On this episode of The Six Five On The Road, hosts Daniel Newman and Patrick Moorhead welcome Intel’s Greg Lavender, CTO, SVP and GM of Intel SATG for an inside preview of the upcoming Intel InnovatiON 2023 event happening this week in San Jose, California.
Their discussion covers:
- An inside look at the content Intel has lined up for InnovatiON around their next-gen services, client and edge cloud, with a big focus on AI
- What Intel CEO Pat Gelsinger meant when he coined the term “Siliconomy” at Intel VisiON last year
- Why software is not a shift for Intel, even though its most known for its hardware
- A preview of what to expect at Intel InnovatiON 2023, what trends will be in the spotlight and what the future holds for research and technology at Intel
Be sure to subscribe to The Six Five Webcast, so you never miss an episode.
You can watch the full video here:
Or Listen to the full audio here:
Disclaimer: The Six Five webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.
Patrick Moorhead: Hi, this is Pat Moorhead, and the Six Five is live on the road, talking right before Intel Innovations, their premier developer event. Dan, it is great to see you, my friend. I am excited to land on the ground. But we are going to talk, do a little bit of a preview of the show here.
Daniel Newman: Yeah, it’s another week on the road, going to be a busy week. But this is a great event, Pat. We’re bringing together Intel’s community of developers and we’re going to be having conversations with a number of the company’s executives and leaders about where the business is heading, and of course, that critical role that the developers play in driving the future.
Patrick Moorhead: Yeah, I’m super excited. It’s funny, I used to be Intel’s largest OEM partner a couple careers ago, and they had a conference called Intel IDF. And everybody came, all the developers, everybody in the ecosystem. The fun part about evolution and Innovation is we’re going to see so much more software. I mean, we’re going to see hardware, of course, but we’re going to see a ton of software and a ton of services and core technologies at the show. Now, we’re going to talk tech, we’re going to talk software. We are going to bring in, introduce Greg Lavender’s-
Daniel Newman: AI.
Patrick Moorhead: I know, and AI. Of course, it’s AI. We’re going to bring in Greg Lavender, Intel’s CTO and GM of the Software and Advanced Technology Group. Greg, welcome to the show.
Greg Lavender: Thank you, Dan. Thank you, Pat. Glad to be here and happy to talk about Innovation. I was a big attendee at Intel Developer Forum back in the day and this will be our third Innovation since I joined the company a little over two years ago.
Patrick Moorhead: That’s so exciting, and we have had so many great folks from Intel on the Six Five. Your fearless leader, Pat, a lot of his business leaders. And it is a special treat to get you on right on the cusp of this Intel Innovation Conference.
Greg Lavender: Great.
Daniel Newman: So let’s jump in, Greg. I spent some time looking at the content that’s coming our way and of course, we’re going to be hearing about next gen services in client, at the edge. But as I whispered into the microphone right before Pat introduced you, I said we’re going to see a lot of AI. Artificial intelligence seems to be a big focus. Tell us what’s going on there.
Greg Lavender: Yeah, I’ve actually got a long history in this. When I was first starting out my graduate career in the mid-1980s, I actually was thinking about going into AI. But networking was more exciting to me and I made a good choice, I think, 40 years ago roughly, jumping into the networking space. But networking is now enabling all this broad connectivity, high bandwidth computing, et cetera. And so AI has really emerged, I think, here. Obviously with generative AI, we all see it. It’s pretty amazing, the speed at which it’s basically getting adopted, the amount of compute, networking, bandwidth, storage, is really just changing the landscape.
I started working on AI the day I started at Intel in June of 2021, so this isn’t a new area for us. It’s taken us a while to get our hardware into the market and it’s now there, and we’re ready to go and we’ve got our software. Because as you know, and heard me say before, software is the soul of the machine. And that’s what’s going to make all this AI and generative AI and large language model training and inference the big thing that’s going to drive computing for the next decade.
Patrick Moorhead: Well, listen, I think your timing for AI is great. It wasn’t as exciting back in the eighties, and it was even less exciting back in the sixties when the fundamental algorithms were being looked at and invented. But one thing that I think the industry finally understands, and particularly in the semiconductor industry, is that it’s one thing to crank out a piece of silicon and hope and pray that developers show up. But software more and more is becoming so much more important about the overall value proposition, how to get different pieces of acceleration. And Pat brought you in a couple years ago to lead software and ATG, but maybe everybody doesn’t fully understand this massive shift in the marketplace. Can you talk a little bit about this, and maybe talk about what that means for Intel?
