role? LISA SU: Yeah, so I've been at about 12 years, CEO for almost 10 years, and one of the things that is true in every business around the world is that you have more opportunities than you have people or resources or leadership bandwidth. And so for us at AMD, it was deciding, really, what are we going to be best at? And our heritage has been one of high-performance computing and really building at the bleeding edge of technology. And that was, really, our focus site.
So there were some things that we had to choose not to do. Like, for example, mobile phones are a very interesting part of semiconductors. There are lots of great needs in that area.
That wasn't the perfect area for AMD, and we just had to really choose the things that we were best at. So our focus was high-performance computing before high-performance computing was sexy. And now, we can say between high-performance computing and AI, we are in, perhaps, one of the most exciting areas, if not the most exciting area in semiconductors, and it has a lot to do with our heritage and focus. ADI IGNATIUS: Yeah.
So I know your goal is to stay at the cutting edge of technology, the next innovation, this is a competitive industry. And when you're up against big players like NVIDIA, how do you do that? LISA SU: Well, the beauty of technology-- and I like to say this very much-- it is about building great products. And to really do that, we actually have to see the future. We need to decide, hey, where is the industry going over the next three to five years, and we need to place big bets on technology. And I think from that standpoint, it is one of those areas that is very rewarding if you make the right big bets.
And we've made some very good bets, I think, as we look at technology going forward. I'm super excited about what we're doing in AI. It's sort of a confluence of events. I mean, generative AI has come into fruition, and the fact is, everybody needs compute technology, and we're one of the very few companies in the world that can do that. And we've been really investing in this space for the last 10 years. So it is one of those places where you have to see across the horizon, and with that, we invest very heavily in R&D in the key technologies to enable the next generation of products.
ADI IGNATIUS: I love your observation, who knew this industry would be sexy, but you're having your moment, so that's great. So here's another audience question. This is Gajan Yogeswaran who's asking, how expensive versus accessible will AI technology be in the medium term? And the point is, given the large cost of materials required for building semiconductors, for employee headcount at the big producers, like AMD, do you see the cost of accessing this technology will limit the ability of certain people, certain companies to take advantage of what it could offer? LISA SU: Yeah, the great thing about technology, especially when you think about usage curves is, we're very cognizant of the fact that for technology to be most broadly adopted, you do actually need to get the cost to a very, very reasonable point.
So one of the things that we're working on today are things like if you think about there are all kinds of large language models that are used in AI. There are some who are the most advanced, the largest which require tens of millions, hundreds of millions of dollars, maybe even billions to train. But frankly, there are ways to really access more fine-tuned models that don't require that kind of investment.
Or if you think about how much it costs to ask a question to ChatGPT or one of your copilots these days, we call that an inference opportunity. We're absolutely looking at reducing the cost of that by factors over the next couple of years. So I don't believe that this is going to be an overall issue, where the cost is prohibitive.
I think it is an issue of you have to decide where your return on investment is and where are you going to see the largest productivity enhancements. And that is very much what we're driving, as we look at advancing the technology going forward. ADI IGNATIUS: So your industry seems very complex, and the supply chain seems very complex.
On top of that, you have the uncertainty of political and trade issues. As I said, it's a sensitive industry. China recently said, at least, it was prohibiting AMD and Intel chips from government computers.
How do you respond to that? Can you do anything to try to move the needle on policy issues like this? LISA SU: Well, I would start with the notion that, look, every country has to do what they believe is in the best interests of their national interests. That being said, the particular question that you have about China's policies around government procured processors, that actually wasn't new news. That was telegraphed, actually, late last year. And so it is something that, again, we look at the breadth of the market that we have. We are a global company. We work in all markets.
China is a large market for us. And so within that, as long as we can plan across the different markets, I don't see it as a significant factor in business. I think the more important conversation is, we're very much about driving deep partnerships across the globe, and that's with both large companies, as well as small companies, startups, and companies are very regionally focused and will continue to drive deep partnerships across the world. ADI IGNATIUS: As I mentioned at the start, you may well be the most prominent woman in the technology industry. How do you think the industry is doing now in terms of gender equity? LISA SU: Well, that's very kind of you to say that, Adi. I appreciate that.
