Episode Transcript
[00:00:03] Speaker A: Welcome to AINA Insights, where prominent leaders and influencers shaping the industrial and industrial technology sector discuss topics that are critical for executives, boards and investors. INA Insights is brought to you by INA AI, a firm focused on working with industrial companies to make them unrivaled segment of ONE leaders. To learn more about INA AI, please visit our website at www.ina.AI.
[00:00:40] Speaker B: Welcome everyone to another episode of Titanium Economy podcast. We're joined here today by Chris Burgi, Executive Vice President at arm's Edge AI business unit. Now, ARM is one of the most consequential companies in the world today.
Their design and architecture runs 99% of smartphones, the driving Edge AI across consumer to industrial landscape. And 50% of new hyperscaler server chips have ARM architecture in it. So one of the more fundamental companies today, Chris, comes with 30 years of semiconductor experience. He's had roles across large companies like AMD, Western Digital, SanDisk, Broadcom, and then a whole set of communication based startups. And even at arm, you've held a couple of roles. You started in the infrastructure division, then the client, and now you're leading the Edge AI. But Chris, it's our fortunate to have you on the podcast and we're really happy to host you.
[00:01:43] Speaker C: Thanks, Akshay. It's my pleasure to get back with you. It's always good to see a friendly face from the past.
[00:01:48] Speaker B: It absolutely is. So, Chris, before we dive in, I think if I think about ARM, it's hit $4 billion in revenue. You've grown 25% year on year on the top line, about 150% on your operating income.
For the audience who are less aware of arm, can you just tell us what does ARM do and what is driving this accelerated growth?
[00:02:10] Speaker C: Sure, Akshay.
So ARM has been around for a little bit over 35 years now, and the majority of our business over that period has been to develop IP that's used in the building of semiconductors and specifically really the cpu. The ARM CPU around the ARM instruction set has been foundational to that. And then we do a lot of complementary things as we've grown out our business from there.
ARM started back in the early days around a lot of the idea of battery powered computing and portable computing. It was in the first Nokia GSM smartphone, Nintendo ds, even the not so successful Apple Newton were early adopters of ARM technology. But that really started the foundation of getting in with many of these industry leaders. Whether it was the first product was successful or as we moved on, obviously we've seen tremendous success as you've highlighted.
And what we've seen is that this ubiquitous computing platform has served us obviously so well.
We were a big part of the smartphone revolution and that's one of the things we're probably most known for that, as you mentioned, a lot of smartphone platforms are built on it. But because smartphone is such a ubiquitous platform for driving technology that's gotten replicated into TVs, into tablets, into wearables, and now data centers and automotive and autonomous driving and all the things you kind of highlighted. So that's been kind of fundamental to our growth.
A lot of that was we were purchased in 2016 by SoftBank and really what that acquisition allowed us to do was to make some significant investments in a lot of these other growth areas right outside of maybe the more consumer smartphone area. And so we really increased our investments in things like data center, in things like automotive and in AI, actually in many of the things that we saw kind of at the forefront. So we're really seeing now is that explosion.
We have a pretty long gestation period from our concepts of the architecture and what we want to go do, creating that into code that then gang get built into semiconductors that can come to market. So that can be a three, five, even a seven year kind of a period. And so you're really seeing that growth now from those investments. And we continue to even invest more in some additional products. We're seeing a lot of interest in kind of building out those subsystems. So we have now css, which are compute subsystems. We've expanded that even to introducing actually doing our first ever chip that we announced back in March, the ARM AGI CPU for data centers. And so it's a really exciting pipeline of technology.
We've got an amazing ecosystem of partners that we work with.
[00:05:10] Speaker B: So, Chris, exciting story. I think one of the elements of the growth story is the ARM V9 architecture that you've launched last year that's grown significantly. Now this is also where you command almost twice the amount of royalty versus the prior generation.
And then it's got a completely new compute stack with an NPU built in. Can you talk a bit more about what this is? Where are you in the transition? Because it's already growing, it's almost 30% of your revenue.
