The Vergecast - How 5G and edge computing fit into the future of Intel’s traditionally chip-focused lineup

Episode Date: February 4, 2020

Intel is one of the biggest names in the tech world, with chips that quite literally are the brains behind most of the computers and servers that we use every day. But the world of computers is expanding and Intel is changing, too, with a focus on both edge computing that puts processing resources in the cloud and the power that’s available directly on the physical device.  And at the head of that is Dr. Venkata (Murthy) Renduchintala, the chief engineering officer and group president of the Technology, Systems Architecture and Client Group at Intel. Renduchintala joined Intel in 2016, having previously headed up competitor Qualcomm’s chip business.  Renduchintala is the person in charge of almost all of Intel’s hardware, from design to engineering to manufacturing. He joined Verge editor-in-chief Nilay Patel and news editor Chaim Gartenberg for an interview episode of The Vergecast this week to discuss the present and future of Intel, including the company’s place in the development of 5G, the changing landscape of personal- and cloud-based computing, and what the next-generation of processors could look like.  Learn more about your ad choices. Visit podcastchoices.com/adchoices

Support for the show comes from Retool. Too many companies run critical operations on duct taped spreadsheets, Slack workflows, and whatever else they could cobble together. Not because they want to, but because building internal tools means weeks of waiting on someone else's backlog. That's where Retool comes in. Build custom internal tools just by describing what you need. Prompts something like,

Build me a revenue dashboard on our Salesforce data. And Retool actually builds it on your company's data, in your cloud with enterprise security built in. Go to retool.com slash Verchcast. We all need to retool how we build software. What's up, y'all. I'm Skylar Diggins, seven-time WMBA All-Star, Olympic gold medalist, and mom. And I'm Cassidy Hubbard, host and reporter for nearly 20 years,

covering the biggest names and stories in sports and mom. And this is Am Mom, a community for athletes, game changers, and moms of all kinds. dropping May 14th. Tap in with us. Hey everybody. It's now from the Vergecast. On this week's interview show,

our own Heim Gartenberg joined me to talk to Intel chief engineering officer Dr. Murthy Rendu Chintala. This one gets really deep. We talked about the state of the Intel roadmap, how he's managed to reset the company

after jumping over to Intel from Qualcomm. This one gets pretty deep. We talked about the state of the Intel roadmap, how he's managed to reset the sort of engineering side of the company after jumping over from Qualcomm, how he manages that. that engineering team. And of course, my favorite question, which I'm kind of asking everybody,

when he finds time to actually get work done. Check it out. Dr. Murthy, Rendu Chantala, chief engineering officer, Intel and group president. Welcome to the Redcast. Thank you. Good to be here. So a lot going on in Intel world. Haim Gartanberg is here with me. Hello. But let's start just at the start. When did you get to Intel? You were at Qualcomm before. You've been there for a few years now. Just give me that story. So I started in Intel at the beginning of 2016, after having spent, 12 years at Qualcomm. So moved up to the Bay Area in early 2016 and been there for coming up to

four years now. And when you first got to Intel, you made some waves, right? There's been a big shift. There's like a new CEO. You were the new chief engineering officer. The company needed to change. Tell me about sort of making that change happen. Well, I mean, clearly I wasn't hired by Intel to maintain the status quo. But at the same time, Brian Krasanich, the CEO at the time, had very clear vision of how he wanted to transform the company and move it on from a PC-centric to a data-centric company. And I think he was very much looking for leadership that could help him promote and accelerate that transition. And given my experience in the silicon space outside of Intel, I thought, I think Brian thought that my addition to the leadership team would

be a good addition to the overall chemistry of the company and start bringing some outside thinking in. So very much towards moving Intel towards embracing and understanding how the outside world did things in terms of trying to mix the two ecosystems together in a way that meant a net positive progress for the company. Let me push you. What does it mean to be a data-centric? For a shipmaker, what does it mean to be data-centric? It's to be immersed in everything to do with the storage, the processing, and the distribution

of data. It's to fundamentally understand that data is exploding around us and the ability to provide the processing capabilities to deal with that data explosion, whether it be to process to distribute it or to store it is going to be the key areas of growth for the semiconductor industry in general over the next few years. These are my favorite kinds of questions. I understand what you're saying, but in practice, you show up, it's your first day, you brought into this task, everyone understands it. Do you just, like, roll into the conference room with the KB Lake team, and you're like, understand the data exists and, like, roll out again? Like, how does that take place?

