 Welcome to this CUBE Conversation. In this segment, we're going to talk about the future of IoT and the critical role semiconductor technology plays in shaping this exciting space. As we've reported on our breaking analysis segments, the fabulous chip company enabled by the ARM ecosystem has permanently changed the semiconductor industry. Intel's fateful decision in the mid 2000s to pass on the chip design for the Apple iPhone was an ironic reminder of IBM's decision to outsource the microprocessor for the original IBM PC to Intel. In both cases, the market leader didn't appreciate the tectonic industry shifts that were possible. And importantly, the impact that volume economics would have on the power structure of the entire industry. Now fast forward today, and we believe ARM wafer volumes are 10x those of the general purpose x86. This means that the ARM ecosystem is on a cost curve that is unmatched in the business. Moreover, as we've reported, the ARM ecosystem is blowing away the historical performance curves that we've seen in the chip industry, aka Moore's law. Whereas for years, the x86 performance curve grew roughly at 40% per annum is now moderated to the low 30s. Over the past five years as evidenced by the progression of Apple's a series chip based on ARM, when you observe the combined performance of the CPU, the GPU, the NPU, the XPU, DSPs, accelerators, et cetera, the alternative processors in combination have driven the average annual performance improvement to over 100% per year. This is an astounding achievement. Why is this so important to IoT? Well, the edge is projected to be the next trillion dollar market. We believe we'll see a world with more than a trillion devices and as we've reported, IoT use cases are going to require specialized and distributed processing power and lots of it. AI inferencing at the edge will enable real-time action and embedding intelligence in the chips that win the edge will be high performance, low power, inexpensive and programmable with a much faster time to market profile than historical semiconductor cycles. We're already seeing that with companies like AWS, Apple, Tesla, Ampere and others going from design to tape out in under two years versus the historical norm of let's say four years to be generous. And with me to discuss innovation in IoT and some big news from the 2021 ARM Summit is Mohammed Awad who's the vice president of IoT and Embedded at ARM, Mohammed. Good to see you, thanks for coming on. Thank you, thanks. I really appreciate the opportunity, thanks for having me. So you're welcome. So tell us about your role at ARM. I know you were responsible for infrastructure previously and now sort of extended to IoT and embedded. Tell us more about that. Yeah, sure. So I've been with ARM for a little over three years now. I started working with the infrastructure team where we worked on a lot of different initiatives. One of the things that we launched was our ARM NeoVerse and we went on to do some interesting things there. As you mentioned, we're making some great traction in the infrastructure space. About a year ago, I took on the role to head up ARM's IoT and embedded business. And it's interesting because my career really started in IoT and embedded. I was in the Boston area working for companies like Lucent, Nortel and then eventually Ember. Very early IoT startup. So that was 25 years ago now, but I still got roots in the Boston area. So I like your hat in the background now. Yeah, right. Go socks. Go socks. So how do we get here? You've had a lot of experience in embedded IoT, which is relatively new term to most people. It sort of evolved from a period of, you had instrumentation for at least some components of the system. And then we focused on connectivity. But as I was saying in my upfront narrative, we're really now embedding AI and its intelligence. So there's phases. How do you see the progression in terms of how we got to where ARM is today in IoT? Yeah, it's really interesting because if you think about, if you think about ARM, and then you really just think about IoT, as you said, INT started off with, hey, let's stick a microcontroller in everyday devices. Let's stick a microcontroller to something like a vending machine. Then we went on and said, well, hey, what if we could remotely control that device or gather data from that device? And then we, so we entered this phase of, what we like to call intercom activity, right? And that was all about connecting these devices with things like low power Bluetooth or even now low latency 5G. And what's interesting is that, together the work that the ARM ecosystem has done over the years has really solved the problems of how to add microcontrollers or connect to the devices. I mean, those problems have largely been solved for a lot of the reasons that you described earlier, which is, we focused on lowering the barrier for folks to come in and innovate around sort of a core technology and lots of innovation happened as a result of that. We're entering this new phase now, which is really about, you get all these devices out there which can easily be connected. They've got microcontrollers or technology in them, which allows them to be intelligent. But how do we really extract the level of kind of AI intelligence out of those devices? Ultimately, what we're trying to do is, is the industry needs to figure out how to derive intelligence from the smallest sensor all the way up to the largest cloud data center. And that means local intelligence. It means regional intelligence and it means global intelligence. The potential is enormous, but the challenge is pretty enormous as well because of all those diversified use cases, all the diversified devices, all of the sort of scale and sort of number of platforms that we're talking about. That's really what we're excited to kind of go work on that. It is exciting. I