1 00:00:00,000 --> 00:00:02,560 Speaker 1: Bloomberg's Tom mackenzie got a chance to catch up with 2 00:00:02,680 --> 00:00:06,440 Speaker 1: Armed Ceolready Hass. The pairs spoke about the company's future 3 00:00:06,440 --> 00:00:10,080 Speaker 1: and the generative AI revolution, and how Hass's role has 4 00:00:10,280 --> 00:00:12,760 Speaker 1: changed since Arm's public listing night. 5 00:00:12,840 --> 00:00:14,319 Speaker 2: Thank you so much for your time. 6 00:00:14,520 --> 00:00:17,160 Speaker 3: I look back on this year looking from the outside 7 00:00:17,239 --> 00:00:19,960 Speaker 3: in at ARM, and it seems like this is an 8 00:00:20,040 --> 00:00:25,280 Speaker 3: historic year for ARM. And you've navigated many different moments 9 00:00:25,280 --> 00:00:28,480 Speaker 3: in this story, being previously at a Nvidia, and then 10 00:00:28,520 --> 00:00:30,160 Speaker 3: of course ARM when it was privately held, and then 11 00:00:30,200 --> 00:00:32,599 Speaker 3: of course the IPO. This is an historic year for ARM. 12 00:00:32,640 --> 00:00:36,479 Speaker 3: How do you look back on this year for the company? 13 00:00:36,520 --> 00:00:37,880 Speaker 4: An amazing year? Right? 14 00:00:38,600 --> 00:00:41,519 Speaker 5: We had the Nvidia transaction which ended in twenty twenty two. 15 00:00:42,120 --> 00:00:43,720 Speaker 5: We thought we were going to go public in twenty 16 00:00:43,760 --> 00:00:46,879 Speaker 5: twenty two. The markets thought otherwise, so twenty twenty three 17 00:00:46,960 --> 00:00:49,800 Speaker 5: ended up being the year and just in the phenomenal 18 00:00:49,840 --> 00:00:52,640 Speaker 5: you know, achievement. We are so proud of the history 19 00:00:52,640 --> 00:00:55,640 Speaker 5: of the company, and I think the IPO represents not 20 00:00:55,720 --> 00:00:58,800 Speaker 5: early a fantastic milestone for the employees of today, but 21 00:00:58,920 --> 00:01:01,680 Speaker 5: all the employees that got us here. So a truly 22 00:01:01,720 --> 00:01:04,920 Speaker 5: momentous year, both for the company. For myself personally, I 23 00:01:04,920 --> 00:01:08,000 Speaker 5: also became a grandfather two months prior to the IPO 24 00:01:08,120 --> 00:01:11,679 Speaker 5: to the day, so September fourteenth was the IPO. My 25 00:01:11,720 --> 00:01:15,200 Speaker 5: granddaughter was born on July fourteenth, so it it's made 26 00:01:15,200 --> 00:01:16,040 Speaker 5: for quite a year. 27 00:01:16,040 --> 00:01:18,960 Speaker 3: Well congratulations, Yeah, very very big deal, on a very 28 00:01:19,000 --> 00:01:21,039 Speaker 3: big year indeed on a personal level, and of course 29 00:01:21,040 --> 00:01:23,160 Speaker 3: for the company. And you'll role here, how do you 30 00:01:23,160 --> 00:01:26,280 Speaker 3: think it's changed then? Your role as CEO, going from 31 00:01:26,560 --> 00:01:30,120 Speaker 3: privately held to publicly listed ARM, how has it changed 32 00:01:30,200 --> 00:01:30,800 Speaker 3: your role? 33 00:01:31,319 --> 00:01:35,360 Speaker 5: We've been public now for a few months, so I 34 00:01:35,360 --> 00:01:37,200 Speaker 5: would say I feel like the eyes are upon me 35 00:01:37,319 --> 00:01:40,840 Speaker 5: a little bit more than they were prior. But on 36 00:01:40,880 --> 00:01:43,160 Speaker 5: the other hand, I try to think about it in 37 00:01:43,240 --> 00:01:47,080 Speaker 5: terms of the future of ARM is really not measured 38 00:01:47,080 --> 00:01:48,400 Speaker 5: in what we are today. It's what we're going to 39 00:01:48,400 --> 00:01:50,240 Speaker 5: be in a few years. I spend most of my 40 00:01:50,320 --> 00:01:53,320 Speaker 5: time thinking about where Arm's going to be three years 41 00:01:53,320 --> 00:01:56,600 Speaker 5: from now, five years from now. So on one level, yes, 42 00:01:56,640 --> 00:01:58,560 Speaker 5: it's changed because the eyes are on us. We're a 43 00:01:58,560 --> 00:02:00,680 Speaker 5: public company, we have to meet our at spectations, we 44 00:02:00,720 --> 00:02:03,240 Speaker 5: have to do exactly what we say we're going to do. 45 00:02:03,840 --> 00:02:05,920 Speaker 5: But on the other hand, I'd like to think maybe 46 00:02:05,960 --> 00:02:09,280 Speaker 5: not so much change, because I'm thinking about five years away, 47 00:02:09,400 --> 00:02:11,799 Speaker 5: just like I was before the IPO, same way I'm 48 00:02:11,800 --> 00:02:12,600 Speaker 5: thinking about it today. 49 00:02:12,639 --> 00:02:15,880 Speaker 3: And how do you balance the demands of investors in 50 00:02:15,919 --> 00:02:18,840 Speaker 3: the company now, those new investors and the legacy investors, 51 00:02:18,840 --> 00:02:23,680 Speaker 3: and of course massiyoshi'san SoftBank, being the big force, gravitational fool, 52 00:02:23,720 --> 00:02:26,079 Speaker 3: So how do you balance those demands. 53 00:02:26,480 --> 00:02:28,640 Speaker 5: So ARM is really a company that is hard to 54 00:02:28,639 --> 00:02:32,160 Speaker 5: look at quarter to quarter because the technology we're working 55 00:02:32,160 --> 00:02:35,240 Speaker 5: on today is really technology that's going to end up 56 00:02:35,280 --> 00:02:36,360 Speaker 5: in products. 57 00:02:36,000 --> 00:02:37,680 Speaker 4: Three years from now, four years from now. 58 00:02:37,720 --> 00:02:40,680 Speaker 5: So while the financial results are important, they're really the 59 00:02:40,720 --> 00:02:43,160 Speaker 5: results of strategies we put in place a number of 60 00:02:43,200 --> 00:02:45,440 Speaker 5: years ago, which we're quite successful. This is when we 61 00:02:45,960 --> 00:02:48,920 Speaker 5: created our newoverse product line for the servers, we created 62 00:02:48,919 --> 00:02:51,880 Speaker 5: our automotive product lines, and you're seeing those things really 63 00:02:51,919 --> 00:02:54,720 Speaker 5: come to beir So for us, what we try to 64 00:02:54,720 --> 00:02:57,000 Speaker 5: do is just make sure that the investors understand that 65 00:02:57,040 --> 00:03:00,160 Speaker 5: while the near term results of your important, it's really 66 00:03:00,200 --> 00:03:01,640 Speaker 5: also important to really think about. 67 00:03:01,440 --> 00:03:03,120 Speaker 4: The long term and where we're going directionally. 68 00:03:03,240 --> 00:03:05,800 Speaker 3: Do you think there's still a process of learning from 69 00:03:06,200 --> 00:03:08,919 Speaker 3: four investors to kind of catch up with that view 70 00:03:09,320 --> 00:03:12,240 Speaker 3: that you're projecting out. They're looking at the need it's 71 00:03:12,520 --> 00:03:16,480 Speaker 3: quarter by quarter and you're looking out ahead. You want 72 00:03:16,480 --> 00:03:17,359 Speaker 3: to close that gap. 73 00:03:17,680 --> 00:03:18,720 Speaker 4: Yeah, I think so. 74 00:03:19,760 --> 00:03:21,799 Speaker 5: ARM was a public company in twenty sixteen, so we're 75 00:03:22,120 --> 00:03:24,280 Speaker 5: not new to the markets. But that was a long 76 00:03:24,320 --> 00:03:27,000 Speaker 5: time ago. That was seven years and we were not 77 00:03:27,080 --> 00:03:29,960 Speaker 5: a public company. We didn't talk to investors. So there 78 00:03:30,000 --> 00:03:32,320 Speaker 5: is a bit of re education about who ARM is, 79 00:03:32,400 --> 00:03:35,760 Speaker 5: how our business works, and our business has changed a 80 00:03:35,800 --> 00:03:39,240 Speaker 5: bit than we were in twenty sixteen. Our royalties are 81 00:03:39,240 --> 00:03:42,240 Speaker 5: a much larger percentage of our business, We're much more diversified, 82 00:03:42,280 --> 00:03:45,440 Speaker 5: we're in broader markets. So it's continuously sort of educating 83 00:03:45,440 --> 00:03:48,360 Speaker 5: the investors about who we are, where we're going, and 84 00:03:48,360 --> 00:03:49,920 Speaker 5: what's changed since they las knew us. 85 00:03:49,960 --> 00:03:52,040 Speaker 3: And look, ARM was set up in the nineteen nineties. 86 00:03:52,080 --> 00:03:55,960 Speaker 3: You've been the building blocks of our digital globe, digital world, 87 00:03:56,040 --> 00:03:59,320 Speaker 3: digital economy since then and now at this point where 88 00:03:59,320 --> 00:04:02,520 Speaker 3: we're focused on generative AI and that is reshaping this 89 00:04:02,600 --> 00:04:06,040 Speaker 3: digital ecosystem that we now live in or that we 90 00:04:06,080 --> 00:04:09,760 Speaker 3: will be living on. What role do you see ARM 91 00:04:09,880 --> 00:04:14,160 Speaker 3: playing in that transformed digital space in the years ahead? 92 00:04:14,680 --> 00:04:17,320 Speaker 5: Yeah, So ARM in the thirty plus years has just 93 00:04:17,360 --> 00:04:21,000 Speaker 5: been foundational to computing in a way that no other 94 00:04:21,040 --> 00:04:24,200 Speaker 5: computer architecture before it, and I dare say after it 95 00:04:24,240 --> 00:04:28,000 Speaker 5: has been seventy percent of the world's population touches ARM 96 00:04:28,040 --> 00:04:31,960 Speaker 5: today in some way. We were well known for smartphones 97 00:04:32,880 --> 00:04:35,719 Speaker 5: and before that non smartphones, feature phones, if you will. 98 00:04:36,240 --> 00:04:38,919 Speaker 5: The ARM of today is not only much more diversified 99 00:04:38,960 --> 00:04:42,919 Speaker 5: around as I mentioned before, data centers, automotive, IoT, but 100 00:04:43,000 --> 00:04:46,160 Speaker 5: now as AI and AI workloads are finding their way 101 00:04:46,160 --> 00:04:50,400 Speaker 5: into every single application, whether it's your thermostat or your 102 00:04:50,480 --> 00:04:53,400 Speaker 5: data center, ARM will be there. So I think for 103 00:04:53,560 --> 00:04:55,599 Speaker 5: us in the upcoming years and decades, ARM will be 104 00:04:55,640 --> 00:04:57,360 Speaker 5: foundational to everything going on with AI. 105 00:04:57,640 --> 00:05:00,880 Speaker 3: So you can assure investors that you will. You're essential 106 00:05:01,000 --> 00:05:03,120 Speaker 3: right now to this digital world that we live in. 107 00:05:03,400 --> 00:05:07,880 Speaker 3: You will be essential in that generative AI reshaped world 108 00:05:08,080 --> 00:05:08,960 Speaker 3: in the years ahead. 