Well, first of all, you spend a lot of time listening, learning, and understanding. I mean, I think what's very, very exciting about the opportunity at Intel is it's a huge opportunity to learn. It's the first time I've ever worked for an IDM. It's the first time I've ever been in this industry space. So, you know, when you walk into a company like that, you spend your first six months listening, learning and understanding and getting to grips with what the company is all about, understanding the new culture, understanding how it works. And then trying to figure out how the experience set and the skill set you have could be applied complementary. to the assets you have around you. One thing that just struck me when you're talking about data

and sort of the shift towards data from the individual processors is that it sort of mirrors the shift that we've been seeing a lot in the general tech industry, which is sort of a more cloud-based approach. Your games are no longer, you know, start on your computer, your music. Everything is on a server. We're seeing, you know, game streaming services in particular where the processing power is offset elsewhere.

And I wanted to kind of get your take on that as it sort of fits with that approach. Yeah, I actually think that the end state is going to be a harmonious interplay between cloud and local processing. I think the demands for instantaneousness and low latency in a deeply immersive experience means that any experience going forward, whether it be gaming, VR or applications of that nature are going to be a well-architected mixture of on-device processing. and storage as well as cloud-based capability. I mean, I think if you look at your smartphone, for example, it's the ultimate cloud-based device. At the same time, there's more semiconductor processing in that smartphone

that I see across the desk from me than was maybe in a laptop two years ago. So I think it's a fallacy to assume that as the cloud grows and the cloud environment hosts more and more of what we've hitherto done on the device, the device becomes dumber. I think in many respects, the areas that you specialize in the client devices just tend to be more different. So display, graphics, storage tend to be much more amplified at the point of the experience consumption than maybe otherwise would be imagined. I'm glad you brought up smartphones because it's another topic that I want to ask you about in particular. It's an area that Intel hasn't so much been a part of.

The company just sold it. It's 5G mode in business to Apple. And I'm curious where in this data-centric approach, Intel kind of sees itself fitting into this mobile space, just especially as that's becoming an increasingly central part of the computing life. Sure. Well, again, smartphones really were the, I think, the iconic devices that created the mobile internet age that were all experiences now. And we have 8 billion connected devices that really are key to the way we function today. But if you wind forward in the next five to ten years, you're talking about a world that's going to have a hundred billion connected devices way beyond the smartphone. Devices talking to devices without even human interaction in between.

And therefore, it creates a completely different landscape where the amount of data that's going to be generated is going to just exponentially grow. And the infrastructure required to deal with that is going to have to completely transform. So I think what you're really seeing now is the movement of people thinking about. the mobile internet towards the internet of everything, and that's going to have seismic transformations in everything around us, not least of which is going to be the networks that service all of that. The edge and the compute at the edge is going to become more and more important because the amount of data that will be generated just can't be consumed by traditional backhaul

architectures. So I think it's going to be a transformative period for us. In fact, I think this transition to 5G is going to be more profound on the networking side than on the device side. And I think we'll see a transition that's a significant in the network as the transition was from analog to digital. But I don't think of Intel as a traditionally networking company. I think if Qualcomm is a traditionally networking company, is that part of the idea here that you're going to move to more directly to compete with Qualcomm? I mean, when I joined Intel, Intel was considered to be a company that made PCs

and servers and basically had 90 plus percent share in a $60 billion tab. What we really tried to do is reposition Intel's position in the overall landscape to be a company that has about a 25% share in a $300 billion, of which two-thirds is data-centric. And it's all about the transformation of the network and the cloudification of the network, where the technologies that were in the data center get distributed and integrated into the fabric of the networks we know of today. And where you go from a network delivered by bespoke pieces of silicon, each working together, with single function piece of silicon towards a paradigm of what we call network function virtualization, the rise of the software-defined network, where workloads can be virtualized and run on a general

purpose compute engine, such as a Zeon processor for that matter, which is why, as you think of the future, you should think about the assets that Intel have becoming more and more relevant across the continuum of the network landscape. That's the vision of kind of network-based compute that I've heard from Verizon, for example, I've heard it from AT&T. I haven't heard it so much from other 5G vendors, right? The carriers are kind of doing this in different ways. But you're saying that's an overall vision?