So where are you in terms of the transition and what's the future look like for this platform?
[00:05:48] Speaker C: Yeah, so v9 actually is actually a journey that we started much longer than a year ago. In fact, we brought our first V9 cores to market, announced those to the market, four to five years ago. And actually that means the technology gestation has almost been almost a decade. And so V9 has been around a couple key things.
One is AI acceleration. We can talk more about that, but a lot of it's also been around security. So it was really a architectural shift around making sure that we had the most secure, highest performance, still very efficient architecture, but then adding in some of the elements that we saw coming around AI and some of the advanced math type activities.
So that has really been adopted tremendously by the ecosystem as you mentioned. It now significantly has changed the economics of our royalties and it's really the most advanced architecture out there because of the markets that we serve.
We are super religious on driving efficiency and also driving evolution. So for example, one of the first architectures to drastically deprecate 32 bit and move fully to 64 bit in our implementations and that just means efficiency. It means efficiency and performance. And so because of the way that our architecture and our ecosystem is set up, we can be super dynamic in driving that innovation while also still having that compatibility of our complete ecosystem. And so that has been kind of our secret sauce. And we just keep driving that forward.
[00:07:39] Speaker B: On the topic of secret sauce, you touched upon ecosystem very briefly right now, but we've discussed that in multiple forums. Like the ecosystem is a secret sauce for ARM. You've got 22 million developers, thousands of partners on it.
Now, as you think about AI coming to four, what does it take for ARM to continue being the center of that ecosystem versus getting routed around?
[00:08:03] Speaker C: Yeah, well, so first off that 22 million developers, what that means is that you're just able to leverage so much work that's been done, so many tools, so much code, all those kinds of things. And when you think about the innovation that needs to happen, whether you're, you know, adding AI to your washing machine or you're building a next generation trading system for a data center, you really want to focus on that innovation and the new stuff. And you really want to stand on the shoulders of something that's solid and something that's extensible. And so that's really kind of where what we've been able to leverage.
And so you know, we, we love to be very specific and we're very religious about that compatibility and the non fragmentation of the architecture so that people can write once and understand it's going to run on the broad, broad ecosystem of ARM devices, but on the other side allow for innovation and customization. So allow for the ability to tightly take AI acceleration and couple it to the computing clusters or take advantage of next generation AI acceleration. That's absolutely happening in the architecture as well.
And then we've expanded that now to our compute subsystems and beyond. So that really is what we allow, is that we give an amazing technology foundation for people to build on top of. But then our partners just be able to trade their innovation and what they bring to market and what they're best at, their special sauce. And when those things come together, that's really why you see our footprint in the ecosystem that it is today.
[00:09:51] Speaker B: I mean, essentially the sum of parts is greater than the intuitive.
[00:09:55] Speaker C: Absolutely.
[00:09:56] Speaker B: Now in. Chris, you're a prolific speaker. I've seen you at Mobile World Conference ces. Latest one was the Embedded conference.
I think one thing which you've said is around how inference is where the value is versus training in the cloud is more than training in the cloud. Can you talk a bit more about where does that conviction come from and how does it shape the strategy for ARM going forward?
[00:10:25] Speaker C: Well, I think that obviously it's a yin and a yang, right? I mean, I think training and the training that's happening and how quickly the models are getting more and more intelligent or getting smaller and smaller or being able to do more and more things. And that's the incredible part of training and why we need to keep driving training. But when you actually see, when you put it to action, that's actually the majority of that is inference. So whether it's your car self driving down the road, well, that's actually an inference workload that clearly is taking advantage of the model that was trained based on all the driving data.
When you are training a model, you're training it once and then it may run a billion times if you think about searches or things that are happening in your phone or on your computing devices. And so it really is about when. How that AI kind of gets monetized or how it becomes physical or how it becomes, whether it's a robot or how it becomes physical in the way that you're interacting. And wow, that just did something really intelligent for me or it helped me learn how to study for my test this afternoon or whatever, those kinds of things. So that's the inference part of AI. And of course you can imagine that that happens a lot more often.
