The Next Wave with Young Sohn – Unleashing New Frontiers – From Chips to Industries

Show video

Hi Simon, good to see you again! Thanks  for joining us for The Next Wave!   You know, during this pandemic, it's really hard  to do things face to face, but I'm glad at least   we are able to do this virtually. So, let's start  with Arm. You know, I think I am of course privileged   and have been involved with Arm for many years  as a customer, as a board member, as a partner,   but not everybody knows Arm. So I think it'll be  worthwhile for you to educate the audience about   the role of Arm, what Arm does and (give) a  little historical perspective as well.   Well, first of all thanks for  having me today, it's great to catch up!   We've been friends for a long time and it's  good to see you. So, Arm is a   slightly unusual business. We often get described  as a chip company but that isn't what we do.  What we do in our core business is design  microprocessors, but instead of putting them inside   chips and selling chips, what we do is license the  designs to sending it up to companies for them to   use as a building block within much more complex  devices that they're making. The company will be  

30 years old in November, so we're about to celebrate  a big milestone. And at the time the company   got formed, semiconductor manufacturing  technology was just getting to the point   where it was feasible to start building much more  complex chips that became systems in their own   right, up to that point if you wanted to build  a system you took a bunch of chips which had   individual discrete functions and you built the  system on a printed circuit board. With each   generation of process technology, we've been able  to cram more and more transistors into a chip   and that gave rise to the ability to put those  systems inside a single piece of silicon. So when   the company was formed the idea was, well, we have  this low-power energy-efficient microprocessor,   I wonder if it's going to turn out to be useful  to embed that inside a more complex chip. We didn't   know what the killer application was going to  be, but it turns out that in pretty much every   application embedding a microprocessor adding  intelligence turns out to be incredibly useful.  Yeah, I mean I remember the discussions at  the boardrooms, and we're looking at number   of applications and it's just incredible how  widely Arm's architecture has been adopted   around the world. And obviously, for  consumers that don't know much about Arm,  

they probably have no idea that they actually  are living with Arm every day. Absolutely yeah,   the business model of licensing  technology to other chip companies to   enable them to put intelligence into a chip  to build an ecosystem around it has proven to be   incredibly successful. We have today over 500  different companies around the world building   chips based on the Arm architecture there's a  big ecosystem around that providing software   and other design services to make that all  happen, but the real impact   though is in all the consumer devices, all the  devices that go into enterprise networks and   drive the internet. There's just so  many of them that are now based on the   Arm architecture. Over the history of the company  there have been now over 180 billion chips built   using the Arm architecture, which is quite  an expanding number. It took something like   11 or 12 years to get to the first billion  chips and now our licensees – our partnership –   more than 22 billion a year. I mean it's quite  incredible, there's something like over 700 per  

second being manufactured. The scale is just  astronomical. That is an incredible number and an incredible reach for a small company that came  out of Cambridge and today it has a huge footprint   that all of us in the world are living with.  Anybody that has a mobile phone or anybody   that has automotive applications or  embedded applications like robotics has   Arm in it. You may want to give some perspective  around what did it take for you to build such a   what I call the digital architecture of choice  in today's world that came out of Cambridge, UK.  

I mean I have to sort of pinch myself to  think about just exactly what we've achieved.   When I joined the company right at the  beginning, I was lucky enough to be the 16th   employee and we were in a converted turkey barn  in the middle of the Cambridge countryside and   the technology span out of a British computing  company called Acorn. It was a joint venture with   Apple, Apple wanted this processor to put  inside the Apple Newton – if you remember that,   the first digital personal assistant – and, you  know, we just kind of looked for other applications   where this technology might be useful. We went knocking on doors and trying to   convince people that, if they wanted  to put a processor inside their chip,   it would be a good thing to help try and establish  a global standard, a standard around which we could   build this ecosystem that would drive down the  cost of implementing a processor inside a chip.  We knew from day one there was no way we could do  everything ourselves and so we set out to build a   network of partners who were going to be able to  use the technology and to prosper from using   their technology by building their businesses  around this standard. So the goal really   genuinely from day one was to create a  global standard that would make it very   efficient to put this intelligence inside chips,  and that's turned out to be, you know, incredibly   successful. That's the kind of philosophy that  we've taken in to build a company based  

around partnership and ecosystem. The Samsung team really  appreciates the partnership that Arm has given   and all of our mobile phones and TVs and others  have an Arm architecture in it, like multiple   of them in actually even one phone. So it just  shows the power of the processors and the ecosystem   that the Arm team has driven. You also took this  strategy of partnership and building this ecosystem   for others to join in, which made it even stronger.  But also you had an angle of providing low power,   which I think, if you think about traditional  players – they focus on more on performance than power   and are always driving a huge power footprint, and  I think what you have done in some ways is    idea of this innovation coming from below and  just hitting it on the incumbent not being able to   even deal with the kind of equation that you've  been working on. This is an interesting lesson for how  

this new architecture can come about. Of course, it took  time, but it does show the new architecture can come   at a right time when you are patient and if you  build an ecosystem together with partners. It's an interesting   discussion that I'd like to get your perspective on. At the time that the company was formed,  