109 00:05:09,000 --> 00:05:10,320 Speaker 2: You will continue to play that as. 110 00:05:10,200 --> 00:05:13,839 Speaker 5: Such, you can't run AI without ARM. It's foundational and 111 00:05:13,960 --> 00:05:16,720 Speaker 5: AI is going to find its way into every single 112 00:05:17,120 --> 00:05:20,080 Speaker 5: electronics device that we use. Again, whether it's the smallest 113 00:05:20,080 --> 00:05:22,680 Speaker 5: device in your home or the largest data center that 114 00:05:22,760 --> 00:05:24,440 Speaker 5: sits out in the out in the wild. 115 00:05:24,600 --> 00:05:27,320 Speaker 3: You used to work at video of course, in Video 116 00:05:27,360 --> 00:05:28,640 Speaker 3: have backed ARM. 117 00:05:28,720 --> 00:05:29,320 Speaker 2: The IPO. 118 00:05:29,839 --> 00:05:34,000 Speaker 3: Jensen Kwang, the CEO, has been very optimistic and supportive 119 00:05:34,160 --> 00:05:36,479 Speaker 3: of ARM. Have you been surprised at how successful in 120 00:05:36,560 --> 00:05:40,560 Speaker 3: VideA have been at carving out that niche around generative 121 00:05:40,600 --> 00:05:43,800 Speaker 3: AI GPUs and chips. Has that come as a surprise 122 00:05:43,880 --> 00:05:44,680 Speaker 3: or did you see it coming? 123 00:05:45,160 --> 00:05:48,839 Speaker 5: Nothing Jensen does surprises me. He's been at this for 124 00:05:48,880 --> 00:05:51,760 Speaker 5: a long time. Around accelerated computing. But the whole notion 125 00:05:51,800 --> 00:05:55,320 Speaker 5: about accelerating computing is quite interesting because obviously you need 126 00:05:55,360 --> 00:05:58,960 Speaker 5: some computer to accelerate. And one of the things that 127 00:05:59,000 --> 00:06:01,480 Speaker 5: we work very closely within VideA is the idea of 128 00:06:01,520 --> 00:06:04,520 Speaker 5: having the central computer being ARM and the accelerator being 129 00:06:04,560 --> 00:06:06,880 Speaker 5: in VideA. One example of that is a new product 130 00:06:06,880 --> 00:06:10,240 Speaker 5: that they announce called Grace Hopper Grace which uses up 131 00:06:10,279 --> 00:06:13,359 Speaker 5: to one hundred and forty four ARM CPUs inside the 132 00:06:13,360 --> 00:06:17,120 Speaker 5: super chip, then uses an h one hundred GPU as 133 00:06:17,120 --> 00:06:20,800 Speaker 5: the accelerator. So we actually work very closely within VideA, 134 00:06:20,839 --> 00:06:23,880 Speaker 5: and it's a very good example. In the most stringent 135 00:06:24,080 --> 00:06:27,960 Speaker 5: of AI applications. They chose ARM as a compute partner 136 00:06:28,000 --> 00:06:28,679 Speaker 5: to their GPU. 137 00:06:29,640 --> 00:06:32,640 Speaker 3: Microsoft have also come out with Cobol, which is a 138 00:06:32,800 --> 00:06:37,560 Speaker 3: cloud focus chip underpinned by ARM infrastructure. A lot of 139 00:06:37,600 --> 00:06:39,520 Speaker 3: people saying this is very good news for ARM. It 140 00:06:39,560 --> 00:06:41,400 Speaker 3: shows that you can continue to build out your market 141 00:06:41,400 --> 00:06:43,799 Speaker 3: share in data sentences. Can you give us a flavor 142 00:06:43,839 --> 00:06:48,240 Speaker 3: of any other innovations products that may be coming down 143 00:06:48,279 --> 00:06:51,200 Speaker 3: the pipe that may continue to build out your market 144 00:06:51,200 --> 00:06:52,040 Speaker 3: share in that space. 145 00:06:52,200 --> 00:06:55,679 Speaker 5: We saw the start of this with AWS who announced 146 00:06:55,680 --> 00:06:58,000 Speaker 5: their Graviton processors a number of years ago, and now 147 00:06:58,000 --> 00:07:02,000 Speaker 5: they've announced Graviton four, fourth generation product where they are 148 00:07:02,000 --> 00:07:07,480 Speaker 5: integrating armcpus to get an extremely power efficient platform forty 149 00:07:07,480 --> 00:07:11,120 Speaker 5: percent better than the competition in terms of performance per wat. 150 00:07:11,720 --> 00:07:14,080 Speaker 5: Those are the kind of reasons that companies like Microsoft 151 00:07:14,120 --> 00:07:16,400 Speaker 5: with Cobalt, which is a terrific product one hundred and 152 00:07:16,400 --> 00:07:20,760 Speaker 5: twenty eight arm CPUs into a custom sc The benefit 153 00:07:20,840 --> 00:07:23,120 Speaker 5: that the OEMs get by doing the own silicon is 154 00:07:23,120 --> 00:07:27,160 Speaker 5: not only around the forty percent better performance per watt, 155 00:07:27,400 --> 00:07:29,920 Speaker 5: but also building out the entire data center. That is, 156 00:07:30,040 --> 00:07:33,160 Speaker 5: they can customize the rack, they can customize the blade, 157 00:07:33,160 --> 00:07:36,640 Speaker 5: they can customize the footprint. So I think the Microsoft 158 00:07:36,640 --> 00:07:40,080 Speaker 5: announcement and the AWS announcement are are a couple that 159 00:07:40,120 --> 00:07:43,240 Speaker 5: are great proof points. I'm pretty confident you'll see more 160 00:07:43,240 --> 00:07:43,560 Speaker 5: of those. 161 00:07:44,440 --> 00:07:47,560 Speaker 3: Every AI executive that I speak to, whether it's Deep 162 00:07:47,640 --> 00:07:51,400 Speaker 3: Mind or coher or Inflection or others say, the scramble 163 00:07:51,480 --> 00:07:54,400 Speaker 3: for compute is front and center along with the fight 164 00:07:54,440 --> 00:07:57,440 Speaker 3: for talents. How do you see that fight for compute 165 00:07:57,480 --> 00:08:00,520 Speaker 3: going forward? How does that evolve in the years ahead. 166 00:08:00,640 --> 00:08:03,720 Speaker 3: Do we get to see more products coming onto the 167 00:08:03,760 --> 00:08:06,760 Speaker 3: marketplace that's going to address some of that tightness and 168 00:08:06,800 --> 00:08:08,240 Speaker 3: some of the demand for computer. 169 00:08:08,200 --> 00:08:10,720 Speaker 5: I think right now what we're seeing with these large 170 00:08:10,760 --> 00:08:14,600 Speaker 5: language models, for example, that are very very large in 171 00:08:14,640 --> 00:08:17,120 Speaker 5: terms of their data capacity, they take lots of computer 172 00:08:17,200 --> 00:08:20,200 Speaker 5: cycles to train, is why you're seeing such a demand 173 00:08:20,200 --> 00:08:23,400 Speaker 5: for compute resources. And that's not just the accelerator, that's 174 00:08:23,440 --> 00:08:27,160 Speaker 5: the core computer aka what ARM does. So I think 175 00:08:27,200 --> 00:08:29,520 Speaker 5: that trend is going to just continue over the next 176 00:08:29,600 --> 00:08:33,320 Speaker 5: number of years because as sophisticated as the models are today, 177 00:08:33,400 --> 00:08:37,320 Speaker 5: whether you run chat Chept or coming up with Gemini 178 00:08:37,559 --> 00:08:40,720 Speaker 5: or what inflection AI is working on, they're terrific products, 179 00:08:41,320 --> 00:08:43,440 Speaker 5: but they're still limited in terms of how quick they 180 00:08:43,440 --> 00:08:46,480 Speaker 5: can learn, how quickly they can adapt. That needs more 181 00:08:46,520 --> 00:08:49,720 Speaker 5: and more training, which means more and more compute, and 182 00:08:49,760 --> 00:08:51,480 Speaker 5: then you have to run the inference that is the 183 00:08:51,520 --> 00:08:54,640 Speaker 5: actual use of those models, which requires even more compute. 184 00:08:54,920 --> 00:08:57,040 Speaker 5: So I do think for the next number of years 185 00:08:57,080 --> 00:09:00,320 Speaker 5: we are going to see a increasing demand are more 186 00:09:00,320 --> 00:09:03,760 Speaker 5: and more compute now. I think that's also going to 187 00:09:03,840 --> 00:09:06,520 Speaker 5: drive is a high degree of need for power efficiency 188 00:09:07,120 --> 00:09:11,720 Speaker 5: because these data centers require hundreds of watts megawatts up 189 00:09:11,720 --> 00:09:14,679 Speaker 5: to the gigawatts of type of energy. We don't produce 190 00:09:14,720 --> 00:09:18,439 Speaker 5: that much energy as a planet. We are having limitations 191 00:09:18,480 --> 00:09:23,880 Speaker 5: relative to fossil fuels, limitations sustainable, so a rush to 192 00:09:24,000 --> 00:09:27,320 Speaker 5: get to power efficiency around these compute models is going 193 00:09:27,360 --> 00:09:30,640 Speaker 5: to be very very significant. That's great for ARM because 194 00:09:30,720 --> 00:09:33,920 Speaker 5: the DNA of our company is around power efficiency. We 195 00:09:33,920 --> 00:09:36,520 Speaker 5: were born building our first CPUs off of a battery, 196 00:09:37,000 --> 00:09:39,120 Speaker 5: So for us, I think we're in a very very 197 00:09:39,120 --> 00:09:40,800 Speaker 5: good place to help with that transition. 198 00:09:40,840 --> 00:09:43,280 Speaker 3: What do you think it does for pricing going forward? 199 00:09:43,280 --> 00:09:46,960 Speaker 3: Prices very elevated already around compute. Does that inflation around 200 00:09:47,000 --> 00:09:50,160 Speaker 3: prices compute prices for generative ali does that continue as 201 00:09:50,200 --> 00:09:52,480 Speaker 3: a trend that we starting to peak in terms of price. 202 00:09:52,640 --> 00:09:54,960 Speaker 5: It's hard to say, because on one hand, people will 203 00:09:54,960 --> 00:09:56,880 Speaker 5: look at it and say well, the pricing is incredibly high. 204 00:09:57,280 --> 00:09:59,760 Speaker 5: On the flip side, it's now delivering capability we've never 205 00:09:59,760 --> 00:10:04,000 Speaker 5: seen before, and as that capability gets transitioned into real 206 00:10:04,040 --> 00:10:08,160 Speaker 5: productivity gains. For example, for ARM, one of the things 207 00:10:08,200 --> 00:10:09,760 Speaker 5: that we would love to see is how can we 208 00:10:09,800 --> 00:10:12,520 Speaker 5: shrink the amount of time it takes to build our products. 