And I think we now are obviously getting into the age of agentic AI where, you know, we're now starting to kick off multiples of these inference workloads in parallel and they're all leveraging each other, etc. So definitely inference is a orders of magnitude bigger than training, but it is a very feedback loop relative to the better the training happens, the more powerful, the more capable, the more proliferated we can make the inference happen.
[00:12:14] Speaker B: On that note, you launched Lumex, your CSS platform, last year, sometime in September, and you're projecting this will be in 3 billion devices over the next five years by 2030.
Could you talk a bit more about what does this particular platform enable and what does it mean for user experience going forward?
[00:12:36] Speaker C: Yeah, so I think so. Lumix is our premium platform for personal computing in smartphones and laptops and tablets, all those kinds of devices.
I think the 3 billion thing we were talking about specifically was around actually SME2, which is actually the matrix enhancements that we brought to the V9 architecture.
And so that now is actually shipping today in all of the premium smartphones, whether they're coming from Android or iOS.
And so that is becoming the foundation now for on CPU AI.
We're seeing a lot of opportunity to take that elsewhere. But definitely doing CPU AI is very attractive, both from a latency point of view, but also from a ease of use from a programming model.
When we get into AI, you know, everyone talks about GPU computing or there's NPU computing, and the reality is, is that, you know, we think about heterogeneous computing. There's different tools for different jobs, there's flexibility, there's ubiquity, there's performance per watt, all of those things are all trade offs that you kind of go through. And so it ends up coming all together in these platforms.
But yeah, we're super excited about where we see our SME technology going and the uptake has been, quite frankly, tremendous.
[00:14:06] Speaker B: Got it. I think shifting attention to the position of ARM in the AI stack, I think more importantly, you're sometimes partnering and sometimes competing with the same folks. Right.
So if I think about the Nvidia Blackwell, that GPU that Blackwell runs on ARM architecture, where you have the Nvidia GPU sitting next to an ARM cpu.
How do you manage this balance between competing and enabling with the same ecosystem partners?
[00:14:41] Speaker C: Well, so, I mean, it's really not something that honestly keeps us up at night because at the end of the day it's all built in the ARM architecture.
And so our partners are what make us great and have caused us to have the success we've had over these 35 years.
But what we've allowed people to do is again build on top of what we provide. I think during COVID when Jensen was doing his talks in his kitchen or he was doing these presentations in the kitchen, he had this awesome few slides that I love to reuse where explain why they were moving to ARM from x86. In the past, Nvidia GPUs data center always ran on x86 CPUs.
And what that meant is that they were limited to PCIe type interfaces, memory interfaces and memory bandwidths that were optimized around a CPU centric world.
And that wasn't the AI centric world. The AI centric world is that in many ways the accelerator and the CPU need to be of equal importance or may have different set of requirements. And so by using the ARM CPUs and partnering with ARM, they were able to build a system that was balanced and it had the right amount of bandwidth for the CPUs and the GPUs. They were no longer I o bound. They could create those kinds of things. So that's really where that kind of one plus one equals three really comes to bear. Because if they didn't have access to ARMS building blocks, that kind of innovation wouldn't have happened. And so we're all about driving computing forward and that's why we think we've obviously had tremendous success and we also think we're just very well positioned for the future of computing.
[00:16:34] Speaker B: Got it. And I think on the topic of the X86 market, I think one of the things you've stated is 40% of PC and tablets this year would be ARM based.
So you're structurally getting into a market which was owned by the x86 architecture. If you think about this, how does ARM continue to win in this architecture and what are some of the things that need to go right for ARM to continue growing?
[00:16:58] Speaker C: Well, it is things that we've said and we believe those will become true.
And it's really about delivering what users want.
Right.
I've worked closely with the PC industry back from my AMD days two decades ago, and we used to think about laptops that you would, I would call mobile went off, right, because they were many pounds and you had, not only do you have the laptop, but you had the extra bag with their huge DC charger and everything else. And now, you know, you expect these thin and light laptops, you expect them to be fanless, you expect all day battery life and you expect to be able to charge them with your USB C charger that you also use for your phone. And so that drive to kind of what people expected or what was possible.