Look, I consider myself extremely lucky to be where I am. This is my dream job, to be a part of an industry that is so important and essential to the world and be leading a company like AMD. In tech, look, there are not enough women. I mean, I think we can say that.
It's one of those areas where we're consistently trying to drive more gender diversity, as well as just overall diversity of thought. And the reason for that, frankly, is we want to build the best business, and we want to build the best products. And to do that, you do need diversity of experiences and thoughts. I'm a big believer in the best thing that we can do is give people opportunities.
I was very lucky in my career, and I got a chance to really experience many things early on in my career, which helped give me some great experiences. And so that's very much what I'm focused on doing is giving women more exposure to the industry overall and then opportunities to shine and demonstrate their capabilities going forward. ADI IGNATIUS: So there's also the question of, I guess, age diversity.
David Dawson a viewer asked, do you see any clear opportunities or gaps where new perspectives-- and I think by that he means new graduates, young workers-- will be beneficial and let's say, particularly, in AI? LISA SU: Yeah. Look, we are always looking for new talent. I mean, we've significantly grown as a company.
When I first started as CEO, we were about 8000 people. We're now about north of 25,000, so lots of growth over the last 10 years. And I think the key for that is a diversity of perspective is super important.
What I like to say, especially when we're looking for new graduates, we don't view hiring somebody out of school as job training. We're not looking for that exact software skill to plug into a software team. What we're looking for is people who are great thinkers, who are great problem solvers, who are here to build a career, and here to learn a lot of different things.
And along the way, we're going to need your hardware skills and your software skills and your problem solving skills. And so yes, I think diversity of thought is really important. We love new graduates out of school, and we hire across the world, new hires every year, and we'll continue to diversify our talent base going forward. ADI IGNATIUS: So you can't do an HBR interview without getting at least one classic HBR question. So here's my classic HBR question. In your decade as CEO, what's the most important lesson that you've learned in these 10 years? LISA SU: Yeah, so I think the most important lesson that I've learned is to really be very ambitious in the long-term goals that you set for a company.
I mean, if you think about where we were, we were a $4 billion company in 2015, and we're now north of $22 billion, $23 billion last year. I think setting very ambitious goals for the team, while having very clear milestones for how we show progress along the way. Certainly, in our business, it's about long-term thinking and charting a strategy for that, but everyone needs some near-term milestones as well.
ADI IGNATIUS: So there's an audience question that has gotten lots of upvotes. You can do with it whatever you want. This is from Bashar. I don't know where Bashar is from, but Bashar's question is, what are you reading now? LISA SU: Wow, that's a great question.
I read a lot of things online, actually, and believe it or not, I'm a pretty avid user of both Reddit and X, because they actually helped me get very good real-time information of what's going on in the world. ADI IGNATIUS: OK. And last question. So a couple of people have asked, on the topic of sustainability, is AI helping AMD achieve its goals for sustainability? And then more broadly, do you see AI playing a role in influencing sustainability or CSR efforts for you or for others? LISA SU: Yeah, let me turn it around the other way.
I mean, our technology is actually very focused on sustainability. So the idea of where technology is going, think about it as not just about high performance, but it's about what performance can you get at a certain power point. So we're all about, when you think about today's limitations, frankly, power will be a limitation as you go forward. And so we're constantly looking at how can we be more efficient with our products, which help the overall sustainability conversation. Now, as it relates to AI, I do absolutely believe that AI will help us in sustainability from the standpoint of, it will get us to answers more efficiently, and with that, you need less power for that. That being the case, there's also the reverse trend, which is we're using a lot more computing to help us modernize our businesses.
So a lot of focus on sustainability. What I would definitely say to this audience is that the newer the technology, frankly, the more sustainable it is, because you do have all of the benefits of newer technologies being much, much more power efficient. So you need much less power to get the job done. ADI IGNATIUS: So Lisa, I want to thank you for being at this event. I have long admired you and long admired AMD, so it's really nice to have this conversation.
So thank you for being here. LISA SU: Thank you so much for having me this morning.
2024-12-14 15:43