209 00:10:13,000 --> 00:10:15,600 Speaker 5: For us to design a product, for example, we spend 210 00:10:15,640 --> 00:10:19,360 Speaker 5: more time verifying that the product we design is correct 211 00:10:19,400 --> 00:10:22,840 Speaker 5: and actually designing it. That verification work can really benefit 212 00:10:22,960 --> 00:10:26,160 Speaker 5: from large AI models. So one could argue, well, the 213 00:10:26,200 --> 00:10:28,480 Speaker 5: pricing is kind of high, but if at the end 214 00:10:28,559 --> 00:10:30,439 Speaker 5: it actually saves me a lot of time and cost 215 00:10:30,440 --> 00:10:33,160 Speaker 5: in product development, it's not so high, so I think 216 00:10:33,240 --> 00:10:34,520 Speaker 5: it'll work itself out over time. 217 00:10:34,600 --> 00:10:36,640 Speaker 3: Well, you touch on something that's really interesting, which is 218 00:10:37,040 --> 00:10:40,920 Speaker 3: obviously you're providing components, chips, designs, the blueprints they're going 219 00:10:41,000 --> 00:10:44,680 Speaker 3: to fuel this AI revolution, but you're using AI in 220 00:10:44,760 --> 00:10:46,960 Speaker 3: house as well, and so I wonder in terms of 221 00:10:47,200 --> 00:10:50,600 Speaker 3: verification you can envisage a role where AI helps with 222 00:10:50,679 --> 00:10:53,280 Speaker 3: that process and cuts down the time. Talk to us 223 00:10:53,280 --> 00:10:57,840 Speaker 3: about talent and the focus on engineering talent at ARM, 224 00:10:57,880 --> 00:11:00,800 Speaker 3: are you still able to attract that talent and talon 225 00:11:01,000 --> 00:11:01,480 Speaker 3: at this point. 226 00:11:01,559 --> 00:11:04,080 Speaker 5: Yeah, So back to the productivity gains, you know, right 227 00:11:04,160 --> 00:11:05,880 Speaker 5: right now it's very early stages. 228 00:11:06,240 --> 00:11:07,560 Speaker 4: We're experimenting with it. 229 00:11:08,120 --> 00:11:09,960 Speaker 5: One of the things that needs to happen for this 230 00:11:10,040 --> 00:11:13,920 Speaker 5: to really be unleashed is privacy of data and integrity 231 00:11:13,960 --> 00:11:17,920 Speaker 5: of data. We can't have our engineers innovating on models 232 00:11:17,920 --> 00:11:20,120 Speaker 5: that all that data goes out to the World Wide 233 00:11:20,120 --> 00:11:22,160 Speaker 5: Web and then the public can use it. So right 234 00:11:22,200 --> 00:11:25,480 Speaker 5: now we're limited a bit by the capability of the enterprise tools, 235 00:11:25,520 --> 00:11:27,720 Speaker 5: but that will get sorted out over time, and then 236 00:11:27,720 --> 00:11:30,320 Speaker 5: I think you see a big productivity gain. So then 237 00:11:30,360 --> 00:11:32,680 Speaker 5: people say, well, gosh, isn't that going to eliminate jobs 238 00:11:32,720 --> 00:11:34,120 Speaker 5: or may and you're going to have less and less 239 00:11:34,200 --> 00:11:37,240 Speaker 5: verification engineers. Well not really, because then maybe I can 240 00:11:37,280 --> 00:11:39,959 Speaker 5: see maybe ten times more products, and I can develop 241 00:11:39,960 --> 00:11:42,000 Speaker 5: more and more products that I can put in the 242 00:11:42,040 --> 00:11:45,360 Speaker 5: marketplace because I can move engineers into different phases of work. 243 00:11:45,840 --> 00:11:49,200 Speaker 5: So I think generally speaking, when people talk about AI 244 00:11:49,360 --> 00:11:52,480 Speaker 5: and productivity gains and moving jobs around, I think for 245 00:11:52,600 --> 00:11:55,079 Speaker 5: our world, there's going to be no shortage jobs. Then 246 00:11:55,120 --> 00:11:57,760 Speaker 5: to your question, can we can we find the talent? 247 00:11:58,800 --> 00:12:01,679 Speaker 5: It's always challenging we have a lot of people here 248 00:12:01,720 --> 00:12:03,959 Speaker 5: in the UK. That's where the majority of our people 249 00:12:03,960 --> 00:12:07,680 Speaker 5: are in Cambridge. We have locations all across Europe. We 250 00:12:07,720 --> 00:12:11,280 Speaker 5: have locations in the United States, also in India and 251 00:12:11,320 --> 00:12:14,120 Speaker 5: parts of Asia. The fight for talent for what we 252 00:12:14,160 --> 00:12:18,360 Speaker 5: do is quite fierce. Thankfully, people like working at ARM. 253 00:12:18,440 --> 00:12:20,880 Speaker 5: The company is doing well, so we're able to retain 254 00:12:20,920 --> 00:12:24,040 Speaker 5: our talent, but finding new talent is always a challenge. 255 00:12:24,880 --> 00:12:28,360 Speaker 3: AI engineer is better than human engineers, or could they be? 256 00:12:29,440 --> 00:12:31,480 Speaker 5: Today we were having a chat around this where the 257 00:12:31,679 --> 00:12:33,760 Speaker 5: number folks at lunch were one of the folks at 258 00:12:33,840 --> 00:12:36,680 Speaker 5: lunch describe them to an intern in other words, the 259 00:12:36,679 --> 00:12:39,360 Speaker 5: work is decent, but I need to check it. And 260 00:12:39,440 --> 00:12:41,760 Speaker 5: that's I think where we are today. But I think 261 00:12:41,800 --> 00:12:44,760 Speaker 5: over time they will become very very good assistance. 262 00:12:45,320 --> 00:12:48,160 Speaker 3: Okay, so we're still a ways away from that. You've 263 00:12:48,160 --> 00:12:50,280 Speaker 3: taught us about being in an R and D supercycle. 264 00:12:51,040 --> 00:12:54,040 Speaker 3: How long does that supercycle continue? 265 00:12:54,920 --> 00:12:58,160 Speaker 5: I think for where we are with AI, and AI 266 00:12:58,280 --> 00:13:01,400 Speaker 5: to some level is not new. When you look at 267 00:13:01,720 --> 00:13:04,920 Speaker 5: on a website and you do Google Translate, that's AI 268 00:13:05,360 --> 00:13:08,520 Speaker 5: on some level. But what's happened, particularly with the chat 269 00:13:08,920 --> 00:13:13,320 Speaker 5: GPT moment. We've elevated the capability, and now what we 270 00:13:13,400 --> 00:13:18,000 Speaker 5: see is a real opportunity to get to your point 271 00:13:18,360 --> 00:13:22,040 Speaker 5: about can artificial engineers be as good as real engineers. 272 00:13:22,520 --> 00:13:25,800 Speaker 5: We're very early in that area. So I think back 273 00:13:25,800 --> 00:13:28,360 Speaker 5: to your question around the supercycle. I think we are 274 00:13:28,520 --> 00:13:33,080 Speaker 5: in a accelerated investment cycle where these models will get smarter, 275 00:13:33,320 --> 00:13:36,320 Speaker 5: the data sets will get sharper, and the outputs will 276 00:13:36,360 --> 00:13:38,959 Speaker 5: get better, and I think we're very early in that. 277 00:13:39,400 --> 00:13:42,280 Speaker 5: One of the analogies I like to talk about is 278 00:13:42,360 --> 00:13:45,600 Speaker 5: it reminds me a little bit around of the Internet 279 00:13:45,640 --> 00:13:50,600 Speaker 5: of nineteen ninety five, where different browsers, different portals were 280 00:13:50,640 --> 00:13:52,760 Speaker 5: all coming out and the magic of oh my gosh, 281 00:13:52,760 --> 00:13:55,680 Speaker 5: I've got a guy and now I can access information 282 00:13:55,720 --> 00:13:57,520 Speaker 5: around the globe. If you look at the Internet of 283 00:13:57,600 --> 00:13:59,840 Speaker 5: nineteen ninety five to what the Internet of twenty twenty 284 00:13:59,840 --> 00:14:03,520 Speaker 5: five tree is, obviously we've come a long long way. 285 00:14:03,840 --> 00:14:06,959 Speaker 5: I feel like that's where we are with this large 286 00:14:07,000 --> 00:14:09,079 Speaker 5: language model sort of transition on AI. 287 00:14:09,400 --> 00:14:11,640 Speaker 3: You went back to the early nineties, I want to 288 00:14:11,640 --> 00:14:14,680 Speaker 3: go back to twenty seventeen. You diversified the business, you 289 00:14:14,800 --> 00:14:20,880 Speaker 3: broke out these four different divisions, IoT Cloud infrastructure, autonomous vehicles. 290 00:14:21,520 --> 00:14:24,680 Speaker 3: Why did you see that as necessary back in twenty seventeen, 291 00:14:24,680 --> 00:14:27,520 Speaker 3: you were trying to diversify away from smartphones or at 292 00:14:27,560 --> 00:14:28,920 Speaker 3: least make sure you had a product mix. 293 00:14:29,040 --> 00:14:30,240 Speaker 2: Yeah, but why was it essential? 294 00:14:30,280 --> 00:14:30,480 Speaker 4: Then? 295 00:14:30,840 --> 00:14:32,960 Speaker 5: One of the things that we were seeing at that 296 00:14:33,120 --> 00:14:36,480 Speaker 5: time was the hallmarks of what was making are very 297 00:14:36,520 --> 00:14:42,520 Speaker 5: successful in mobile phones, high performance software ecosystem power efficiency 298 00:14:43,280 --> 00:14:46,760 Speaker 5: was becoming table stakes through these other markets. The cloud 299 00:14:46,840 --> 00:14:48,880 Speaker 5: data center, as we talked about before, in terms of 300 00:14:48,920 --> 00:14:52,200 Speaker 5: power efficiency, you just don't have all the energy required. 301 00:14:52,880 --> 00:14:57,600 Speaker 5: Automobiles that were transitioning now to evs or a digitized 302 00:14:57,600 --> 00:15:00,000 Speaker 5: cockpit where your car starts to look like a computer. 