Many of Those innovators that are driving that kind of innovation are doing so with the ARM architecture.
Secondly, you've seen the PC morph into leveraging many of the things that the smartphone was so good at and pioneered.
Covid drove you to wow. Video conferencing wasn't a thing that happened every once in a while, it was the thing. And you wonder why people don't have their camera on now. Well, guess what that meant you need to have great subsystems, you need a great camera, subsystems, all those kinds of things.
And you needed to have great audio because you were now listening to those devices.
Now AI is the killer workload and we need to accelerate AI again. The first copilot PCs were ARM based that Microsoft introduced. So all of these waves of the future are largely built on ARM because of the industries they're coming with, because of the partners. And it's really about just user experience, both relative to choice and innovation.
[00:18:54] Speaker B: So shifting from you're disrupting x86 architecture, RISC v is becoming a bit more real and gaining some real traction right now it's an open source alternate to some of your own work.
How do you think about the competitive dynamics in this place and what's ARM's answer to this?
Let's just say justify getting traction.
[00:19:21] Speaker C: Well, so computing is a huge space, right? And we're very fortunate it's growing the way it is.
We are being challenged to provide more performance, more capabilities in so many of the markets that we've talked about. And so we're very focused on quite frankly delivering those people are looking for a non fragmented architecture. They want to leverage all these assets that developers, the tools, all those kinds of things we talked about.
And so for the computing platforms that really matter, we really are focused on driving things forward from an innovation point of view.
And we're seeing that be still, that is the driver of the future. And so we're just excited about that
[00:20:05] Speaker B: opportunity, I think shifting a bit away from your own personal journey.
So because you, you worked across the entire spectrum of semiconductors, right? You, you were at amd, you were at Western Digital and Sandis Memory company, you were at Broadcom. Now you're at arm. Just the spectrum of all of the industry that you've seen, right?
Can you talk about what was the learnings in those early years that you continue to sort of carry with you through each of these leadership roles?
[00:20:41] Speaker C: Well, there's a lot of different ways I could go with that actually. I'm actually passionate about talking about this because it's you know, it's been an amazing journey for me.
You know, I think you take something different from every role, right. I think that, you know, the startups is a very interesting one just because, you know, if you don't do it, no one else will. Right? And so you've got to really carry, get things done. It's not like you can hide out in a big organization.
And so I bring many of those concepts to my team and saying, look, how are you making a difference today?
How are you making an impact today? And really trying to get the team to feel like they have that onus on them, that they are entitled to make a difference and they shouldn't always ask for permission, but drive things and tell people what they're doing and why they're doing it. And those kinds of things, I think that urgency and those kinds of things that come from that startup mentality, you know, I think you, you learn something different from each industry. I mean, I think you mentioned my experience in the memory industry.
It's super interesting. You know, you're working in a commodity industry, yet you're trying to drive differentiation.
You go through boom and bust cycles that are tremendously humiliating and keep you honest and keep you on your toes, but it also just drives you to the, you know, you can't be right all the time, but how are you adapting? How are you moving quickly?
And then I think working for some industry leaders kind of when I was at amd, we were a distant number two to intel. And so you get that scrappiness, that chip on your shoulder of hey, we're going to make this happen. And it's fun to fight and it's fun to not be the incumbent. And then at a company like Broadcom, we just had what we thought were some of the smartest engineers and they wanted to make, you know, really make, drive a change in the industry and take some, you know, some technical risk. And we just knew we could make it work and make it happen. And so we really kind of reshaped the industry. So each, each role you just, you take different things from it and it's, it's been an amazing journey for me and I'm, I'm just excited to give back to, to the teams that I get to work with. And arm, you know, ARM kind of takes the cake relative to having just an amazing innovation, kind of a mindset, deep roots in academia and our Cambridge heritage, and now just driving that out to the amount of impact that we can make. So it's been fun and that's a
[00:23:24] Speaker B: fabulous leadership arc across all of these various semiconductor companies.