mean, just it's mind boggling the capabilities, the processing capabilities of this distributed world that we're evolving towards. Let's talk about the hard news. Why are you announcing what you're announcing? I mean, what are the trends that are sort of informing that? Maybe you could hit some of the highlights of the announcement and give us the key details. Yeah, sure. So what we announced is arm total solutions for IoT and that's really made up of three things. My favorite part is on virtual hardware. Our virtual hardware is all about making available a virtual representation of devices in the cloud for lots of developers to use. And I'll get to that in a minute, but I think, in order to understand that, you have to kind of understand the broader context of what arm total solutions are. It starts with pre-integrated, pre-verified IP package. You talked earlier about how design cycles we're looking to accelerate and people were looking to develop a silicon much faster. Part of what we're doing at ARM is we're actually taking pre-integrated, pre-verified IP packages, we call those ARM core zone. We're making those available to the market. So we give those to our silicon partners and then they can use that. They might include a neural processing unit, a CPU. They might include an interconnect, all the kind of the base IP. And then our silicon partners can use that as a jumping off point so that they can quickly get silicon to market. That's the first part of the news, which is, we're doubling down on that now. In the last three years, we've had over 150 different designs which have used our ARM core stone products. So moving forward, we're going to make that foundational to how we deliver IoT technology to the market. But the second part of it, which is super exciting is that, not only are we going to accelerate the time of market for our silicon partners, we're also making a virtual representation of that underlying core stone design available in the cloud for software developers all around the world to use at the same time that IP is ready. So at the same time we hand IP to our silicon partners, we're making a virtual representation in the cloud so software developers can start. Now, let me just take a step back here and make sure that everyone kind of understands how big of a deal this is, right? Before the way this used to work, I would hand IP to a silicon partner. It would take them 18 months, maybe two years to get a piece of silicon in market. And then a board manufacturer would have to go off and then only maybe three or four years later could the software developers start five years to get a product to market. What we're doing here with our total solutions we cut that five years down to three years so we can massively accelerate time of market. And then the third part of what our total solutions is is something we call Project Centauri. Project Centauri is about putting in place a set of standards to ecosystem initiative which puts in place a set of standards, reference software and specifications around things like security and how devices should communicate with the operating system or cloud service providers that allows software developers to get a level of reuse and leverage. So today in the IoT, every time you develop a piece of software you're going to develop it over and over again. But what we're talking about here is they can develop it once and be able to apply and reuse a lot of that software over and over again. The same way they do in other markets like infrastructure and mobile. Love it, so okay, I wanted to ask you if there's a blueprint there that we can learn from but before we do that. So if I go back to the three items that you mentioned. So for example, one of your licensees can say, okay, I want to take just the standard components, the CPU, whatever. But I might want to customize the neural processing unit as you said and they have the flexibility to do that at the same time when they bring it to the foundry because it's a standard platform that you know it's going to work. That's kind of a nuance that maybe people maybe don't fully appreciate. Am I getting that right? That standard platform has dramatically changed the industry. Yeah, that's right. I mean, the idea is that we take these IPs, we integrate them together, we verify them, we design it as a subsystem, we target specific use cases and then we make them available. Our partners are certainly free to go off and make modifications to it and what they see fit. But when we hand it to them, it's ready to go. And that's the other idea. Yeah, and then the point about being able to give developers access in the cloud, we've often said that developers are going to shape IoT. And so I think what you're saying is essentially instead of this linear process where you can get dependent on the previous one being done, you're actually parallelizing, if you will, the innovation. Yeah, that's exactly right. And I'd actually take it one step forward. There's a subtlety there, which I didn't comment on, which I think it's important to call out, which is not only are we parallelizing, but we're enabling what I'll call modern development methodologies, right? The way that development is done in areas like mobile and in the cloud data centers, they use agile workflows, things like continuous integration, broad-based testing as they go along, that's very different than the way that embedded development is done today. Embedded development today is done the same way it was 25 years ago. You get a board on your desk, you mess around with a bunch of jumpers and cables and wires, hope you did it right, and then you write your software and you hope the hardware guy doesn't want to revise the hardware, and then you're going to start all over again, right? You know, the last thing that you'd want to do is set up a hardware form, right? Lots and lots of different hardware