303 00:15:00,440 --> 00:15:03,840 Speaker 5: We were seeing again demands for lots of software but 304 00:15:03,920 --> 00:15:07,280 Speaker 5: also power efficiency. So we were very lucky in twenty 305 00:15:07,320 --> 00:15:11,280 Speaker 5: seventeen we were privately held, so as a result, we 306 00:15:11,360 --> 00:15:14,640 Speaker 5: could make some investments into new products that were much 307 00:15:14,680 --> 00:15:18,480 Speaker 5: more purpose built for the end markets that we were 308 00:15:18,480 --> 00:15:22,440 Speaker 5: competing in. So previous to twenty seventeen, we would design 309 00:15:22,520 --> 00:15:26,440 Speaker 5: a mobile processor and then really attempt to squeeze it 310 00:15:26,560 --> 00:15:29,840 Speaker 5: or peanut butter it, if you will, into the server space, 311 00:15:29,960 --> 00:15:32,680 Speaker 5: or into the automotive space, or into different areas of IoT. 312 00:15:33,680 --> 00:15:35,480 Speaker 5: It just wasn't sufficient, and we had a lot of 313 00:15:35,520 --> 00:15:38,520 Speaker 5: questions early on, why isn't what's taking arms so long 314 00:15:38,560 --> 00:15:42,640 Speaker 5: to be successful in the servers. Many different reasons contributed that, 315 00:15:42,680 --> 00:15:44,480 Speaker 5: but one of them was candidly, we just didn't have 316 00:15:44,480 --> 00:15:47,720 Speaker 5: the right products, and we were able to create neoverse. 317 00:15:47,760 --> 00:15:49,000 Speaker 4: We were able to add. 318 00:15:48,880 --> 00:15:53,520 Speaker 5: Custom extensions, things like confidential compute that are necessary for 319 00:15:53,600 --> 00:15:56,400 Speaker 5: building high grade data centers. We could put those in 320 00:15:56,400 --> 00:15:59,600 Speaker 5: the products, and now with neoverse, we have a world 321 00:15:59,640 --> 00:16:02,520 Speaker 5: class competitive entry that you're seeing the benefits of it, 322 00:16:02,560 --> 00:16:06,600 Speaker 5: whether it's the Microsoft, Cobalt or AWS. We had to 323 00:16:06,640 --> 00:16:09,360 Speaker 5: do that in twenty seventeen because it was clear that 324 00:16:09,400 --> 00:16:13,000 Speaker 5: these other markets were demanding it and the selfware ecosystem 325 00:16:13,080 --> 00:16:14,280 Speaker 5: needs to be able to support it. 326 00:16:14,440 --> 00:16:16,920 Speaker 3: What kind of revenue mix would you need to see 327 00:16:17,280 --> 00:16:20,960 Speaker 3: to kind of declare a success in that diversification push. 328 00:16:21,560 --> 00:16:23,600 Speaker 4: I'm going to declare some moticum of success. Now. 329 00:16:24,160 --> 00:16:27,160 Speaker 5: We were over sixty percent revenue in smartphones, maybe even 330 00:16:27,240 --> 00:16:31,600 Speaker 5: higher prior to this change. Smartphones now, while a very 331 00:16:31,680 --> 00:16:35,440 Speaker 5: large market for US, are about forty percent and declining, 332 00:16:35,840 --> 00:16:39,680 Speaker 5: and our largest growth in terms of revenue are coming 333 00:16:39,720 --> 00:16:43,520 Speaker 5: from these new markets. As I mentioned cloud and automotive. 334 00:16:43,800 --> 00:16:45,400 Speaker 2: Which of those is most exciting to you? 335 00:16:45,480 --> 00:16:48,800 Speaker 3: Cloud automotive, which has the greatest prospect in terms of 336 00:16:48,880 --> 00:16:50,400 Speaker 3: driving revenues into the business. 337 00:16:50,840 --> 00:16:51,640 Speaker 4: They're quite different. 338 00:16:52,200 --> 00:16:55,840 Speaker 5: The cycles on automotive are much much longer, so it 339 00:16:55,880 --> 00:16:58,800 Speaker 5: will take a longer amount of time, but there's huge 340 00:16:58,840 --> 00:17:01,400 Speaker 5: potential there. Just when you look at the massive shift 341 00:17:01,800 --> 00:17:06,360 Speaker 5: that's taking place on automobiles, combustion engines to EV those 342 00:17:06,440 --> 00:17:11,040 Speaker 5: require processors. Your car is becoming a computer. The IVI, 343 00:17:11,280 --> 00:17:14,159 Speaker 5: the digital cockpit, that is all moving to something that 344 00:17:14,240 --> 00:17:17,040 Speaker 5: runs on software. All good for ARM. And then these 345 00:17:17,080 --> 00:17:21,040 Speaker 5: electronic control units, your power brakes, your power windows, those 346 00:17:21,080 --> 00:17:23,960 Speaker 5: are all integrating into a common network. That's all moving 347 00:17:23,960 --> 00:17:26,800 Speaker 5: to ARM. So I'm so excited about where the automotive 348 00:17:26,800 --> 00:17:29,480 Speaker 5: market can go. I do think the cloud data center 349 00:17:29,920 --> 00:17:33,240 Speaker 5: is even potentially more exciting from the perspective of when 350 00:17:33,240 --> 00:17:36,880 Speaker 5: you layer in the AI component. We talked about all 351 00:17:36,880 --> 00:17:39,760 Speaker 5: the growth that ARM is seen in the cloud, but 352 00:17:39,880 --> 00:17:43,240 Speaker 5: now as we're seeing the shift towards AI, a product 353 00:17:43,280 --> 00:17:46,840 Speaker 5: like grace Hopper again in Vidia taking their GPU and 354 00:17:46,920 --> 00:17:50,560 Speaker 5: our CPU and creating the world's best super chip for AI. 355 00:17:51,400 --> 00:17:54,520 Speaker 5: That wasn't on the roadmap in twenty seventeen when we started, 356 00:17:54,520 --> 00:17:56,840 Speaker 5: because we weren't thinking about the AI market. So I 357 00:17:56,880 --> 00:17:59,199 Speaker 5: think for us cloud could be a terrific opportunity. 358 00:17:59,280 --> 00:18:01,119 Speaker 3: What is the market sharing cloud that you would be 359 00:18:01,160 --> 00:18:03,720 Speaker 3: targeting in the years ahead? Do you have a number 360 00:18:03,800 --> 00:18:04,359 Speaker 3: in mind? 361 00:18:04,560 --> 00:18:05,040 Speaker 2: Rene? 362 00:18:05,280 --> 00:18:07,040 Speaker 5: You know we're in double digits now, yeah, And I 363 00:18:07,040 --> 00:18:09,280 Speaker 5: always like to say, when you get to double digits 364 00:18:09,320 --> 00:18:12,920 Speaker 5: in our space, you're now You're in the party. Can 365 00:18:12,960 --> 00:18:16,520 Speaker 5: we get to over fifty percent over time? I think 366 00:18:16,560 --> 00:18:20,000 Speaker 5: we can. I'm a very very ambitious guy. I want 367 00:18:20,000 --> 00:18:22,200 Speaker 5: to see ARM do what we did in the data 368 00:18:22,200 --> 00:18:23,919 Speaker 5: center in all the other markets. I think we have 369 00:18:23,960 --> 00:18:24,919 Speaker 5: all the ingredients to do. 370 00:18:25,000 --> 00:18:27,560 Speaker 3: So what is the opportunity What is the potential around 371 00:18:27,600 --> 00:18:30,399 Speaker 3: autonomous driving? We move from L three, L four, L 372 00:18:30,520 --> 00:18:33,760 Speaker 3: five to autonomous, that is an opportunity for ARM as well. 373 00:18:33,760 --> 00:18:36,040 Speaker 3: What is the potential growth to the top line from that? 374 00:18:36,520 --> 00:18:39,000 Speaker 5: I think that again a little bit longer out in time, 375 00:18:39,280 --> 00:18:41,720 Speaker 5: but can be very very significant. We have a lot 376 00:18:41,760 --> 00:18:45,880 Speaker 5: of engagements today already around autonomous We're our AID ass 377 00:18:45,880 --> 00:18:50,840 Speaker 5: share today already is fifty percent and growing. We're ninety 378 00:18:50,880 --> 00:18:54,880 Speaker 5: percent of IBI, which is a digital cockpit, so between 379 00:18:54,960 --> 00:18:57,919 Speaker 5: IVI and aid ASS and particularly as those getting melded together. 380 00:18:58,640 --> 00:19:02,280 Speaker 5: And then again the real strength environment is the software ecosystem, 381 00:19:02,840 --> 00:19:06,359 Speaker 5: software developers, middlewale developers who are writing all this software 382 00:19:06,400 --> 00:19:09,719 Speaker 5: for a common platform. This is what the automobile industry wants. 383 00:19:09,840 --> 00:19:13,919 Speaker 5: So I think the autonomous not only leverage itself interesting 384 00:19:13,960 --> 00:19:16,960 Speaker 5: in terms of automotive, but to other areas such as robotics, 385 00:19:17,080 --> 00:19:18,639 Speaker 5: which I think with AI are going to be a 386 00:19:18,760 --> 00:19:21,280 Speaker 5: very very large market growing forward. I think the robotics 387 00:19:21,320 --> 00:19:26,600 Speaker 5: market in terms of large language models, programming robot in 388 00:19:26,680 --> 00:19:28,760 Speaker 5: English and telling it what to go off and do 389 00:19:29,320 --> 00:19:31,440 Speaker 5: is going to be reality in the next number of years. 390 00:19:31,680 --> 00:19:33,639 Speaker 5: That leverages very heavily. From the work that we do 391 00:19:33,680 --> 00:19:35,120 Speaker 5: around autonomous driving. 392 00:19:35,800 --> 00:19:39,040 Speaker 3: On China geopolitics, you get still about twenty five percent 393 00:19:39,080 --> 00:19:42,040 Speaker 3: of your revenues and on from that market. How do 394 00:19:42,080 --> 00:19:45,359 Speaker 3: you ensure you continue to get market access, protect your 395 00:19:45,359 --> 00:19:49,200 Speaker 3: intellectual property in that market, and don't run a foul 396 00:19:49,240 --> 00:19:52,439 Speaker 3: of those restrictions from the US. It feels and sounds 397 00:19:52,440 --> 00:19:54,480 Speaker 3: like a very fine line to walk. 398 00:19:55,040 --> 00:19:58,520 Speaker 5: Yeah, So I think, like most tech CEOs, we all 399 00:19:58,880 --> 00:20:02,359 Speaker 5: run very fine lines these days. I think the CEOs 400 00:20:02,680 --> 00:20:05,800 Speaker 5: ten years ago did not talk to government officials with 401 00:20:05,880 --> 00:20:09,439 Speaker 5: nearly the frequency that we do today. So we comply 402 00:20:09,560 --> 00:20:11,800 Speaker 5: with all the export controls, whether it's from the US 403 00:20:11,880 --> 00:20:14,240 Speaker 5: or other other parts of the world. China is a 404 00:20:14,280 --> 00:20:16,119 Speaker 5: big market for US, as you said, about twenty to 405 00:20:16,160 --> 00:20:20,040 Speaker 5: twenty five percent today. China looks like the rest of 406 00:20:20,080 --> 00:20:23,040 Speaker 5: the world though in terms of the ecosystems they build, 407 00:20:23,400 --> 00:20:26,000 Speaker 5: the software that runs on the processors and the markets 408 00:20:26,040 --> 00:20:28,560 Speaker 5: that they want to address. So the big markets that 409 00:20:28,560 --> 00:20:31,920 Speaker 5: are going for US in China, unsurprisingly are cloud and automotive. 