I think Krish closer to home in ARM itself, you started in the infrastructure space, then you went and you ran the client division. Now you're leading edge AI.
I mean, across each of them, you had to sort of build credibility in a new domain with new customers. Can you talk about what this transition entails just within arm, and then what do leaders typically get wrong going within different domains?
[00:23:55] Speaker C: Well, actually, you know, I. I'd love to say that I had a perfect plan, but I got to say, you know, I think like most of my career, you. You take kind of things as they come. And I actually spent a good amount of my career both at, you know, in endpoints as well as in.
As well as in infrastructure. And so actually, in my career, it uniquely kind of set me up to maybe make some of these transitions. But, you know, I think at the end of the day, if you're passionate about technology like I am, there's a lot of leverage. And in fact, that's fundamental to even ARM's business, right? I mean, we do create CPUs for many different markets, but they're highly leveraged. Right? And so those core capabilities, those things do cross over, but the applications and those things, I think, are quite different. So I think the key is, first off, you need to put very intellectually honest people around you.
And by the way, customers are awesome relative to a truth serum, right?
They kind of aren't too shy, usually to tell you where you're doing well and where you're not doing well, or at least where their pain points are. And so I think you really need to listen to them.
And so for me, it was, you know, come in, learn a bunch, surround myself with people that complemented my capabilities.
And then I think you go through some of these scenarios where the technology didn't matter, but you kind of understood, how do we win in this scenario? How do we really understand what our value proposition is?
What part of the market do we really think we can penetrate? Should we go where the market was or should we go where it's going to be?
Quite frankly, when we came in do, it started Neoverse, we just said we're going all in cloud, right? Even though that wasn't the biggest part of the market. I mean, you look back now and go, wow, Chris, you were brilliant. No, but that's true. We literally said, sorry, when the big OEMs call, we're not answering their phone calls. We're focused on Amazon, we're focused on These cloud providers. And so we placed that bet and said, we think that's the future of computing, that's where we want to be.
We thought that we had unique attributes and it paid off because we didn't have the inventor's dilemma that some of the incumbents had around. Well, wait a minute, we do all these things over here, we just want to reuse those things. We were like, no, we're doing something new. And we found key leverage points.
We build great CPUs for data centers, but quite frankly, we built an amazing fabric for data centers. And with this AI acceleration and the coupling of that, that has been the magic of you can have amazing bandwidth, you have amazing multi compute type of a solution. So really finding those key differentiators.
We're very fortunate that we work with some of the most innovative companies out there and they like to partner with us because they think our engineers are smart and we give them access to those engineers and we listen and we really try to deliver value, I think, to them.
And so it's really about that kind of, I think, being intellectually honest on the technology and then understanding the dynamics of the industry, because each industry is quite different. And even in the six and a half years now that I've been at arm, we've seen these waves, even in those time periods of what's important, are you going through consolidation, are you going through expansion, are you going through.
New applications? Quite frankly, even in one of the areas that you would say maybe is more boring, like industrial or IoT, that is going through a huge revolution right now because we're going from computing to AI. And so now how do you leverage AI in all these endpoints? And now how do you orchestrate all those things? And many of the initiatives where we've tried to orchestrate these devices, it's worked somewhat. But now with AI, we can, I think, create these orchestration layers that never existed before. We tried to do with standards bodies. Now we can have AI agents figure out how to orchestrate these things. So I just, you know, to me, this is what gets me up in the morning, is that, you know, even though these industries are, you know, from a 20,000 foot level, look very stagnant, when you actually look at what's happening underneath the water, you can actually see that, wow, there's a lot of fundamental change that's going to happen and that'll really separate, I think, the future winners and losers.
[00:28:48] Speaker B: Chris, I think it's refreshing to hear about the unencumbered approach. You went all in into Cloud, Right.