to go off and test over and over again. Now with virtual hardware, you can move all of that to the cloud, all that complexity goes away, and you've massively reduced the investment required for software developers to get going and allow them to take on these more modern techniques. Well, Mohamed, thank you for clarifying that nuance because we're going to see a renaissance in the way that embedded development occurs. And I'm curious as to how you think about that in terms of, because you're going to have a whole new breed of developers come in with the cloud developers, if you will. They have, they see IoT as a massive opportunity. As well, you're going to see the, I would presume the embedded ecosystem upskill, much in the same way you're seeing Opsdev or DevOps or IT people learn Python to upskill. So you're going to see like a two vectors of innovation in terms of developers coming together. How do you see that? Yeah, that's exactly right. And that's exactly what we're driving to. And when we talk about this, we talk about changing the economics of IoT. That's exactly why. Because what we're saying is that, hey, you can have all this massive innovation that can be unleashed from all these developers that didn't have access to these devices before. And you can also take all these embedded devices, embedded developers can make them so much more efficient with these new modern development methodologies. Combination of those two things is going to, not only is it going to lower the cost of development, but it's going to spur a massive amount of new innovation in all new products and services, right? We really think an unleasable potential right to. So step back a little bit, help us understand kind of how you came to this, your strategy. I mean, what are the friction points that you see in IoT and embedded in terms of being able to scale this capability? Yeah. Yeah, it's a great question. And I got to tell you, when I came into this role, the first thing you do is you go off and talk to customers and partners and you try to understand how people are using. But most of the time when people think of ARM, they think of us as, hey, they're the guys that are off talking to the Silicon partners or talking to the hardware guys. And we absolutely do. We have strong relationships with all the Silicon partners. But because of our place in the ecosystem, as a company, which we've got shipped over 70 billion Cortex-M devices to date, we underpin, the IoT basically runs on us. And so a lot of what we do too is we talk to the software ecosystem, we talk to OEMs and we talk to service providers looking to capitalize on all of that on the depth and breadth of our ecosystem. And when I talked to OEMs and when I talked to software service providers, two things became really clear. The OEMs wanted to find a faster path to market. They're like, it just takes too long for us to get our products to market. We need to figure out how to streamline it. So that was one. When I talked to the software service providers, they came to us with a little bit of a different problem. What they said is like, hey, we really want to deploy software and services across this IoT edge space, but it's just so diverse and so massively complex. Everybody's got a different view on things. Can you help us? Where's the, you're the common denominator. Can you help us figure out how to attack this problem? And that's really what drove us, right? Awesome. Let's talk a little bit more about some of the announcement details. Project Centauri in particular, what are some of the things that you want people to really appreciate? And specifically, what does it mean to the ecosystem? I mean, you touched on a little bit, but I don't know if you have any examples or customers and maybe also Mohammed, if you could help us understand how it relates to other ARM projects like Cassini. Yeah, sure. So project, so two things. So first of all, let me just talk to what projects Centauri is. So Project Centauri is really looking to help enable a level of software leverage across that diverse M-class devices that are out there, M-classes that are microcontroller devices that are out there. And so it's really made up of three parts. One part is all about security. So it uses PSA and our PSA certified framework, including TFM, Trusted Firmware M. So this is our security framework that we put forward. And then the PSA standards initiative that's out there in the marketplace, all of the efforts that we bring to bear around that. The second part of it is around open CMSIS and open CDI CMSIS, which is really about standardizing aspects of how software is delivered to an IoT device, packaged and delivered. It's also about things like how any RTOS, so any real-time operating system or any cloud service provider can be accessed from the device. So the idea is that today, if you think about the way that this works, if you're a silicon provider or your hardware manufacturer, you have to go off and support multiple different cloud service providers. You may want to support multiple different operating systems depending on which particular OS you're interested in. And what we're trying to do with Project Centauri is to specify key attributes of the services that exist down on your silicon so that you can more easily integrate with whatever OS you want, whatever service variety you want on whatever hardware you want. It's still allowing plenty of differentiation. So it's not like we're saying, hey, this is how we actually do over-the-air updates, for example. Rather, what it's saying is that, hey, this device supports over-the-updates. If you're going to ask for that service, here's how you present yourself. And that allows a level of software portability that you just didn't have in the IoT space previously. Right, and then the licensee can tune that to their specific use case and add their own