Is that the vision you're pushing Intel towards? Is that where you go out and market and say, hey, Sprint, build our network this way because our chips will be great for it? Or is this where the entire industry is going? I think this is where the entire industry will ultimately go. And Intel is catalyzing and accelerating that trend. think what you're hearing from network operators is a different pace at which they're approaching that transition based on the legacy that they have. For example, if you're a network operator

that's putting something in from the ground up today as a day one exercise, I think the approach you'd take would be very different and very radical. For example, if you're an entrenched operator that has a huge amount of sunk capital in 3G and 4G legacy. So I think what you're really seeing is sometimes it's much easier to move when you have less baggage of the past to carry with you. It's a bit like, for example, if you look for the rise of an operator like geo in India, where it's just basically completely swept the country,

it's really because it's put a ground-up network in that's been digital and no circuit switch capability from day one, as opposed to the legacy carriers that had to deal with circuit switch 2G and 3G as well. So I think at the end of the day, the economics and the ability to deliver network evolution swiftly and rapidly is going to mean that the movement towards software-defined networking is going to be inevitable. The pace of which various operators will ultimately execute is going to be based on a function

of how much they've already sunk in legacy networks, be it 3G or LTE, and how much are being started from a greenfield approach. So I want to make sure we talk to you about the chips and the laptops and how many cores they have, which is of paramount interest to me. But I want to just stick with networks for a minute. I ask everybody this question. I hope you can answer it. Is 5G a race?

No, I think 5G is a... No one says yes, by the way, so you're fine. I just don't understand it. What happens when we lose? So 5G in my mind is an inevitable transition that will affect the globe and the technology around the globe at various paces. I think what you'll find is that smartphones

will basically be the thing that catches the, the eye immediately because they're they're in people's hands they're personal but ultimately what will be the truly transforming experience is what happens to the network behind it and how that transforms to deliver a variety of services that you hitherto couldn't have even imagined of where you're going towards situations where the network latency can support real autonomous services as opposed to you dial a button and you wait for three seconds and then you get a circuit switch connection it's going to completely transform the experience and it's not a race because at the end of the day

The victories come in so many different shapes, forms, and sizes. There isn't one game we're playing. It's a completely transformative experience of our digital lives. And that's going to happen at different paces at elements to our lives that have different importance to different people. But this doesn't sound like you're putting chips in consumer products. It sounds like you're selling a lot of gear to operators. I think I think of it differently. I think that what you're really seeing is a greater and greater interconnectivity

and coupling between what's happening at the network and what's happening in the devices. And the way in which the services are delivered and the way in which applications are started and where they're run is going to become a much more harmonious interplay between what's running the network and what's run on the device. So I think you have to start thinking about a much more continuous type of mindset in terms of the whole kind of like client server model we're used to. This is going to be a scenario where the client, the edge and the network are going to work in harmony and in a synchronized manner to deliver a really immersive user experience.

So I think it's going to be a much more unified concept. So you're saying that Intel wants to be not just the name on the box that you're holding in your hand, but the name on the box that you're not seeing that that box is talking to. I think, you know, the vision we have of Intel is being a solutions provider that powers your digital world. That's kind of like the way we're really like trying to think of ourselves. Yeah, how do we affect the evolution of the core network, the data center, the end, edge, the client environment, the inauguration of more and more autonomous agents into an environment where everything speaks to everything else and creates a harmonious, synchronized

continuum. Do you think of Intel is an actual operator, equipment vendor on the scale of a Huawei or a Siemens? Did Siemens own Erickson now? They're all merging and unmerging. But that class of companies that the operators rely on, do you think that you need to build that class of equipment? I think we need to think about how we expand our customer base to include those kind of customers. So, for example, today, Erickson and Nokia are going to be our, and will continue to be very, very important customers for Intel, beyond our traditional PC customers. So as you expand from a company that's focused on a $60 billion tam to a $300 billion tam, then the industries and the customers that we're going to service is going to expand exponentially. and therefore we see ourselves constantly engaging with customers that are new and bringing in new knowledge to us

and we have to serve them in in different ways than we've served our traditional customer basis. That's how we grow and how we expand. All right. That's the future. I get it. You're going to do cool network stuff. Let's talk about the extremely real present, which is this laptop I have in front of me. I'm going to just let Heim drive because he has a long list of questions about your current set of chips.

Okay. So Intel's a company is on a generational edge right now. We're finally seeing the first wave of the 10 nanometer chips, which have been anticipated for years. There's been a lot of drama, a lot of delays and rollout for these chips, and we're finally starting to see them. Just give me a quick snapshot of where the company's at right now in terms of that hardware and where it's going in the immediate future. Well, Haim, you're quite right. We're now seeing the first fruits of our labor. on 10 nanometers with the isolate trips that came out, the 10th generation stuff. And we've been very pleased with how that's been received by the market. We're seeing not only the benefits of 10 nanometer technology that brings different power versus performance points for the products, but we've also got a whole barrage of new IP that had been node lock to 10 that now sees the light of day.