410 00:20:32,680 --> 00:20:35,600 Speaker 5: Automotive and evs are going very very fast in China. 411 00:20:35,640 --> 00:20:38,480 Speaker 5: You just have to travel to China. I was there 412 00:20:38,520 --> 00:20:40,920 Speaker 5: for the first time in four or five years back 413 00:20:40,960 --> 00:20:44,040 Speaker 5: in the spring. Was amazed how a how many Chinese 414 00:20:44,080 --> 00:20:46,960 Speaker 5: branded automobiles we saw, and then how many were EV. 415 00:20:47,680 --> 00:20:50,560 Speaker 5: Most of those today run on ARM, similar to the 416 00:20:50,600 --> 00:20:53,240 Speaker 5: Cloud Data center. The large players there are the ten 417 00:20:53,320 --> 00:20:56,879 Speaker 5: Cents to Bite Dances, and the reason for that is 418 00:20:56,920 --> 00:21:00,719 Speaker 5: that the software ecosystem that runs in China is very 419 00:21:00,720 --> 00:21:02,639 Speaker 5: similar to what runs the rest of the world, and 420 00:21:02,760 --> 00:21:05,800 Speaker 5: since ARM is so closely linked to what goes inside 421 00:21:06,280 --> 00:21:09,120 Speaker 5: the software ecosystem in terms of what runs on our processors, 422 00:21:09,560 --> 00:21:12,000 Speaker 5: it's natural that our China business would look a lot 423 00:21:12,040 --> 00:21:13,560 Speaker 5: like the rest of the world, which. 424 00:21:13,359 --> 00:21:18,600 Speaker 3: It does when it comes to the I guess the 425 00:21:18,640 --> 00:21:22,040 Speaker 3: question around there's a couple of questions around that picture 426 00:21:22,040 --> 00:21:27,480 Speaker 3: on China. The restrictions that coming through from the US. Clearly, 427 00:21:27,560 --> 00:21:29,720 Speaker 3: does it concern that some of these technologies to end 428 00:21:29,840 --> 00:21:33,720 Speaker 3: up with these kind of duel use properties? How do 429 00:21:33,760 --> 00:21:39,040 Speaker 3: you mitigate future additional restrictions being put in place from 430 00:21:39,160 --> 00:21:40,520 Speaker 3: the US and other jurisdictions. 431 00:21:40,560 --> 00:21:41,640 Speaker 2: How do you think about that? 432 00:21:42,240 --> 00:21:43,960 Speaker 5: But it's a little hard to predict what's coming around 433 00:21:44,000 --> 00:21:47,400 Speaker 5: the corner where we tend to play is in the 434 00:21:47,480 --> 00:21:51,480 Speaker 5: components to build up in a system on chip. Generally speaking, 435 00:21:51,520 --> 00:21:54,400 Speaker 5: the restrictions that get put on ARM are either around 436 00:21:54,440 --> 00:21:56,520 Speaker 5: companies that are on the entity list, and that's rather 437 00:21:56,560 --> 00:21:58,760 Speaker 5: black and white. If the companies on the entity list, 438 00:21:59,480 --> 00:22:01,960 Speaker 5: then the rules are very clear and simple in terms 439 00:22:02,000 --> 00:22:05,000 Speaker 5: of how to comply, in which we do. Now, it's 440 00:22:05,040 --> 00:22:08,200 Speaker 5: interesting in those cases because most of our products are 441 00:22:08,240 --> 00:22:11,200 Speaker 5: not designed in the US, we actually don't get impacted 442 00:22:11,280 --> 00:22:14,640 Speaker 5: quite as much. Since most of our products are non 443 00:22:15,359 --> 00:22:19,040 Speaker 5: non US origin. Now, in the areas around corners where 444 00:22:19,160 --> 00:22:21,919 Speaker 5: we might get impacted. Right now, what we've been seeing 445 00:22:21,960 --> 00:22:27,240 Speaker 5: are things around performance thresholds and memory bandwidth. And again 446 00:22:27,320 --> 00:22:29,679 Speaker 5: in those areas where the restrictions are laid out, we 447 00:22:29,760 --> 00:22:32,679 Speaker 5: will comply. But because we don't build a chip like 448 00:22:32,680 --> 00:22:36,600 Speaker 5: an Nvidia H one hundred or an AMD MI I 449 00:22:36,720 --> 00:22:39,920 Speaker 5: three hundred, where the chip has to hit a certain 450 00:22:39,960 --> 00:22:44,119 Speaker 5: spec we do some of the underlying technology. Our products 451 00:22:44,160 --> 00:22:46,199 Speaker 5: are really not impacted very much. We have a handful 452 00:22:46,240 --> 00:22:46,960 Speaker 5: that are okay. 453 00:22:47,359 --> 00:22:48,840 Speaker 3: I think it's fair to say that there is a 454 00:22:48,840 --> 00:22:52,560 Speaker 3: bit of confusion from some as to the relationship between 455 00:22:53,560 --> 00:22:56,760 Speaker 3: ARM China and soft Bank, So can you just explain 456 00:22:56,840 --> 00:22:59,040 Speaker 3: for us how that relationship looks. 457 00:22:59,320 --> 00:22:59,600 Speaker 4: Sure. 458 00:23:00,280 --> 00:23:03,160 Speaker 5: Armed China is a joint venture that is forty nine 459 00:23:03,160 --> 00:23:06,919 Speaker 5: percent owned by ARM slash SoftBank and fifty one percent 460 00:23:07,000 --> 00:23:09,800 Speaker 5: owned by China investors, and there are a number of 461 00:23:09,840 --> 00:23:13,080 Speaker 5: just private investors are who are ownership in that stake. 462 00:23:14,200 --> 00:23:17,840 Speaker 5: That joint venture has two primary objectives. One is, and 463 00:23:17,840 --> 00:23:20,280 Speaker 5: that's the area that we care most about, is the 464 00:23:20,359 --> 00:23:24,399 Speaker 5: reseller of ARMED technology into China. So Armed China is 465 00:23:24,880 --> 00:23:28,360 Speaker 5: the exclusive distributor, if you will, of all the technology 466 00:23:28,440 --> 00:23:31,720 Speaker 5: that ARM has to companies that are headquartered in the 467 00:23:31,760 --> 00:23:34,639 Speaker 5: PRC and in that model it works very closely the 468 00:23:34,640 --> 00:23:38,199 Speaker 5: way a distributor would look. They sell the product, We 469 00:23:38,359 --> 00:23:40,399 Speaker 5: deliver the product to the end customer. So if no 470 00:23:40,480 --> 00:23:43,000 Speaker 5: IP goes into the joint venture, it actually goes from 471 00:23:43,600 --> 00:23:47,800 Speaker 5: arm UK into the China partner and then in a 472 00:23:47,840 --> 00:23:52,200 Speaker 5: classic distributed relationship, they will pay us a certain portion 473 00:23:52,280 --> 00:23:54,919 Speaker 5: of the of the dollars back. That's very simply in 474 00:23:54,960 --> 00:23:57,320 Speaker 5: terms of how that portion of the business works. That's 475 00:23:57,320 --> 00:23:59,480 Speaker 5: the part that's about twenty to twenty five percent of 476 00:23:59,480 --> 00:24:02,040 Speaker 5: our revenue. That's the part of the business that we 477 00:24:02,119 --> 00:24:05,480 Speaker 5: and I care about most. The other charter of ARMED 478 00:24:05,560 --> 00:24:11,119 Speaker 5: China is to be essentially an independent entity to develop 479 00:24:11,119 --> 00:24:14,320 Speaker 5: products for the China market. Now, they can develop products 480 00:24:14,320 --> 00:24:16,880 Speaker 5: that are based on ARM or they can develop products 481 00:24:16,920 --> 00:24:20,399 Speaker 5: that are independent. They can do their own controllers or 482 00:24:20,440 --> 00:24:23,760 Speaker 5: display et cetera, et cetera. For that portion of the business, 483 00:24:23,760 --> 00:24:25,680 Speaker 5: we really have nothing to do with it. They are 484 00:24:26,240 --> 00:24:28,000 Speaker 5: free and clear to develop what they want. We have 485 00:24:28,040 --> 00:24:31,280 Speaker 5: some guidelines in terms of product roadmap. I happen to 486 00:24:31,280 --> 00:24:34,240 Speaker 5: sit on the board of that joint venture, so I 487 00:24:34,320 --> 00:24:36,800 Speaker 5: wear two hats. I see what's going on with that 488 00:24:36,880 --> 00:24:39,840 Speaker 5: side of the house. But that's essentially the model. So 489 00:24:39,880 --> 00:24:42,760 Speaker 5: as it applies to soft Bank, that forty nine percent 490 00:24:42,800 --> 00:24:46,040 Speaker 5: that's owned by soft Bank and ARM ninety five percent 491 00:24:46,160 --> 00:24:49,399 Speaker 5: is soft Bank owned, five percent is ARM owned. But 492 00:24:49,480 --> 00:24:52,919 Speaker 5: from a revenue standpoint, all the revenue consolidates back into ARM. 493 00:24:52,960 --> 00:24:54,360 Speaker 5: There's no revenue in the soft. 494 00:24:54,160 --> 00:24:56,720 Speaker 2: Bank, Okay, very very simplistically. 495 00:24:57,440 --> 00:25:01,320 Speaker 3: Revenue for ARM comes from license fees on royalties, and 496 00:25:01,760 --> 00:25:05,600 Speaker 3: pricing has has gone up. We know that have you 497 00:25:05,720 --> 00:25:08,639 Speaker 3: have you faced pushback from from some of your major 498 00:25:09,000 --> 00:25:12,119 Speaker 3: clients and customers on this. How what is the willingness 499 00:25:12,119 --> 00:25:15,520 Speaker 3: to absorb those high prices and the prices continue to rise. 500 00:25:15,680 --> 00:25:15,879 Speaker 4: You know. 501 00:25:15,920 --> 00:25:17,680 Speaker 5: One of the things that we needed to do when 502 00:25:17,720 --> 00:25:23,159 Speaker 5: we created these new product lines around the cloud and 503 00:25:23,160 --> 00:25:26,240 Speaker 5: and automotive was also re evaluate the business structure and 504 00:25:26,280 --> 00:25:31,800 Speaker 5: business models, because the value proposition of a server based 505 00:25:31,800 --> 00:25:35,080 Speaker 5: core is very different than the value proposition of a 506 00:25:35,119 --> 00:25:38,840 Speaker 5: small core going into a security camera. So the changes 507 00:25:38,880 --> 00:25:41,320 Speaker 5: that we made around pricing, where we have very specific 508 00:25:41,440 --> 00:25:45,399 Speaker 5: royalty rates that exist for the server market we have 509 00:25:45,520 --> 00:25:48,680 Speaker 5: very and the server products we have very specific royalty 510 00:25:48,720 --> 00:25:51,000 Speaker 5: rates for the automotive profit products that go in the 511 00:25:51,000 --> 00:25:54,000 Speaker 5: automotive market. And now we are also building something that 512 00:25:54,000 --> 00:25:58,680 Speaker 5: we call compute subsystems, where we take these independent components 513 00:25:58,800 --> 00:26:01,719 Speaker 5: of a of IP key and then deliver them as 514 00:26:01,760 --> 00:26:05,399 Speaker 5: a full finished solution. So the way to think about it, 515 00:26:05,440 --> 00:26:08,720 Speaker 5: the analogy I like to talk about is legos. So 516 00:26:09,080 --> 00:26:12,439 Speaker 5: if you think about our products as Lego blocks and 517 00:26:12,480 --> 00:26:15,480 Speaker 5: you put them together, maybe to build Big Ben will 518 00:26:15,480 --> 00:26:18,479 Speaker 5: now sell you the entire Big Ben kit, so when 519 00:26:18,520 --> 00:26:20,840 Speaker 5: you put it all together, it'll look just like Big Ben. 520 00:26:21,040 --> 00:26:24,280 Speaker 5: The benefit of that being well, I can build Big 521 00:26:24,280 --> 00:26:26,760 Speaker 5: Ben a lot faster if all the parts just fit, 522 00:26:27,240 --> 00:26:30,760 Speaker 5: and that's huge benefit to the end customer. And if 523 00:26:30,800 --> 00:26:32,880 Speaker 5: you go to the Lego store, the Big Ben Kit 524 00:26:32,960 --> 00:26:35,879 Speaker 5: probably costs a little bit more than the separate Lego blocks. 