And the whole point about listening to your customers, being close to it, throwing in an additional question. Around industrials, right? Like in Cloud and in Compute, there are some large customers you can sort of work with. But as you think about going towards industrials, which is a more fragmented space, is that a place where you've decided to sort of, how are you working with customers and industrials, how do you view that space? Like, who's the, the partner to sort of get you into that edge AI adoption at industrials?
[00:29:23] Speaker C: Well, I mean, we're fortunate in that, you know, most of the industrial companies, at least at the, at the building block level. Right.
You know, that's been a foundation for ARM for a long time. And so, you know, companies like ST Microelectronics, ti, Infineon, and then you got newer players, you know, that, that are evolving and kind of coming from say, wireless backgrounds and power backgrounds and all these kinds of things. So we've got this amazing breadth of partners.
But I think the key thing in industrial is it's more fragmented. You don't have units or a single unit that sells in the orders of 100 million units or whatever kind of a thing.
So it's really about trying to build building blocks that can be highly leveraged and customized and optimized for the different use cases. And so I think it is a market where there's probably.
You wouldn't want to necessarily listen to a single customer, but you have to be a little bit more passionate about what you think the key building blocks are for the future and allow people to kind of build on top of that. Because it is a market that looks for stability, it looks for kind of capabilities that it can extend over time and those kinds of things. It's not a consumer market or anything like that. So I think it's about bringing in innovation, but understanding that innovation will be not as rapid as a consumer market. But what we still do is we do a lot of the fundamental plumbing to try to make it easy. Right. I think about, you know, frictionless AI, right? How do we make AI frictionless? How do we make it easy for people to make that transition from a compute centric to kind of an AI centric world?
[00:31:29] Speaker B: And you put the ecosystem and the developer partners that you have and you've got the building blocks and the accelerators put together.
[00:31:36] Speaker C: Correct.
[00:31:38] Speaker B: One thing did stick with me at Computex. You said AI is changing the playing field and ARM is basically laying out the board.
So if you think about looking at decade out, what does the board look like for you and then what has to go right for AI to fully realize its potential in the AI native world?
All built on the fact that he said the field is changing and ARM is building the board.
[00:32:08] Speaker C: Well, I think when I talked about that, I was really talking about how we're allowing our partners to kind of innovate. It kind of goes back to my Jensen example earlier about, hey, you can build a platform the way that you want to build it because sometimes you need amazing memory bandwidth, sometimes you need I O bandwidth, not one size fits all. And so really that allows us to be this foundational element behind a lot of these different things.
I think that it is, I mean, it is just mind boggling to me to see how fast it's moving. And I think any, even those of us have done this for a long time, this innovation cycle is amazing.
So I mean, clearly there's some fundamentals you can look at from a physics point of view of, you know, it's very clear that, you know, power, memory bandwidth, you know, these things are just, it's going to just keep challenging those metrics. And so how can you do things more efficiently? How can you do things, you know, how can you manage how you can build better models that don't need as much memory bandwidth? And you know, you're kind of just squeezing the balloon of just trying to figure out how to optimize things. I think things that are maybe not as obvious to, to folks is developer friendly.
AI is not easy. Many of these low level programming capabilities and obviously you see that Nvidia has done an amazing job with CUDA and really have invested in that ecosystem to really try to make AI programmable and improve that.
I think that is a big challenge that we're focused on from an ARM point of view of making sure that AI is tremendously developer friendly because, you know, that is key considering how fast things are moving.
And you know, I think it's been one of the reasons why Nvidia has such success on the accelerator side and quite frankly it's one of the reasons we've had some success, so much success on the computing side of AI.
[00:34:16] Speaker B: Chris, one final question before we close and probably be asking you to do a bit of crystal ball gazing here.
Now, you've gone through several cycles. You've seen the PC cycle, the mobile, the cloud and now AI. Now we're right in the middle of the AI cycle and this is probably one of the biggest you've seen if you were to project ahead.
What does this era mean for the semiconductor industries.
And where does ARM continue to sit in that future?
[00:34:48] Speaker C: Yeah, I mean, I think that the, you know, one of the exciting things for me, and I know you've heard me, I've been kind of passionate about this for, I don't know, it's been almost a decade now of how much the CMOS sensor changed the world.