So we've seen the Gen 11 graphics come out where we've seen a significant improvements in the nature of the experience. We've also brought out technologies like integrated Type C and Thunderbolt technology into the base architecture. And we've introduced for the first time AI instruction sense into the core instruction set of our computers to be able to accelerate AI tape capabilities. So I think what you're seeing with Ice Lake is the first in a very exciting lineup that's going to be rolling out through 2020. and we're putting in more and more differentiated IP together with evolutions of our 10 nanometer technology

as we go through 2020. So you ain't seen nothing yet. I think the products we'll have in the course of 2020 are going to be very exciting, where we bring bigger CPUs, better graphics, better power versus performance and battery consumption, really continuing to elevate the user experience with every platform we launch on an annual.

cadence. So I have to say two words, which is Moore's Law, followed by a question mark. It is one of the questions that has come up, especially with the much longer development time for 10 nanometer. I know Bob Swain has already said that Intel's hoping to return to the two and a half year cadence for 7 nanometer and 5 nanometer beyond. Is that realistic? And how important is that that you fit to those timescales? So first of all, I think Moore's Law is a line. and well. It's not becoming any easier with time because, you know, developing deep sub-micron geometry at that two-year clip takes more and more innovation. But Bob's absolutely right to reiterate that we want to get back to a two to two and a half year cadence. And, you know, the whole 10 nanometer

experience for us is more of a big hiccup than in any way a reset on what we're really thinking. I would say that 10 nanometers was more a over-conflience of risk that made our ability to deliver on a two-year cadence. We missed a step, effectively. But we're really, really focused on getting back onto that two-year, two-and-a-half-year cadence. And the lessons we've learned on 10 nanometers have really figured into how we create that equation going forward to deliver that predictability. So with seven and with five, we're looking towards basically finding a much more appropriate balance point between, scaling, cost, power, performance, and schedule predictability. And we're optimizing our challenges in each of those areas to really up the probability that we can nail the two to two and a half year cadence.

That's going to be a very big part of our roadmap construction going forward. What we're also doing is between those two and a half years delivering meaningful intranode improvements as well. So that what you'll see is less of a kind of like a step function improvement in performance and more of a continuum where you'll have, essentially the guide rails of kind of node geometry changes every two years. But we want to basically do incremental performance improvements during those years. That was a core lesson we learned when we were in the middle of the trials and tribulations with 10. We were able to generate a lot more performance out of 14. And what we did on 14 now becomes the new normal going forward.

One of those things that Intel did with 14 is something that's very interesting to me. Over the generations of the 14 lineup, we saw an increasing number of multi-quart. processors come out from Intel and not just Intel across the industries where we once had, you know, four cores was the upper limit. We're now seeing hexacore and eight core laptops as standard consumer devices. On the desktop side, we're seeing, you know, 28 core. It's a core arm race across the whole industry. I wanted to know if you could talk a little bit about kind of the decision to branch off into that

and while the course. Well, you know, to some degree it's a little bit of deja vu to me. because I went through this in the smartphone industry, some a few years earlier than the PC. So we went from kind of like single core CPU architectures to four core and then to eight cores. And the logic really applies that as you get towards the diversity of workloads we're dealing with today, in the digital world we live with, there's a greater capability to be able to split workloads up

and run them more efficiently and in parallel multiple processes. You know, there's one school of thought, which was essentially you could basically run a single CPU with a higher and higher clock rate. And you get to a point of thermal limiting at that point, then the next scenario is you split up the workloads and run them over a number of cores at a lower frequency, which essentially gives you more throughput at a more efficient power level.

And it's the same logic coming through that what we're really seeing is the way the PC applications of today and the future of being architected, whether they be cloud-based or native, being architected in a way where they take advantage of this concept of multi-threadedness, where they can or being able to be split up across multiple cores, and therefore it catalyzes multi-core designs. So I think it's been the advent of multi-core designs

with the advent of applications that can take advantage of those multi-cores that are really fuel this going forward. And I think what's really happening now is that you're seeing more and more workloads that just need more and more compute, and to deliver that in a power efficient and a heat efficient way, it makes sense to split the workloads up across a number of cores. So, you know, what you're seeing in the laptop and what you're seeing in the data center was pre-informed

by what we saw in the smartphone many, many years ago, where essentially you go from single-core high-frequency to now multi-cores going forward. And the future is going to be not just multi-cores, but what we call multiple XPUs, where it's not just CPUs, but CPU's working in continuum with GPUs and specific function ASICs for things like AI. Support for this show comes from Shopify. Starting something new isn't just hard. It can be really scary, too.

So much work goes into this thing that you're not entirely sure will even work. But here's a better thought. What if it did all work? What if your instincts were actually right all along? Shopify wants to help you get there. They're the commerce platform behind millions. of businesses worldwide and nearly 10% of all e-commerce in the U.S. From established brands like Allbirds and Heinz to companies just getting started.