525 00:26:35,920 --> 00:26:39,000 Speaker 5: But that's essentially where we've gone in pricing. We've increased 526 00:26:39,160 --> 00:26:42,119 Speaker 5: the pricing or changed the pricing around the neoverse and 527 00:26:42,200 --> 00:26:45,840 Speaker 5: automotive products, and then these new subsystems have different pricing 528 00:26:45,880 --> 00:26:48,160 Speaker 5: as well, and that's primarily around the royalties. 529 00:26:48,600 --> 00:26:51,320 Speaker 3: Okay, I love the Big Ben analogy. I love the 530 00:26:51,359 --> 00:26:54,640 Speaker 3: Big Ben analogy. It is becoming more challenging, more complex 531 00:26:54,680 --> 00:26:57,159 Speaker 3: to deliver these solutions. So is it fans assume that 532 00:26:57,160 --> 00:26:59,680 Speaker 3: those price rises will continue. 533 00:26:59,480 --> 00:27:01,879 Speaker 5: Or do you see them peaking? I see the pricing. 534 00:27:01,920 --> 00:27:04,480 Speaker 5: It's a different model altogether. And again going back to 535 00:27:04,520 --> 00:27:07,760 Speaker 5: the Big Ben model, the benefit is twofold for the 536 00:27:07,840 --> 00:27:12,480 Speaker 5: end customer. One is their design time shrinks. What may 537 00:27:12,480 --> 00:27:15,639 Speaker 5: have we had one customer say that you have basically 538 00:27:16,320 --> 00:27:19,800 Speaker 5: taken off eighty man years of design by doing this, 539 00:27:20,880 --> 00:27:22,919 Speaker 5: or where it may have taken them two years to 540 00:27:23,000 --> 00:27:25,720 Speaker 5: develop the product after we deliver the initial LEGO blocks 541 00:27:26,119 --> 00:27:28,359 Speaker 5: in the Big Ben model gets down to thirteen months. 542 00:27:28,760 --> 00:27:31,520 Speaker 5: That's huge savings for a company in terms of entering 543 00:27:31,520 --> 00:27:36,080 Speaker 5: the market. The other driver behind that is the processing 544 00:27:36,160 --> 00:27:39,040 Speaker 5: cycle times to build the chips. So when we deliver 545 00:27:39,119 --> 00:27:41,920 Speaker 5: IP to a customer, they then need to put together 546 00:27:41,960 --> 00:27:45,280 Speaker 5: in their SoC and then essentially send it to a 547 00:27:45,280 --> 00:27:48,720 Speaker 5: fab to be taped out. From the time that they 548 00:27:48,800 --> 00:27:52,159 Speaker 5: deliver their netlist or their file to the fab to 549 00:27:52,280 --> 00:27:55,399 Speaker 5: the time their chips come back sixteen to eighteen weeks. 550 00:27:56,080 --> 00:27:59,400 Speaker 5: As you get down to smaller transistor geometry sizes from 551 00:27:59,480 --> 00:28:02,879 Speaker 5: five nan meter to three to two, those sixteen to 552 00:28:02,920 --> 00:28:05,679 Speaker 5: eighteen weeks go from twenty six to twenty eight weeks. 553 00:28:06,440 --> 00:28:09,240 Speaker 5: So the partners come back to us and say, ARM, 554 00:28:09,280 --> 00:28:12,240 Speaker 5: can you please deliver the product to us eight weeks sooner. 555 00:28:12,640 --> 00:28:15,479 Speaker 5: It's far easier for us to deliver the subsystem and 556 00:28:15,560 --> 00:28:19,320 Speaker 5: save them the eight weeks than deliver them the individual solutions. 557 00:28:19,359 --> 00:28:24,480 Speaker 5: So combination of cycle times being much longer to build 558 00:28:24,480 --> 00:28:28,000 Speaker 5: the product and design cycle times taking longer, there's a 559 00:28:28,000 --> 00:28:31,000 Speaker 5: lot of value that we bring in terms of the 560 00:28:31,040 --> 00:28:34,320 Speaker 5: overall solutions. So because we may come back and say, well, 561 00:28:34,359 --> 00:28:36,800 Speaker 5: your prices went up, I like to think that, well, actually, 562 00:28:36,800 --> 00:28:38,520 Speaker 5: the value that we deliver has gone up. I've actually 563 00:28:38,560 --> 00:28:40,960 Speaker 5: saved you a lot of money by delivering the product 564 00:28:41,000 --> 00:28:42,160 Speaker 5: the way we're now delivering it to you. 565 00:28:42,480 --> 00:28:45,080 Speaker 3: When it comes to competition, A couple of the analysts 566 00:28:45,120 --> 00:28:47,719 Speaker 3: and others have focused a lot on RISK five, So 567 00:28:47,800 --> 00:28:52,120 Speaker 3: the open source platform for building out and designing processes. 568 00:28:52,360 --> 00:28:56,120 Speaker 3: How do you see risk five evolving in the years 569 00:28:56,160 --> 00:28:58,120 Speaker 3: to head? How much of a threat could that be 570 00:28:58,600 --> 00:28:59,680 Speaker 3: in the future if not now? 571 00:29:00,160 --> 00:29:02,600 Speaker 5: I think one of the things that we cheered Chad 572 00:29:02,640 --> 00:29:05,040 Speaker 5: early about ARMS a thirty three year old company, and 573 00:29:05,800 --> 00:29:07,920 Speaker 5: in those thirty three years, one of the things that 574 00:29:07,920 --> 00:29:11,120 Speaker 5: we've done an amazing job of is building a software 575 00:29:11,120 --> 00:29:15,760 Speaker 5: ecosystem like no other. Fifteen million developers, ten million apps 576 00:29:15,800 --> 00:29:21,000 Speaker 5: that are all tuned and run on ARM. The power 577 00:29:21,040 --> 00:29:25,200 Speaker 5: of any CPU is in the software. Lots of CPUs 578 00:29:25,200 --> 00:29:28,680 Speaker 5: income have gone over the years and decades, but it's 579 00:29:28,760 --> 00:29:32,920 Speaker 5: really the software ecosystem. It's running iOS, it's running Android, 580 00:29:32,960 --> 00:29:35,920 Speaker 5: it's running Linux, and it's not just running the software, 581 00:29:36,320 --> 00:29:38,160 Speaker 5: it's knowing that the app is going to work the 582 00:29:38,200 --> 00:29:40,320 Speaker 5: first time. It's not going to be buggy. It's going 583 00:29:40,360 --> 00:29:44,000 Speaker 5: to be the best performance that takes decades to it 584 00:29:44,040 --> 00:29:46,320 Speaker 5: to create. And that's why there's very few companies that 585 00:29:46,640 --> 00:29:50,760 Speaker 5: build microprocessors today. So when people ask me about RISK five, 586 00:29:51,280 --> 00:29:54,080 Speaker 5: I think for certain applications where a software ecosystem is 587 00:29:54,120 --> 00:29:58,240 Speaker 5: not critical, risk five may be an okay solution, But 588 00:29:58,360 --> 00:30:02,160 Speaker 5: where the software ecosystem really really matters, I think it's 589 00:30:02,160 --> 00:30:04,160 Speaker 5: going to take some time for risk five to be 590 00:30:04,480 --> 00:30:05,280 Speaker 5: a viable option. 591 00:30:05,960 --> 00:30:09,960 Speaker 3: And you touched on this the relatively new product from 592 00:30:10,120 --> 00:30:12,720 Speaker 3: m which is ARM Total Design. So that kind of 593 00:30:12,840 --> 00:30:16,240 Speaker 3: end to end chip development that you guys have laid out. 594 00:30:16,440 --> 00:30:17,080 Speaker 2: And pushed down. 595 00:30:17,600 --> 00:30:20,440 Speaker 3: Has there been any reaction from big clients and big 596 00:30:20,480 --> 00:30:23,480 Speaker 3: customers to that, Like, I'm thinking Qualcom you could be 597 00:30:23,520 --> 00:30:27,160 Speaker 3: eating their lunch with this kind of end service and offering. 598 00:30:27,000 --> 00:30:27,520 Speaker 4: It varies. 599 00:30:27,800 --> 00:30:30,880 Speaker 5: If you're a system company and your core business is 600 00:30:30,920 --> 00:30:36,200 Speaker 5: not assembling chips, you look at this product offering and 601 00:30:36,520 --> 00:30:40,760 Speaker 5: you can't get enough of it. Automotive OEMs, cloud data 602 00:30:40,760 --> 00:30:44,760 Speaker 5: center OEMs, wireless base station OEMs, even mobile phone OEMs. 603 00:30:45,280 --> 00:30:48,080 Speaker 5: Now when you go one layer down to people whose 604 00:30:48,320 --> 00:30:51,000 Speaker 5: end business are making chips, is there some tension in 605 00:30:51,040 --> 00:30:53,280 Speaker 5: the discussion relative to Hey, that's the kind of thing 606 00:30:53,320 --> 00:30:57,000 Speaker 5: that I used to do. Initially, Yes, But even now 607 00:30:57,040 --> 00:31:00,320 Speaker 5: across some of the more hardcore areas of even the 608 00:31:00,360 --> 00:31:03,480 Speaker 5: mobile design sector, we're seeing people move to this, and 609 00:31:03,560 --> 00:31:06,320 Speaker 5: it's largely to the reasons I gave you earlier. These 610 00:31:06,360 --> 00:31:09,360 Speaker 5: products are really complicated to build. They take a long 611 00:31:09,440 --> 00:31:14,240 Speaker 5: time to have manufactured. Yet every February at Mobile World Congress, 612 00:31:14,520 --> 00:31:17,440 Speaker 5: the new phone has to be there, and as those 613 00:31:17,960 --> 00:31:21,840 Speaker 5: pressures mount on the design and manufacturing, it's an opportunity 614 00:31:21,880 --> 00:31:23,320 Speaker 5: for ARM to deliver more. 615 00:31:23,480 --> 00:31:26,000 Speaker 3: I want to talk a little bit about Alms Place 616 00:31:26,040 --> 00:31:28,760 Speaker 3: in the UK, almost conceived of cool Sea Ring Cambridge 617 00:31:28,760 --> 00:31:31,200 Speaker 3: in the UK. It's now obviously the majority owned by 618 00:31:31,280 --> 00:31:34,160 Speaker 3: Japanese company so Off Bank, listed in the US what 619 00:31:34,480 --> 00:31:38,520 Speaker 3: role do you see the UK playing for ARM in 620 00:31:38,600 --> 00:31:39,600 Speaker 3: the years ahead. 