Right. If you just think about these CMOS sensors that originally used for camera, like use cases.
And then of course we had the smartphone boom where, you know, they took tremendous advantage of those sensors and then we started putting many of them in and then that became the fundamental for computer vision. And now this is how we have kind of all of these computing things. So I think that there's, there's not a sensor, there's not a device that's created more data on the planet than the CMOS sensor. And I think that that's, that's been one of the kind of unsung heroes of this whole thing and that drove all the computing, all the things that we kind of needed behind that.
I think the next wave is going to be fixed physical actuation.
So as you think. And you can also say that it also drove the lithium ion battery, the whole battery generation thing.
[00:35:54] Speaker B: Right.
[00:35:55] Speaker C: I think physical actuation is actually going to be one of the big waves that we'll see happen is that if we want to bring computing more into the physical kind of actuators and how do you create a, you know, a hand like device that has 33 degrees of freedom or, you know, whatever those things are. And I think that'll be in a tremendous building block of kind of again, as we think about applications and really what is the size of the market.
But I think the computing blocks become, you know, stay the same. Right. I think relative to, you know, silicon will continue to be the computing element.
I think the having a broad ecosystem of computing devices, power, performance will continue to be huge on that.
So I think in many ways there'll be all these new applications and that's what's going to make the TAM be so much larger than we can comprehend today.
But when you get down to some of those fundamental building blocks, it'll be back to the same tried and true reasons because you've got so many things, why go change that?
And that has also evolved into being such an amazing platform because of all the advances that will happen.
Obviously we've got all the advances now around security and what security means.
That is going to continue to be, I think, a huge forefront for the future. Of computing and as well as obviously all the how does AI collaborate and then all the privacy and those kinds of things we've got to deal with as we've always dealt with over the last many generations of computing.
[00:37:40] Speaker B: Chris, you talked about the agility that ARM has because it's primarily an IP and a design company don't have to deal with the fab portion of it. And you, you weren't taping out silicon, but now you've gone to taping out your own silicon.
Is owning a fab next step or is that a future too far?
[00:38:05] Speaker C: That's a future very too far for me.
But just to be clear about why I think that arm, it's not about not building silicon, it's about the fact that we made our technology so widely accessible.
And you know, that sounds like really easy to do, but it's actually not because that meant that we had to create with great recipes, we had to create modularity that allowed us to, you know, make a business out of serving a ton of different players. Right. It wasn't just a one to one thing. And I think, you know, in many ways this is kind of, you know, if you want to look at the TSMC special sauce versus the intel special sauce, right. TSMC had to figure out how to support all of these semiconductor companies and guess what, they got amazingly good at it because there was no way you could call up, you know, 15 year customers, say, oh, you know, we told you the transistor was going to work like this. It doesn't work like that. Can you please tweak your design? Whereas when you had companies like intel, they had a one to one relationship and so they could tweak both sides of it in some ways. Okay, yeah, you can get to maybe an end plane faster, but when you actually look at your scalability, you end up sacrificing that a lot. And so I think it's really that scalability that we were able to build. But no, we have an amazing set of partners, foundry partners across the board, whether it's tsmc, Samsung, intel and the like. And so we are actually very supportive of broad set of foundries. I think that if you look at the, what it makes, what it means to run a great foundry, there's not a lot of overlap with what it means to build a great product. I mean those are very different capabilities. And I think that we're going to keep doing, focusing on what our core business is, which is building great computing elements, putting them together, building more and more valuable subsystems because that's really what our customers want and bringing those to Silicon where it makes sense.
[00:40:13] Speaker B: Well, Chris, on that note, I would want to thank you for the time and for joining us today. It's been an absolute genuine conversation which I believe our audience will enjoy. And thank you very much.
[00:40:25] Speaker C: Thanks, Asha. It's my pleasure.
[00:40:33] Speaker A: Thanks for listening to INA Insights. Please visit INA AI for more podcasts, publications and events on developments shaping the industrial and industrial technology sector.