Their design tools make it simple to create the exact online presence you're envisioning, with hundreds of ready-to-use templates available. And with built-in marketing tools, you can launch full email and social campaigns in just a few clicks. So you can connect with customers wherever they are. It's time to turn those what-ifs into with Shopify. today. You can sign up for your $1 per month trial today at Shopify.com slash vergecast. You can go to Shopify.com slash vergecast. That's Shopify.com slash vergecast. Support for the show comes from Upwork. The days of doing it all, all by yourself, are over. There's no romance and burning out while you're

trying to scale. Instead, you can check out Upwork. Upwork helps grow your business by giving you fast access, to specialize talent across more than 125 categories so you can fill skill gaps, launch projects faster, and scale without committing to full-time headcount. And finding the right talent is easy. You can browse profiles, review past work,

and get help scoping the role so you can get started quickly. Seriously, you could connect with the right freelancer in just a few hours, especially when you sign up with Business Plus. Their AI-powered shortlisting pairs you with the top 1% of talent in under six hours. No endless searcher required.

You can visit upwork.com right now to post your job for free. That's upwork.com to connect with top talent ready to help your business grow. That's upw-w-rk.com. Upwork.com. Another thing I wanted to just touch on is competition. Intel is up against more competition than it's had in possibly the last decade. AMD has this new Risen chips, Qualcomm has desktop chips for the first time. There are more people who are trying to, you know, sort of edge in onto Intel's traditional space at the same time that

Intel is also looking to expand into other areas. So I kind of want to talk about sort of the strategy of the reliance on the client computing group versus other avenues that Intel is working on. You mentioned your, the data-centric approach and how that fits together. Well, first of all, the PC business remains incredibly important to Intel. It's a driver of a lot of value and gives the organization tremendous scale. And we're going to compete as effectively as we possibly can against all comers in that market. Now, first of all, I think competition is extremely healthy because it generally brings out the best of every company when it's faced with stuff competition.

And I can clearly see that Intel has reacted aggressively to the onset of competition. If you look at, for example, the financials of our client business and the operating profit improvement year over year, it's been a stellar performer, primarily driven by a stimulus towards being absolutely agile to respond to the competitive vectors that we see, whether it be the armed community or other X86 areas. What I will say, however, is that that is a business that, while important, is also managed concurrently with our desire to basically move into the data-centric era. You know, in 2013, the PC business amounted to something like 70% of Intel's overall revenue. In 2019, it was more like 50-50 between the PC business and the data-centric businesses, while the business grew some 30% in terms of overall revenue between 2015 and now. So what we're really seeing is a continued focus on the client business and making sure that we maintain competitive ground and investing in order to do.

do so. But we're also making sure that we play into that $300 billion tam that Bob and others have talked about, where we need to play across multiple different fields and in markets where we're not the incumbent, where we're coming into areas where we're competing with strongly entrenched players and trying to pull together disruptive technology in order to be able to gain share there. So for us, it's really a case of holding secure the PC franchise and supporting that with the right amount of R&D to deliver innovation in the area, but making sure that we also transform the company into this data-centric, focused player that we want to be that's playing in that $300 billion town.

But that PC market's changing a lot, right? I mean, I know Apple will try to convince you that the iPad is the future of computers, if you so much as blink at them. Maybe it is. Maybe it isn't. But that is an arm processor. Microsoft just put out a Windows tablet, a Surface tablet with an arm processor. There's a lot of activity in that zone.

towards, hey, we actually do need instant on. We do need integrated connectivity. We do need all the stuff that they tell me, the vendors tell me, comes with arm. By the way, let me be clear. I mean, we absolutely need to innovate in the

PC space to continue to maintain the success of our business. And that's absolutely what we're committed to do. So you're absolutely right. I think, you know, this whole concept of basically making a PC capable of being able to give you an uncompromising full day of productivity and not having to worry about taking a

charger with you, it's fundamental. The ability to drive more and more immersive graphics to be able to have that always on instantly connected capability is taken for granted. But how do we do that in a manner where that continues to drive the thin and light form factors that we appreciate? Wait, how do you do that? Well, again, it comes down to, it comes down to mastering a multiplicity of different technologies. I mean, one of the things that we've learned is that even for the PC, the innovation model that used to be described by Intel as TikTok, which was drive process and drive CPU architecture,

in and of itself, while important, isn't sufficient to drive a leading roadmap going forward. It needs to be augmented with much more mastery in a broader bandwidth of technologies. For example, things like interconnect, things like memory, things like the way we optimize and run software, graphics, these all become just as important. The way we architect our products become just as important. So we have to go from TikTok, which was processed in CPU microarchitecture, towards thinking about this being a game of multiple technology dimensions. Do you feel challenged by the fact that everyone still thinks of Intel as the TikTok company?