Greg Lavender: Yeah, I mean, I think technology obviously has permeated every aspect of our lives now. And there just continues to be this amazing, growing, insatiable demand for processing power, whether it’s at the edge, as you’ve talked about, the client, the automotive, all the way through the core infrastructure that runs through global internet, the clouds, and the computing that we all have in our hands every day. And so this new generation of technology is just driving the convergence of this ubiquitous compute that we have. All the connectivity globally with now 5G, soon 6G, bringing everything together, making information available, wanting intelligence around that information to help us with decision-making.
We also need to be concerned about security and privacy in that realm. And then of course, we’ve got to be concerned about the ethics of how data and information is being used as AI becomes more educated, let’s say, with all this data that we’ve acquired over the last decades. And so really, this whole exponential leap in computing technology has reached this new convergence point. I think it’s the fourth one in my career that’s happened. And we can all see the opportunity and the potential, but I think we still have a lot to learn, and society has a lot to adapt to before this technology becomes more pervasive.
And I think you’ve heard Pat say the word siliconomy. It’s this new global expansion where computing is so foundational to this big opportunity that’s facing all of us. And it’s going to be software that’s going to manage to have to control it because this is not computing in the small, this is computing in the mega scale world. And so software has to manage that in these large-scale distributed computing environments and large-scale data movement and processing that’s going to be happening. Not just to train models, but the scale at which inferencing is going to happen with the client and the edge is going to be tremendous.
Daniel Newman: So Greg, Pat jumped into software, and Pat and I also have this sort of running joke when we talk about how people used to say, “Software will eat the world,” and we say, “Silicon will eat the world.” Because try running your apps on air, you can’t run anything without it. And then you just set me up really nicely when you talked about Pat and his phrase siliconomy. But maybe not everybody out there fully appreciates it. I think in the last few years, what’s happened with the supply chain and other things has definitely brought more attention to it. But share a little bit more about what that means. What does siliconomy mean? What is Pat Gelsinger, your CEO, trying to say when he says that kind of catchy phrase?
Greg Lavender: Well, I think it is a new economic enabler that’s happening. I mean, today, depending on who you read and whether it’s The Economist or Wall Street Journal or other analysts like yourselves, estimated $570 billion semiconductor industry today that’s powering a global tech economy worth, some people estimate as much as 8 trillion and growing. So if you think about those kinds of markets and the kind of expansion that that’s creating in terms of capabilities, in almost every industry, this is going to be quite disruptive. Some people even predict that by the end of this decade, it’ll be a trillion dollar semiconductor industry, which is why we’re investing heavily in our expansion of our factories, our fabs, our packaging facilities, which is also extremely important. The supply chain issues that we’ve got to get, including rare earth metals, to bring all this technology together. I mean, that’s the core business and what Intel is operating on.
And I have the privilege of sitting at the top of that siliconomy of stack, delivering the software aspects, whether it’s the firmware, the BIOS, the memory controller software, the distributed computing, the networking protocols, all the things that are enabling that foundational infrastructure to create these large scale high-performance computing, scale out as well as scale up computing infrastructures that the industry is demanding now, particularly for the training stage of this transition that’s happening. I mean, it’s all converging, as Pat has talked about before, from the compute, the ubiquitous compute, the pervasive connectivity, and then all the way out to the sensing, because we have lots of sensing going on in the world.
And that whole ecosystem is converging. AI is being the forcing function, I think, to bring that together. And we’re going to see a massive capability that’s being unleashed now in the world. And I think the other thing we have to worry about is the sustainability and the power consumption that this is going to create. And do the benefits match the expense of what we’re going to have to do to bring this kind of technology to all humanity?
Patrick Moorhead: So Greg, by the way, I love the term siliconomy. I’ve spent a lot of time in and around Pat at version one when he was at Intel. And now since he took the helm here, I might actually use that term. And I think I’ll give him credit, maybe I will, I don’t know. Hey, what is missing from the AI piece? Part of the siliconomy and bringing it together, you talked about the far edge, you inferred hyperscale data centers and everything in between. What is missing right now from the industry to be able to better enable AI in this siliconomy?