they said that essentially the process technology  was really improving rapidly   and it enabled computing to evolve in  in two different directions at once.   There was the 'how much performance can I squeeze  out of these transistors that i put into a chip'   and my only constraint is how much electricity  can I get from a wall socket, and that was one   way to go. And then there was this idea that, if you can eliminate as much power as possible,   then you open the door to many many  devices which could then run on batteries.   And that was the angle that we took, so we  always looked for how can we simplify the design,   how can we really analyze the design to look  at where the power is being consumed, and use   smart circuit techniques, smart architectural  techniques to make the processors as energy-   efficient as we possibly can. So every time when  we're looking at new design we're thinking 'what  

is the trade-off?', 'how much power have I got?' and 'how  much performance can I deliver with the technology   within that power budget?'. And that's  how we think about the design process.   And then, you know, the way our business model  evolved, it was about licensing to others. So we   really had to think about our success coming when  that licensee of ours actually got into production.   So, as well as our own technology, we were thinking  about who else do we need to work with, what about   the companies that provide EDA software, the  software that's used to actually design a chip.   Let's go and work with them, let's go and think  about how quickly a chip designer can get through   the design process using our IP and their tools,  let's think about the end manufacturing technology.  

So we're always thinking about energy efficiency.  That's important in anything that you're   carrying around in your pocket of course, but it's  actually important everywhere. And then how can we   optimize across an ecosystem to produce the best result. What makes it unique is the fact that, unlike   other players, you're not just making products,  but you are making and building an ecosystem, so   that the customer, the end-customer can be successful  by providing the better tools and better alignment   to the manufacturing skill all those type of  things it's incredible what you have done to   accelerate the development of new electronics by  not being able to design these very traditional   custom compute engines and then you of course  add other blocks like graphics and networking   as well as interface logics and all those  different blocks that people need to build the   chips they made it easier for them to do it so i  think i think this whole idea of creating digital   architecture is really a modular architecture and  i think i'm in a way pioneered it from traditional   compute perspective and you made a migration  to much a um a next-generation approach and the   question i guess i have is now what's next coming  given that some of the announcement that was made   by nvidia that wants to have um so i'd like to get  your perspective of what is your view of synergy   why do you think nvidia will benefit uh by working  with um to take care of and provide benefits to   your customers the applications for computing  have grown and grown and grown over the years the   ability to put so much intelligence into a piece  of silicon for such little cost um has just uh   driven an explosion of the number of end devices  and of course we live our lives today you know   with mobile devices with just all this electronics  that we take for granted but that the and whilst   we've done a lot of efficiency and and advances  in the compute engines that are in these uh these   silicon chips um that the nature of that computing  has been fairly constant over the years they're   it's running conventional code that's been  taught in universities for years and years   there is a transition coming though over the  last little while we've seen massive improvements   in the field of artificial intelligence and  we look at the next generation of computing   being a combination of classical computing and ai  driven computing where data is being used to make   decisions and being used to make decisions in  an autonomous way through algorithms that are   effectively using machine learning and ai  techniques to extract information and look for   patterns within that data so i think the nature  of computing is changing we've been uh studying   that we've been uh expanding our own products  to um provide machine learning accelerators to   provide software frameworks to allow applications  to be built and nvidia have been been driving that   in you know to huge success of thinking about  very high performance applications for ai   now we both companies we see ai just going  everywhere being as pervasive in the future   as conventional computing is today and we believe  that if we take the strengths of arm the ability   to uh create pervasive technology and to be  able to create ecosystems around what we do   we take that and we add and we combine it with the  strengths of nvidia and the all the work they've   done on really high performance ai compute  the work we've done on the the software stack   actually enable applications to be built we put  that together uh we think that creates a really   compelling future in delivering how we deliver  this next generation of computers so that's what's   behind the acquisition given the importance of  um and its architecture for really number of   devices that are based on what we call digital  architecture uh and i think you mentioned 190   billion soon by the time we are done with  our interview maybe it'll be 191 billion   it's accelerating but um you know the safety and  security is something that people are really a a   worried about as the architectural proliferates  and reaching the mass number of devices   so i want to just get your perspective around  what are you what is i'm doing about the security   yeah it's a huge subject for us and it and it's  uh it's a sort of um horizontal um technology   that cuts across absolutely everything that we're  doing to your point you can you can build like   incredibly secure systems and you find it's the  smallest cheapest thing that created the entry   point into the network so specifically around that  we've been looking at how people build uh secure   iot devices for uh some time now we we introduced  uh what we call the platform security architecture   psa which is a set of guidelines for how people  build very small very low cost um iot devices but   have a an architecture that then enables secure  software to be built on top of them so go by   defining the kind of building blocks that need to  be within the chip um how to partition memory up   how to keep encryption keys away uh from from main  memory how to uh partition algorithms uh how to   build a secure operating system that's really  important because in a world of uh a trillion   connected things a world of billions and billions  of iot devices it only takes one for the bad actor   to be able to get into the network and we want  to do everything we can to try and prevent that   we've also been thinking about how software on  those devices get managed through its lifetime   any computer that you might use you can have  operating system updates pushed to it if there's   a security issue we need that to be in place  for absolutely everything that's connected to   the network that's a that's a huge piece of work  that's uh that is ahead of us for for the industry   and then as you say uh security needs to go  everywhere we're thinking about how big cpus   are sitting cloud data centers how are the the  applications within them separated from each other   how does the hardware help support that um how do  you contain um something if there is uh if some   application does get penetrated how do you contain  it so it can't get into the rest of the system   it's driving changes in the base architecture  it's driving changes in the way that software   is written across all of these devices and again  we're taking a very partnership based approach   uh to try and help address these problems working  with people um thinking about what solution is   going to be needed not just tomorrow but five ten  years from now and thinking through the changes in   the architecture that are going to be required to  deliver that so it's a huge piece of work but one   that's crucially important for us as you know when  i was in the boat mobile phone was a major driver   and it was in two ways the volume was going  up and you know it was hitting one and a half   billion units per per per year and then they were  shipping multiple arm processor per phones so this   multiplier effect was really great for arms growth  in during late you know earl late i guess the   2000s and early 2010 era but now with the mobile  phone a bit saturating because everybody has a   phone uh what do you see the next big growth form  going forward yeah i mean the growth of mobile   has been phenomenal for for the company and we've  had two ways of it the initial wave of uh digital   mobile phones and and then smartphones um and  and over the years uh you know we've seen arm   technology being used in many other applications  but i think i think we're at a moment now   where um the the growth of semiconductors looks  really really positive for the for the years ahead   and it's a number of markets that are driving that  growth um we're seeing that the deployment of 5g   5g is way more complex than 4g so it's driving  the adoption of more um processing within all   the chips that build up the 5g network and then on  top of that you you have the growth of iot that's   billions of devices they're really small but  they're enormously high volume devices that's a   driver we have autonomous vehicles highly complex  you know data centers on wheels that need to be driven pardon the pun in a way that makes them  really energy efficient but enables a massive   of compute power to do everything an autonomous  vehicle requires so there are all of these things   going on at once and if this combination of um  ai as a the kind of core element of processing 5g   networks iot this is all happening together uh and  i think is going to drive the growth of the sonar   capture industry for years to come and then the  other place uh the dimension is is the data center   and you know the cloud is is growing like crazy  um and that's another area where through our focus   around energy efficiency we're able to provide  some real benefits there as the cloud expands   um with the performance that we're now delivering  uh we're starting to see more deployments of armed   processes in the cloud as well and that's an  area people you know once upon a time thought   we'd never be able to address but uh but today  we are so a lot of these applications are very   exciting and actually these are all the footprint  of what's driving the demand for the uh devices as   well as the applications that are making our lives  better right uh technology and compute power can   help improve um our day-to-day living but there  are so many other areas where this deployment of   computing the deployment of iot devices the use  of ai can be used to address some of the hardest   problems that the world's going to face over the  coming years and things like climate change things   like growing populations things like pollution  you know everything that's uh encapsulated by   the u.n global goals these are areas where  technology can make a big difference and it's it's  