I mean, aside from actually TikTok, which I don't think anybody confuses Intel with. But, I mean, that has been the story of Intel since I started covering tech. It's still the frame that your roadmap fits into. Is that the big change you're trying to make? Like, stop thinking of it this way? Well, I would say it's not stop thinking about it. Think about Intel as being much more. I mean, Intel has a tremendous heritage that I for one respect tremendously.

So it's not about moving away or redefining. It's about evolving and putting Intel on a number. a broader stage. But just to the example you just gave about okay we the shift to 10 was hard. We learned an awful lot about what we could get out of 14. That actually, to me that flies in the

face of process right? That means we're not optimizing what you have before you move on to the next, which is like the story of the tech industry in many ways. But is there a part of you that just wants to say I'm the new guy ish now but we're going to blow it all up and move

in a totally different direction or is it just a refinement of the existing strategy. Well, I think, I think it's, I think it's basically making a good model even better. I mean, part the reason why I'd like to think Brian brought in people like myself, and I myself were brought in people from the outside, is to bring some of the outside thinking and mel that with what's so good and great about Intel. And I think what, you know, the example you've just talked about, you know, the ability to get incremental improvement out of no transitions was basically taken for granted in the fabulous ecosystem. So myself and others coming in and basically starting to think about that being a modus operandi for Intel was

just natural to us. It maybe wasn't as natural to the Intel folks of the past. But when we sat down and discussed how we actually deal with some of the lessons we learned from 10, it was a natural outcome of that. And starting to think about delivering leadership across multiple technical ingredients just comes from having more vibrant thinking in the company and talking to our customers and thinking about where their roadmaps are going from. So for us, and also reacting to competition and seeing how they're able to gain a voice in the market that we ultimately then wanted to respond to. So I think it's all about making sure that we live by Andy Grove's parable about not being too complacent to lose sight. It was a always be paranoid. That was a very polite

spin on always be paranoid. So is there a world where, you know, in two years, Intel releases a seven nanometer thing, but we also see, you know, 13th generation Intel chips that will still heavily rely on this improved, you know, 10 nanometer scale and kind of have those as a concurrent thing. Like right now I can buy a brand new laptop. I can buy a 10th generation chips. Some of them are 14 nanometers. Some of them are 10 nanometer.

You're mixing the architectures to kind of, you know, maximize on each. Is that a thing that's going to continue to happen going forward? Or is it going to be more of a reversion back to, you know, out? with the old and with the new that we've seen in prior years. I think what you're going to see going forward is a longer interplay between nodes. And it might not even be separate, you know, concurrent 10 nanometer devices, shipping with seven nanometer devices. It might be, for example, seven nanometer and 10 nanometer living within the same SFC construction.

You know, I think one of the things you need to kind of like start thinking about is not being limited by the monolithic paradigm that we live by today. what's going to happen going forward is I think you're going to see SSC construction that uses advanced packaging capability to allow you to have a mixture of process technologies in the same package so that you get benefits of logic scaling where it really matters. And that where it doesn't have much of an influence, you don't have to do all the R&D work that slows you down and delivering all of that new cool processing IP. So I could clearly see a scenario where you could, for example, have 7 nanometer or 4. 5 nanometer or 3 nanometer compute and GP GPU tiles working with I.O and mixed signal technology that might be in an N-minus-1 node and running those in a heterogeneous SOC architecture all delivered within one package.

In fact, we've already sampled the first groundbreaking architecture of that guy's already in the Lakefield product we delivered, where it basically has our I.O. a number of other mixed signal circuitry in a technology we call 1222, with the compute and graphics being in 10 nanometers using a very, very novel, highly innovative 3D packaging construct to be able to deliver that all in one SOC. I think that's going to be the paradigm going forward, that monolithic becomes still an important part, but less of a panacea in the way we'll put our SOC roadmaps together going forward.

And I think you should think about nodes living much more concurrently than the original paradigm of all in one, none, and then all out of it with the next node transition. Let me ask you, it's either a dumb question or it's not. I don't know which one it is. But let me ask you, you talked about packaging, 3D packaging. And just before that, you talked about, hey, I worked at a fabless company before and I came in and now I can just take that, I want Intel to take it for granted. But Intel owns the fab. you work on that part of it more directly than any other chip vendor that I can think of, besides from like TSM, right?