value, right? And so again, to go back to the thing we talked about before, they know it's going to work and they can give it to the foundry and say, make this according to the spec and the foundry's ready for it. That's how we've seen such massive volumes. I wanted to ask you about security. You touched on that. Do you leverage realms in this or is that not in scope? Is that to collect it? No, right now that's more focused on our A-class and B-9 stuff. And you actually asked about Project Cassini a little bit earlier. Project Cassini is really our initiative which is focused on our A-class devices. So our A-class devices typically run what I'll call a rich OS, like a Linux or whatever. And it's really designed for allowing a level of virtualization and allowing a level of shared resources between different containers on an A-class type system so that you can easily deploy and leverage the A-class device resources by different workloads. So the reason I asked, I'm trying to, Mohammed, connect the dots between mobile as kind of a blueprint which can occur for IoT, I think that's maybe. But even some of the stuff that's going on in the data center, it's particularly as it relates to data intensive workloads, some of the work that we've seen that AWS do and offloads, we're seeing all the new, like all the modern storage and networking and security offloads in the data center are moving to ARM. And it just seems like the use cases for ARM are exploding. And I'm wondering if you can help us connect the dots into IoT, which could dwarf all of these markets. Yeah, I mean, what's interesting, what we saw happen in mobile and what we saw happen in the infrastructure, what we see happen in both of those markets is that by creating a level of consistency and how software can be deployed on these devices, whether that's with the mobile phone and the ARM ecosystem and the mobile phone or all the way through to the data center, what you've done is you've unleashed a tremendous amount of innovation. In the mobile space, there's something like three million apps out there today and thousands of different smartphone models. Could you imagine if every one of those app developers had to test their application on every mobile phone in order to be sure that it worked? You'd have a lot less innovation, a lot less scale and a lot less applications. And so what we're talking about here is trying to unleash that same amount of value by creating that consistent. So that's a clear lesson we learned from both mobile and from infrastructure. The other thing that's clear is that a lot of these markets you've got, back to the idea of parallelized development flows and subsystems. And that's directly kind of what we're seeing or what we're putting forth with on total solutions. Yeah, you know, it's kind of buzzwordy and people who watch my program know I'm kind of a fanboy of the ARM model. But you talk about the new IoT economy. In my view, you're actually an underpinning of that economy. I mean, everybody talks about it, this multi-trillion dollar opportunity. But how do you think about this new economy? And we've obviously touched on it. But IoT, the edge, it's really taking shape now and becoming real. Yeah, I think that the idea here when we talk about new IoT economy, very clearly what I'm referring to is this idea that, you know, you've got today, you've got a lot of potential which is lost because you're limited to just the, a vertically integrated solution, software is vertically integrated on the specific hardware and the barriers and the cost investing in that hardware from a software perspective is just too high given the sort of scale that you get with that software after the fact. So we're addressing that in two vectors, we're simplifying it so that lots of different developers, you know, that developer that's sitting at the coffee shop can spin up an AWS instance with our ARM virtual hard running and write an app while they're sitting there. And at the same time, they can access a much broader set of devices than they would have been able to otherwise. If it's not, you know, it's not dissimilar, you know, I hate to keep going back to mobile, but it's not dissimilar from the mobile space where if you think about 15 years ago, when all of the applications that were written on your mobile phone were written by the phone manufacturer, you had a limited number of applications and sure phones were a great thing, but it was nothing like it is today. It was a mobile phone economy. Today, when you think about mobile, you know, mobile really underpins the financial economy, it underpins, you know, the transportation economy, it underpins how we communicate with everybody with social networks and it's really taken a sort of life of its own in lots of different ways. It's not really a mobile economy, it is the economy. And we think IoT can be even larger than that, right? Yeah, you know, I mean, our industry has a tendency to hype a lot of new waves, but they certainly didn't overhype mobile. I mean, everybody, you know, migrated to a mobile. That's why I think it's such a relevant conversation and so adaptable to IoT. I think cloud as well, data as well, you know, they were probably underhyped, if anything. Social, maybe we can put over here in a bucket. There's a lot of friction in social. A lot going on there, yeah, a lot going on there. Right, but those are the three in terms of sort of enterprise and the edge. And I think, you know, from what we can see, ARM has really, in the ecosystem, has completely and permanently altered the shape of the industry. It's a very exciting time. And I think the best is yet to come, Mohamed. I really appreciate you coming on theCUBE. Thanks so much. Yeah, no, I really appreciate it. I think thanks for taking the time. All right, and thank you for watching this CUBE Conversation. This is Dave Vellante. We'll see you next time.