621 00:31:40,080 --> 00:31:43,920 Speaker 5: The UK is our home, it's our headquarters, it's we 622 00:31:43,920 --> 00:31:46,239 Speaker 5: were born and we're always going to be here. So 623 00:31:46,800 --> 00:31:51,080 Speaker 5: the UK is incredibly central to the future of ARM 624 00:31:51,120 --> 00:31:55,160 Speaker 5: in so many ways. We've had fantastic cooperation in the 625 00:31:55,200 --> 00:31:57,880 Speaker 5: two years or so that I've become the CEO from 626 00:31:57,920 --> 00:32:01,600 Speaker 5: the UK Government, had lots and lots of interaction with them. 627 00:32:01,600 --> 00:32:04,360 Speaker 5: I have personally. They've looked for lots of different ways 628 00:32:04,360 --> 00:32:06,960 Speaker 5: to help us grow, whether it's around fast track for 629 00:32:07,080 --> 00:32:10,240 Speaker 5: talent and such, because again that's probably our biggest bottleneck 630 00:32:10,280 --> 00:32:13,560 Speaker 5: for growth, is just getting talent in. But we are very, 631 00:32:13,600 --> 00:32:16,840 Speaker 5: very committed to the UK, and again the UK is 632 00:32:16,840 --> 00:32:18,360 Speaker 5: going to be so critical for our future. 633 00:32:18,400 --> 00:32:20,040 Speaker 3: And the Prime Minister of the Chancellor that they want 634 00:32:20,080 --> 00:32:21,640 Speaker 3: to create they want to sit in the UK into 635 00:32:21,680 --> 00:32:23,280 Speaker 3: the next Silicon valley. 636 00:32:23,400 --> 00:32:25,640 Speaker 2: Is that realistic running Really. 637 00:32:26,200 --> 00:32:27,800 Speaker 4: There's only one silicon valley, right. 638 00:32:28,080 --> 00:32:32,080 Speaker 5: It's a unique area in terms of its ecosystem, the 639 00:32:32,200 --> 00:32:35,240 Speaker 5: universities and how the whole how the whole section works. 640 00:32:35,440 --> 00:32:39,680 Speaker 5: Does that mean that can't be imitated or replicated on 641 00:32:39,720 --> 00:32:41,520 Speaker 5: some level in other parts of the world. No, not 642 00:32:41,600 --> 00:32:43,960 Speaker 5: at all, and I think the government here has been 643 00:32:44,000 --> 00:32:47,680 Speaker 5: doing a fantastic job around that. And Cambridge itself is 644 00:32:47,720 --> 00:32:53,280 Speaker 5: a very very rich community of small companies, incubators, the universities. 645 00:32:53,760 --> 00:32:57,560 Speaker 5: There's a lot of former arm employees who are involved 646 00:32:57,560 --> 00:32:59,400 Speaker 5: in startups here in the community. So I think I 647 00:32:59,480 --> 00:33:02,000 Speaker 5: think Cambridge in particularly and also other parts of the 648 00:33:02,080 --> 00:33:05,240 Speaker 5: UK including Bristol, are very rich for innovation. 649 00:33:05,600 --> 00:33:07,920 Speaker 3: What do you think the government today's government, but previous 650 00:33:07,960 --> 00:33:10,960 Speaker 3: governments as well, have done wrong in terms of failing 651 00:33:11,640 --> 00:33:13,720 Speaker 3: to attract tech companies to actually. 652 00:33:13,400 --> 00:33:15,760 Speaker 2: List here in the UK. 653 00:33:15,960 --> 00:33:18,760 Speaker 3: Or is it that New York just has this gravitational 654 00:33:18,760 --> 00:33:19,880 Speaker 3: pool that's can't be beaten. 655 00:33:20,480 --> 00:33:23,160 Speaker 5: I'm not a financial guy. I'm an engineer, so I'll 656 00:33:23,200 --> 00:33:25,280 Speaker 5: kind of give you just my layman's. 657 00:33:25,640 --> 00:33:28,120 Speaker 4: Thought around that. I think the capital markets and. 658 00:33:28,360 --> 00:33:31,880 Speaker 5: Scale, I guess maybe speak about the arm Our own case, 659 00:33:32,280 --> 00:33:35,080 Speaker 5: very large capital market, the access to capital is very 660 00:33:35,160 --> 00:33:37,480 Speaker 5: large in New York. I think that was one of 661 00:33:37,520 --> 00:33:40,880 Speaker 5: the reasons that that really drove that in our case. 662 00:33:41,320 --> 00:33:44,240 Speaker 5: But I think broadly back to what really matters for 663 00:33:44,360 --> 00:33:48,120 Speaker 5: arm and myself is attracting world class engineers, and the 664 00:33:48,240 --> 00:33:50,400 Speaker 5: UK has always been a fantastic place for R and 665 00:33:50,480 --> 00:33:52,720 Speaker 5: D and innovation, and I think it will continue to 666 00:33:52,720 --> 00:33:53,200 Speaker 5: be that way. 667 00:33:53,320 --> 00:33:56,880 Speaker 3: Okay, you've been meeting, we meet with regulators, you meet 668 00:33:56,920 --> 00:33:58,479 Speaker 3: with the law makers of course around the world. Have 669 00:33:58,560 --> 00:34:02,680 Speaker 3: you been meeting with members of the Labor opposition party 670 00:34:02,760 --> 00:34:05,320 Speaker 3: and if so, what has your message been to them? 671 00:34:05,400 --> 00:34:07,720 Speaker 4: I have not. Yeah, Okay, so far out yet. 672 00:34:07,760 --> 00:34:08,080 Speaker 2: Okay. 673 00:34:08,160 --> 00:34:08,640 Speaker 4: If they were. 674 00:34:09,080 --> 00:34:12,239 Speaker 3: The polls have the Labor Party in a twenty point lead, 675 00:34:12,400 --> 00:34:13,799 Speaker 3: there's a very good chance that they could be in 676 00:34:13,840 --> 00:34:17,759 Speaker 3: government in twelve months time or less. Often governments set 677 00:34:17,800 --> 00:34:20,759 Speaker 3: themselves one hundred days in terms of to achieve big 678 00:34:20,800 --> 00:34:22,640 Speaker 3: things in their firste hundred days. What would you say 679 00:34:22,719 --> 00:34:25,719 Speaker 3: to any incoming government, whether it's the Conservative government or 680 00:34:25,880 --> 00:34:28,239 Speaker 3: a Labor Party government, that you would want to see 681 00:34:28,280 --> 00:34:31,160 Speaker 3: happen in those first one hundred days for your sector? 682 00:34:31,280 --> 00:34:35,200 Speaker 5: Yeah, specifically for ARM As we talked about earlier, we're 683 00:34:35,320 --> 00:34:37,759 Speaker 5: very committed to the UK. This is our home, we 684 00:34:37,760 --> 00:34:40,799 Speaker 5: were born here, we intend to stay here. Please make 685 00:34:40,840 --> 00:34:43,720 Speaker 5: it very easy for us to attract world class talent 686 00:34:44,239 --> 00:34:47,160 Speaker 5: and attract engineers to come work for ARM because that 687 00:34:47,280 --> 00:34:50,239 Speaker 5: is the single largest limitter that we will have for 688 00:34:50,320 --> 00:34:54,120 Speaker 5: growth is access to talent and if that gets any 689 00:34:54,160 --> 00:34:58,680 Speaker 5: more difficult with a change in government that you described 690 00:34:59,080 --> 00:35:01,560 Speaker 5: and with the next election, that would be a head 691 00:35:01,560 --> 00:35:04,520 Speaker 5: winner would want, would not want to have to engage on. 692 00:35:04,640 --> 00:35:07,720 Speaker 5: So I would say my first MESA would be, please 693 00:35:07,760 --> 00:35:10,480 Speaker 5: make it easy for us to bring people in and 694 00:35:10,600 --> 00:35:12,800 Speaker 5: hire and if you can make it even easier, even better. 695 00:35:13,120 --> 00:35:14,440 Speaker 2: You touched on Cambridge. 696 00:35:14,440 --> 00:35:15,799 Speaker 3: This is my last one on the UK, but you 697 00:35:15,840 --> 00:35:17,960 Speaker 3: touched on Cambridge and there is a real and thriving 698 00:35:18,160 --> 00:35:20,880 Speaker 3: tech ecosystem here and arm is a big reason for that. 699 00:35:20,880 --> 00:35:23,560 Speaker 3: It's been a major catalyst. But there's also been concerns 700 00:35:23,560 --> 00:35:27,680 Speaker 3: about planning restrictions, nimbiism not in my backyard, the inability 701 00:35:27,719 --> 00:35:29,680 Speaker 3: for some of these companies. I don't know if armfuls 702 00:35:29,680 --> 00:35:32,880 Speaker 3: into this category to build out their facilities here because 703 00:35:32,920 --> 00:35:36,320 Speaker 3: of some of those issues. I just wonder is Cambridge 704 00:35:36,360 --> 00:35:40,520 Speaker 3: symptomatic of the UK's problems in terms of not being 705 00:35:40,520 --> 00:35:43,880 Speaker 3: able to drive that growth because of things like planning 706 00:35:43,960 --> 00:35:45,719 Speaker 3: constraints because of nimbiism. 707 00:35:45,800 --> 00:35:47,879 Speaker 2: Do you think there's that issue. Do you think that's a. 708 00:35:47,840 --> 00:35:49,160 Speaker 4: Fair I don't know. 709 00:35:49,280 --> 00:35:51,759 Speaker 5: I personally have not felt that You're sitting here in 710 00:35:51,800 --> 00:35:54,799 Speaker 5: a brand new, very large campus that were built three 711 00:35:54,880 --> 00:35:58,600 Speaker 5: or four years ago that we just moved into for 712 00:35:58,840 --> 00:36:01,920 Speaker 5: buildings ABC, and if you notice when you came in 713 00:36:01,960 --> 00:36:05,640 Speaker 5: there's some construction going on with E and F, so 714 00:36:05,840 --> 00:36:07,920 Speaker 5: for us that we want to hopefully be able to 715 00:36:07,920 --> 00:36:10,400 Speaker 5: take advantage of that space. I think that's also a 716 00:36:10,480 --> 00:36:13,000 Speaker 5: little bit of transition because a lot of things got 717 00:36:13,040 --> 00:36:16,319 Speaker 5: quite disrupted with COVID in terms of people coming back 718 00:36:16,360 --> 00:36:18,360 Speaker 5: into the office and did you need all this office 719 00:36:18,400 --> 00:36:20,840 Speaker 5: space and do people. 