What is the advantage of owning the FAB at this point? Is it that you can do innovative 3D packaging across multiple node sizes, or is it you on the FAB and it's cheaper? Like, what is the specific advantage of it at this point that you can leverage? I think I can give you a very objective answer, having lived on both sides of that equation. I think the empowering thing of the environment Intel offers you is really, there's really no technology that we don't have access to that can be applied to the delivery of

a really world-class customer-pleasing roadmap. If you look at what Intel has access to, it has process technology, it has packaging technology, it has memory, it has interconnect. We own the franchise in a cornerstone IP called the CPU. We're moving in to basically become competent in the GPU space. We already have the FPGA space. We have the software. We have the software assets to drive a developer ecosystem of 20 million plus users. So when you bring all of that together, the explosive innovation you can create with those teams working in a harmonized, synchronized way with no silos creates a roadmap that's just totally, extremely difficult to emulate in a foundry, fabulous relationship where you're trying

to do this through contractual relationships where each participant has potentially different P&L motives. So the ability to deliver products like Lakefield, what we're going to be delivering in our roadmap going forward, where we're bringing all of those technology ingredients together to give a product roadmap, but we believe gives truly delightful customer experiences is nigh and impossible to deliver, I believe, if those assets aren't all controlled in a single company mindset. So is that what you thought when you were at Qualcomm? Like, why can't they get it right? Right? Because you were at the other place, and now you're, you have the place with the whole stack, Intel was not leveraging the full suite of capabilities to be ultra-competitive in a number

of zones. Now you're at Intel. You have control over all of that, I assume. Imagine there's some other managers involved, but you have control over it, all of it, I assume. Are you there? Because the evidence suggests that, like, it hasn't, that vision has not come fully to life, right? That it hasn't been ultra-competitive.

Well, we're not there yet. You're quite right in saying that, you know, we've had to basically get over some, some some basic execution floor. So 10 nanometers is clearly one that we need to recover from and recover fast. But that's not to say the innovation around that hasn't been continuing at full steam. And I think we will come to a point where all of those innovations come together at a point in time where we'll basically be able to show a really competitive and compelling roadmap that basically brings everything I've been talking about to stark relief.

So I think, again, what we're really talking about is, a transformational journey for the company that will take years to achieve. It's not just something a click of two fingers that's done within the transition of one product. It's not how ship design works. It doesn't. You think of it today or be in the market three to four years time. So that's really what we're talking about. So when you talk about some of the – for example, I talked about a technology that we use in our advanced 3D packaging.

It's called Fovar Us. It's going to be delivered in Lakefield. That was a research project for 10 years. And it's how much work that needs to go into some of these technologies before they come to fruition. So a lot of the innovations we talk about have deep research precedence that needs to come to the fore to be able to fully exploit it. And I think I'm really excited by the arsenal of technology we have. And as we go through the continuance of our roadmap, hopefully it will come to light for all to see. The opposite argument is that one of the fabs can spend all that time and money investing in that stuff, researching that stuff for 10 years, and then they can sell it to a wide variety of clients, right? And that cost distribution, that changes the nature of the worthwhile nature of that. That changes the nature of how much that investment is worth. Intel, you're going to put 10 years of investment into it and you're your only client. Do you ever think about, hey, we should start doing some fab work for other companies?

Yeah, absolutely. But I wouldn't discount that as a way we expand our relevance and our strategic scope going forward. But that won't mean we compromise what we need to do to deliver the design points we need for our product. So we see the found re-engagement being a strategic option that's interesting for us. But it will basically be integrated into a mindset where we won't compromise the design points that we need for our high-performance compute solutions. You know, the benefits of of being an IDM is that you can focus the optimization of your process technology to the products that you believe you can deliver disproportionate competitive advantage to. You're not really focusing on amassing so much volume that you define a process that there's jack of all trade and master of none. So one of the things that the foundry is always challenged to is how do they design a process that is broad enough in applicability that they get as many customers as possible.

And that basically means you have to have a fairly wide design point for that process, which basically says you're good at many things, but are you really brilliant at any one particular thing? Whereas in an IDM, with Intel's perspective, we can really dial in on the areas that really matter to us, which are high-performance compute, because we're not worrying about filling our fab full of, for example, low-power smartphone devices. Now, there may come a point in time where we'll strategically broaden what we do to potentially service foundry opportunities, and we wouldn't discount that. I think where it makes strategic sense, we're very open-minded to that. But that has to come at the right point in time, and that isn't today, but that might be sometime in the future. All right.