Greg Lavender: Well, I think we all get excited, again, by speeds and feeds, and that’s driven the whole industry really for the last 30, 40 years. And in computing, you have personal computing, you have infrastructure computing. The cloud brought this new era of cloud computing. And really, none of the AI things we’re doing today could have happened had not all those prior innovations of scaling out the internet, scaling out compute infrastructure, going multi-core, multi-socket, GPUs, all these things developing. But again, it all comes back, to me, as a software guy, what’s the software and how does the software enable it?
Most importantly, how do we enable all these developers? Developers have moved from, call it the bit coders that have always been the historical computer scientists at the core foundational level. There’s lots of people that are programming in these ecosystems, open source ecosystem frameworks like PyTorch, TensorFlow, there’s a new one, from JAX, from Google.
So there’s all this software ecosystem that’s exploding as well, to take advantage of all that fundamental physics, science and technology innovation, which Intel is contributing to with others. And so I think really the next step is how do we bring that software together? How do we bring ethics into the picture? Because I think there’s lots of opportunities for abuse and we’ve already seen some of that manifest itself.
And more importantly, I think, the security and the privacy — and I put those two things as sort of two sides of the same coin — really have to go hand in hand, and you’ll hear me talk at Innovation about security for AI, and AI for security. And I think that’s an important concept, where AI is at the apex of making sure that the data and the privacy rules are maintained, but at the same time AI can make us better at managing, using our data and getting more value out of it. I think this is, again, part of the siliconomy, but at the top of the stack, not the bottom of the stack.
And at Intel, I got 19,000 software engineers working across that software stack, and we have over 460 software tools and resources, much of it open source, in our Intel developer catalog at developer.intel.com, to let developers grab that stuff and go create solutions. But you don’t have to be a chiphead to do it. You can operate within these open source AI ecosystems and get a lot of progress and make a lot of value really quickly. And the level of productivity that developers are going to be able to achieve is amazing.
Daniel Newman: It’s interesting that you talked about that inflection or the apex. It’s also a really important moment for AI, Greg, as it relates to where does responsibility, regulation, compliance, security and privacy intersect in such a way where we get it right, where we take care of the users, we take care of the enterprises, but we take care of also these developers that are building and make sure that they’re building things that are safe and secure? I’m kind of curious, because Intel’s been very prominent in building relationships with the government, one of the most, probably what I believe to be one of the biggest beneficiaries of the CHIPS Act, which will be great for the US and our allies around the world.
But I’m curious, as you kind of brought that up, since you brought it to the attention, how can Intel help to drive more responsibility in policy and AI? Is this something that at Innovation when you’re talking to your developer community, when you’re talking to this audience, that we can really make sure that the industry and the developers themselves are going to have to lead the way? Because I just don’t see policy being able to be made quickly enough, coming from our regulators, but really it’s going to come from industry and leaders like Intel.
Greg Lavender: Well, I think we as an industry have to come together on this. And Intel certainly has been a leading public voice in support of the CHIPS Act. And I mean, just to be honest, we haven’t yet received the funding from the CHIPS Act. We’re using our own money at this moment to build out our capacity and capabilities. But we are hopeful that that’s all going to come through as it works its way through the allocation system.
But again, let’s just give credit to the CHIPS Act. I think it’s the first time in 70 years we’ve had industrial policy focused on this particular industry and this capability. So let’s make no mistake about it, the US government is committed to basically enabling the siliconomy. But at the same time, I think we have to make sure that the regulation and the technology co-develop, because I think there is a lot of goodness that’s going to come from this. We already see examples of it, there’s lots of productivity.
There’s also a lot of disruption and there’s social disruption, there’s going to be business disruption. And in any major technology inflection point, whether it was the emergence of the Mosaic browser back in the day that became the Netscape browser, the whole ecommerce factor, the onlineness of commerce through the cloud era. So again, I think AI is going to augment a lot of that, but again, it goes back to the data. Who owns the data? What are the intellectual property rights around the data? Who are the people that are authorized to use that data? And if you make derivative products off of it, is there some custody or chain of ownership that gets respected and you’re not stealing somebody’s IP?
So I think it’s going to test the system and it’s going to drive a lot of innovation. And the way we think about data, the way we think about privacy, the way we think about IP protection. Because this is going to be very, very disruptive, and I think we’re just at the nascent, beginning phases of this. Like I said, it’s going to be a 10-year inflection point in the industry and all these things are going to have to get worked out. And I’m happy to say, I actually know several people who’ve gotten PhDs in ethical AI over the last two or three years from reputable universities. And I think we need leadership in this idea of ethics for AI. Just like we need security for AI and AI for security, we need AI for ethics as well.