something i know it's an area of focus for you  personally young and there is an area of focus for   samsung and it's something that we're big  believers in that the technology can help   play a role i actually do appreciate um also being  supportive of the initiative that i also started   called extremetech challenge startup companies are  using technology and innovation entrepreneurship   to solve the sustainability issues and  i'm hoping that all of us in corporations   they can be able to work with startups to  accelerate these innovations so i'm going to   continue to support and push extremetech challenge  and i also know that you have a 2030 goals that   you've been pushing as a part of the same agenda  in a different scale in the time frame yeah i i   think that the key thing about about what we're  doing in various different ways is um that this   evolution of compute power is it's been incredible  it's it's giving us these you know mobile phones   that we can play games on and do cool stuff but  the the the potential impact goes way beyond that   and what i've seen through arms history is um  if you continue to drive down the cost if you   drive up the performance if you make it easy  for people to get access to these technologies   then what you do is is kind of spark their own  imagination you you enable people to come up with   the ideas that you can have possibly have thought  of no matter how many smart people you've got   within your company because you've got these kind  of business uh objectives that you have to deliver   on but if you can put all those tools in the hands  of other people then they come up and they create   great ideas and through string tech challenging  through what we're doing in 2030 vision   um it helps uh highlight that and and with these  these technologies that we put in people's hands   i'm really optimistic that we're going to  see great solutions coming around and that do   solve some of the world's biggest problems yeah  we need to partner power up and solve some legal   problems and make some impacts right i'm so happy  to hear that i'm so glad also we can be a partner   in multiple levels including this particular  agenda which i believe is really an important one   well simon i really appreciate your time today  we had a very you know variety of discussions   but we have a long history and i believe we can  really work together and i think fundamentally um   really bring the partnership into the ecosystem  and that's what made it possible in my view   well thanks for sharing i really appreciate that  well thanks a lot you know a lot of questions

2020-12-22

Show video