We've only got a few minutes left. I ask almost everybody this question. I'm very interested in how busy people actually expend their days. How is your team organized? So it's a pretty large team. It's about 65,000 people. And about half of it is in our technology and manufacturing organization. And the other half is in our engineering and product development organization. And it's an organization that is really run by people working very closely together in an environment where as a management team we meet for a day, a month, for about a nine-hour kind of. like review of all of our programs and all of our key technologies. Is that where you make decisions or is that a bunch of slideshows? No, it's really a decision-making meeting that essentially the preceding month of work

ultimately culminates and bringing all the issues that we need to sort out as a team and make sure we're on the right trajectory. But it's very much a scenario of how do you create a tight-knit leadership team with an organization that's very large and geographically dispersed. So it means building a lot of one-to-one relationships as well as team-oriented relationships. relationships. And it's the same paradigm that Bob Swan uses at the highest level. So really what I'm doing is just waterfalling a methodology that Bob has created within the executive leadership team and really using those same methods to promote that integrated leadership through the company.

So you get to Intel, you listen. Everybody says they listen first. I believe you listened. What was the first easy win you took off the board? Rationalizing what we were spending our money on. When I came to Intel, we were spending a lot of money on the smartphone development programs. And really, you step back having come from Qualcomm and looking at the momentum and scale of operation that they have. And sometimes when you come in with that kind of like a clear ride outside in perspective, you can actually see that, you can actually see that, yeah, the challenge you've got ahead of you is a lot, lot stiffer than maybe the internal thinking had hitherto understood.

Yeah, like I came from there. Like, don't. Yeah. So it's really a case of you can stand back and you can basically say, hey, listen, it's probably better to cut our losses and just rationalize that this is not an area where we're going to be able to come back in the final quarter and pull off the victory. And we should be spending our money on and wiser things like improving the performance of our PC platforms rather than continuing to try and make a way of it in the smartphone area. So rationalizing that portfolio is something I did within the first few months I was there and really make sure that we move forward. And it was tough, painful decisions, but ultimately, I think, were the right business calls. What's the next easy one on the board? It's to basically get back to a point where we're firing on all cylinders on our process technology. I think 10 nanometers was a big hiccup for the company that we can't allow to be repeated again.

So get the 10 nanometer product volume rolling, delighting customers a little later than we had hoped for, but nonetheless out there delighting customers and make sure that 7 nanometers. follows rapidly behind it back towards this two-year clip we've talked about. All right. Last question. I'm going to ask everybody this question in 2020. That's my promise to myself. Here it is.

When do you work? Right? Because you're busy. You have to blow your time with me in this podcast. You've got to go on stages. You've got to play an idea. When do you sit down and work? I kind of have a philosophy where I have a lot of interests outside of work that basically are also very, very important to me.

I'm a huge sports fan. I love my family, spend a lot of time with them. I'm a huge bookworm, love reading books, big classical music fan. And those things are important to me. And what I find is peak performance at work comes from basically getting work-life balance absolutely right. So in order to work hard, you have to relax hard. For me, it's all about intensity that when you're at work, you're really focused on the job at hand. And it's not about how long you put in.

It's how focused and how intense and how productive you are when you are at work. So for me, it's really a scenario of, you know, from 8 to 6 p.m. It's all about, it's all about intel. No, but tell me about 8 to 6 p.m. When do you read the white papers that your team generates? When do you do your email? Like, when are you alone with the computer doing work? Because I don't know when to do it.

So my goal is to ask everybody else and try to figure something out. So I kind of like, I kind of like, I get a lift into work every day. So I kind of like have 40 minutes where I kind of like read through my email as well as kind of like spend some time thinking about how I'm going to go through my day. The other thing is I try and I try and run my meetings where no meeting lasts more than 50 minutes in the hour. I kind of like fail more than I succeed. But I do try and give myself those 10 minute times. And what I tend to do is I tend to never really sit at my laptop during the day because when I'm in a meeting, I'm in a meeting, right? What I tend to do is kind of like graze on my phone

in terms of anything absolutely mega urgent that needs my attention. And then, And anything else, I know, okay, that's something I need to deal with. And then I kind of like batch process that either in the evenings or a weekend catch-up times. You know, I love my work, so it's not like I feel like I'm sacrificing for it. So what I find is that when I'm at work, it's really about focus and it's about intensity. You keep an eye on things that make sure you don't miss anything urgent. But then you kind of like take an aggregate of what's in your inbox and then batch process it offline. You literally just spread your own time management as a processor.

I'm just letting you know that that's what your processors do. Okay. Dr. Morthy, thank you so much for joining us. I need a hard out. Thank you. I really appreciate that time. Hey, my pleasure. Thank you for the time.

All right, my thanks to Hymme and Dr. Murthy, Rendez Chantala from Intel. That was a great conversation. We'll be back later this week with the chat show, then the interview show on Tuesday, as always. I love your feedback. You can tweet him at Recklist. Tell me who you want me to talk to you. I love figuring out what guests I should have in the show. And we'll talk to you soon.