Patrick Moorhead: Yeah, historically, any big movements of the technology rock has driven a lot of these conversations. You talked about the web, e-commerce, mobile, local, social, cloud. I would say AI version one, and here we are at AI version two. We’ve managed to move across, but the fundamental changes is, I mean, we’re still feeling the changes of mobile, local, social, the amount of time we spend on our phones, the distractions, the kids, and all of that. So even though that started 10 years ago, we’re still working our way through this. So Greg, the auspice of this show was the preview.
I know our viewers and listeners appreciate this conversation coming in, particularly related to AI, but let’s augur in on Innovation and some of the not necessarily specific announcements, because I know you need to keep that under wraps. But without giving too much away, can you tell folks what they can expect at Innovation? Maybe if they’re watching this and they’re not certain on whether they should attend or not, either in person or online, maybe why they should.
Greg Lavender: Yeah. This is our developer conference. We have a vision conference that’s more for business partners and customers. This is really going to focus on the tech, the technology, both the new processor technology, the new accelerator technology. We’ve got our products ready to go, they’re in the market. I’m going to be talking a lot about Intel Developer Cloud, which is a place where developers can go and try before you buy and experiment, camp out for a while and learn about all of our newest technologies.
Just for example, we’ll have our Emerald Rapids, which are our successor to Sapphire Rapids, available in Intel Developer Cloud after Innovation, so they can get the latest silicon from us. It’s not even shipping yet in the cloud or on the server. Obviously, our Max GPUs will be available for people to go experiment with. Our Gaudi 2 processors, we’ll be announcing the successor processors to that.
And then also what we’re doing with AI at the edge and how we have AI hardware accelerators in all the software stacks that we’re doing throughout the ecosystem, whether it’s Windows, Linux, the open ecosystems that I talked about, like PyTorch, et cetera. We’re number three contributors to PyTorch. We’ve earned a governing seat on the PyTorch Linux Foundation board. So we’re contributing to this whole thing. There’s massive amounts of disruption and I’m very excited that we’re going to work hard to drive these open ecosystems, disrupt the incumbents, you know who has a proprietary offering for everybody.
And we think that the developers will gravitate toward open solutions. We want to democratize AI, make AI available for everyone, and again, advance Moore’s law and give people the silicon architectures, the networking bandwidth, the capabilities, and the software that’s going to bring all this together to let people do amazing things.
Daniel Newman: Well, Greg, I think that’s a really great place to leave this. After all, it is a preview, and if we spend a little more time with you, we might end up having to reveal everything that’s coming next week. This is a lot of fun. It’s really good-
Greg Lavender: I’m just giving you a tickle. Come to Innovation, see the rest of it.
Daniel Newman: Don’t you worry, I’m absolutely tickled. Pat and I are silicon geeks. We do love the fact that all this stuff comes together. From chips to SaaS, as we always like to say, there’s a lot of interdependence. But this is going to be a great event and I am very excited, and I’m sure, Pat, I speak for you, to hear about how open democratization of AI is going to become possible. We need more competition, we need growth, we need scale. Intel is a company that we genuinely believe is well-positioned to participate and make a big impact and we look forward to hearing more. Thanks so much for joining us today.
Greg Lavender: Yeah. And just my final comment is I will be talking about some future technology coming out of Intel Labs, which reports to me. Last year, I unveiled a few things about new technology that’s coming. So we’ll do an unveiling of that as well at the end of my keynote talk. As Pat likes to say, the geek is back. If nothing else, tune in to see Pat’s T-shirt during his keynote. And I will have a cool T-shirt, but it’s not as cool as Pat’s.
Patrick Moorhead: Now, are you going to have to do pushups on stage with Pat, or not?
Greg Lavender: No, I think I’ll let Pat do that. He’s got twice the energy I do, even though we’re the same age. And he’s full of silicon, I’m full of software.
Patrick Moorhead: I love it.
Daniel Newman: Well, that’s perfect, we’ll leave it there. Greg Lavender, thank you so much for joining us here on the Six Five. It’s the preview for Innovation. We will be there next week, Patrick and I, the Six Five, and you will be hearing a lot more from us. We’ll have interviews with lots of executives, and CEO Pat Gelsinger will be joining us here on the show. But for this episode, for the preview, we hope you are excited. Follow us for all of the Innovation coverage and all of the Six Five’s coverage. But it’s time for us to say goodbye. See you all soon.