720 00:36:20,640 --> 00:36:21,280 Speaker 4: Work from home? 721 00:36:21,880 --> 00:36:24,480 Speaker 5: For the kind of work we do, particularly around engineering, 722 00:36:24,600 --> 00:36:26,400 Speaker 5: I think it's very important for people to be in 723 00:36:26,440 --> 00:36:29,799 Speaker 5: the office at least a good portion of the time 724 00:36:30,800 --> 00:36:32,960 Speaker 5: because one of the things that when you're developed the 725 00:36:33,080 --> 00:36:36,279 Speaker 5: kind of complex products that we are, two areas that 726 00:36:36,320 --> 00:36:39,760 Speaker 5: are very very critical in terms of people being physically together. 727 00:36:39,840 --> 00:36:43,160 Speaker 5: One is the conception of a product, brand new ideas. 728 00:36:43,360 --> 00:36:46,120 Speaker 5: That's very hard to do in your living room on 729 00:36:46,160 --> 00:36:48,960 Speaker 5: a zoom call. And then on the flip side, as 730 00:36:48,960 --> 00:36:51,520 Speaker 5: you're trying to get the product out the door and 731 00:36:51,560 --> 00:36:54,279 Speaker 5: you've got bugs, bugs and issues and you're trying to 732 00:36:54,360 --> 00:36:58,240 Speaker 5: triage them again having people physically in the building. 733 00:36:58,719 --> 00:37:01,480 Speaker 4: So we were a little bit during COVID. 734 00:37:01,960 --> 00:37:03,640 Speaker 5: Now we've got you know a lot of people back 735 00:37:03,680 --> 00:37:06,040 Speaker 5: in the office at least three days a week or 736 00:37:06,040 --> 00:37:08,760 Speaker 5: three and a half days a week, so and probably 737 00:37:08,760 --> 00:37:11,160 Speaker 5: seventy five percent of the people. So this new space 738 00:37:11,160 --> 00:37:12,680 Speaker 5: that we're building, I think we'll be able to put 739 00:37:12,680 --> 00:37:15,600 Speaker 5: it to good use. But I do think also during COVID, 740 00:37:16,680 --> 00:37:19,080 Speaker 5: back to your nimbiasm and expansion, I think some of 741 00:37:19,120 --> 00:37:21,080 Speaker 5: those things got a little bit derailed because people were 742 00:37:21,080 --> 00:37:22,479 Speaker 5: trying to figure out what. 743 00:37:22,440 --> 00:37:25,359 Speaker 4: Is a new normal in terms of office space, And. 744 00:37:25,320 --> 00:37:27,160 Speaker 3: You think hybrid is probably going to be the new normal, 745 00:37:27,320 --> 00:37:28,799 Speaker 3: going for what are you going to see people back 746 00:37:28,800 --> 00:37:29,839 Speaker 3: full time one hundred percent? 747 00:37:30,160 --> 00:37:32,360 Speaker 5: I you know, I don't know that. We'll let me 748 00:37:32,360 --> 00:37:34,120 Speaker 5: see it this way. Yeah, I'm not sure we were 749 00:37:34,120 --> 00:37:37,120 Speaker 5: ever one hundred percent. There were always folks who've found 750 00:37:37,120 --> 00:37:40,200 Speaker 5: ways to kind of work around that. I'm most interested 751 00:37:40,200 --> 00:37:42,840 Speaker 5: in productivity as opposed to a mandate that says you 752 00:37:42,920 --> 00:37:45,800 Speaker 5: must be in five days a week eight am on Monday, 753 00:37:45,840 --> 00:37:48,360 Speaker 5: five pm on Friday. But on the flip side, I 754 00:37:48,360 --> 00:37:50,440 Speaker 5: would not really want to go to a model that 755 00:37:50,440 --> 00:37:53,040 Speaker 5: says everyone can work from home because I just don't 756 00:37:53,040 --> 00:37:55,520 Speaker 5: think you get the best output. You don't solve problems 757 00:37:55,520 --> 00:37:59,200 Speaker 5: the best way. You don't create a sense of how 758 00:37:59,200 --> 00:38:01,719 Speaker 5: to innovate the best way, and probably one of the 759 00:38:01,719 --> 00:38:03,879 Speaker 5: most important things. We have a lot of young people 760 00:38:03,920 --> 00:38:07,719 Speaker 5: in our company who need to be mentored, taught, coached. 761 00:38:08,520 --> 00:38:10,919 Speaker 5: That's a very hard thing to do remotely. So I'd 762 00:38:10,920 --> 00:38:12,120 Speaker 5: like to see it's maturity back. 763 00:38:12,280 --> 00:38:14,600 Speaker 3: I've got a last couple and then we'll wrap it 764 00:38:14,680 --> 00:38:16,719 Speaker 3: up and really appreciate your time, Rene, I want to 765 00:38:16,719 --> 00:38:18,880 Speaker 3: loop it back to AI. Look, in your role, you 766 00:38:19,000 --> 00:38:22,040 Speaker 3: have to be something of a futurist. It's essential to 767 00:38:22,400 --> 00:38:24,400 Speaker 3: what you do at a company like ARM. So as 768 00:38:24,440 --> 00:38:27,799 Speaker 3: you project out a decade, ten years, a decade, how 769 00:38:27,800 --> 00:38:31,400 Speaker 3: do you see generative AI reshaping our world? 770 00:38:32,600 --> 00:38:36,000 Speaker 5: I think it will find its way into everything everything 771 00:38:36,080 --> 00:38:39,720 Speaker 5: that we do, in every aspect of how we work, live, play. 772 00:38:40,200 --> 00:38:42,239 Speaker 5: I think we are on the cusp of something that 773 00:38:43,360 --> 00:38:46,000 Speaker 5: we probably thought five ten years ago was not in 774 00:38:46,000 --> 00:38:49,080 Speaker 5: our lifetimes in terms of being able to have such 775 00:38:49,480 --> 00:38:53,280 Speaker 5: significant impact, whether it's around productivity, whether it's around health. 776 00:38:54,600 --> 00:38:57,120 Speaker 5: I think it's going to change everything over the next 777 00:38:57,120 --> 00:38:57,759 Speaker 5: five to ten years. 778 00:38:58,000 --> 00:39:01,000 Speaker 3: Is there one part of that that most excites you 779 00:39:01,040 --> 00:39:03,560 Speaker 3: in terms of what could change, what could be disrupted, 780 00:39:03,600 --> 00:39:05,760 Speaker 3: what could come about as a result of generative AI 781 00:39:05,800 --> 00:39:06,640 Speaker 3: that most excites you. 782 00:39:07,200 --> 00:39:11,359 Speaker 5: I think health is an area, whether it's around something 783 00:39:11,400 --> 00:39:15,080 Speaker 5: as simple as your doctor visit or something as complicated 784 00:39:15,120 --> 00:39:18,960 Speaker 5: as having a complete diagnostic map of what's right for 785 00:39:19,080 --> 00:39:22,480 Speaker 5: your personal health situation. I think that's what excites me, 786 00:39:23,120 --> 00:39:26,879 Speaker 5: because when I look at the health situation across either 787 00:39:26,880 --> 00:39:31,359 Speaker 5: the UK or other parts of the world, not much 788 00:39:31,400 --> 00:39:33,080 Speaker 5: has changed. When I was ten years old and I 789 00:39:33,120 --> 00:39:36,080 Speaker 5: went to the doctor and a nurse saw me, and 790 00:39:36,080 --> 00:39:38,520 Speaker 5: then there was a clipboard, and there was erotic questions 791 00:39:38,520 --> 00:39:41,319 Speaker 5: that were about as generic as can be. And we 792 00:39:41,320 --> 00:39:44,919 Speaker 5: were talking earlier about my new granddaughter, and I see 793 00:39:44,920 --> 00:39:47,160 Speaker 5: that her doctor visit looks just like my daughters did 794 00:39:47,800 --> 00:39:51,520 Speaker 5: thirty years ago. I think AI just in terms of 795 00:39:51,560 --> 00:39:54,880 Speaker 5: the capability of not only having the visits being more 796 00:39:54,880 --> 00:39:58,840 Speaker 5: efficient and more productive, but more importantly, a real mapping 797 00:39:59,040 --> 00:40:01,759 Speaker 5: of what exactly is your health history, what exactly is 798 00:40:01,800 --> 00:40:04,759 Speaker 5: your DNA, what exactly are the foods that you've eaten 799 00:40:04,760 --> 00:40:07,280 Speaker 5: for the last number of years, giving you the really 800 00:40:07,600 --> 00:40:10,920 Speaker 5: correct level of diagnostic mapping in terms of what's right 801 00:40:11,000 --> 00:40:13,399 Speaker 5: for you. I think the health area is an area 802 00:40:13,440 --> 00:40:15,319 Speaker 5: of just untapped potential. 803 00:40:15,239 --> 00:40:18,879 Speaker 3: And we've seen the debate around risks, risks around generative AI, 804 00:40:19,000 --> 00:40:21,720 Speaker 3: blow in, blow out into the open with what's happening 805 00:40:21,760 --> 00:40:25,000 Speaker 3: with open AI and Sam Altman, where do you. 806 00:40:25,040 --> 00:40:25,560 Speaker 2: Land on this? 807 00:40:25,680 --> 00:40:28,280 Speaker 3: I just wonder is there an element of this story 808 00:40:28,280 --> 00:40:30,719 Speaker 3: that keeps you up at night that concerns you? What 809 00:40:30,760 --> 00:40:33,880 Speaker 3: are you most worried about within the evolution of generative 810 00:40:33,920 --> 00:40:35,600 Speaker 3: AI that. 811 00:40:36,640 --> 00:40:40,160 Speaker 5: We lose control of the machines I think to some extent, 812 00:40:41,400 --> 00:40:45,160 Speaker 5: and the howser still very much being debated obviously, of 813 00:40:45,239 --> 00:40:48,920 Speaker 5: having the failsafe mechanisms in place that humans can override 814 00:40:48,920 --> 00:40:51,719 Speaker 5: the systems. That's probably the single largest thing I think 815 00:40:51,719 --> 00:40:54,640 Speaker 5: and worry about is that if the machines can be 816 00:40:55,000 --> 00:40:58,120 Speaker 5: or the algorithms or those being generated can be developed 817 00:40:58,160 --> 00:41:00,960 Speaker 5: in such a way that there is no fail safe 818 00:41:00,960 --> 00:41:04,040 Speaker 5: mechanism that it can be overridden by a human. 819 00:41:04,719 --> 00:41:05,560 Speaker 4: That's what worries me. 820 00:41:06,200 --> 00:41:07,880 Speaker 2: Running has thank you very much. Indeed,