WEBVTT - Huawei’s Surprise Chip Raises Stakes in Tech Clash 0:00:02.680 --> 0:00:07.320 You're listening to Asia Centric from Blueberg Intelligence, the podcast 0:00:07.360 --> 0:00:10.600 that pulls back the curtain on global business so you 0:00:10.680 --> 0:00:15.040 can invest better across the Pacific Rim. I'm Tom Corbett 0:00:15.040 --> 0:00:17.160 on assignment in the US. 0:00:17.040 --> 0:00:20.759 And I'm John Lee. US LED sanctions on Chinese technology 0:00:20.800 --> 0:00:23.880 firms were designed to be a retaining wall, putting a 0:00:23.920 --> 0:00:27.240 barrier between China and the critical components needed to make 0:00:27.280 --> 0:00:28.880 the most advanced smartphones. 0:00:29.440 --> 0:00:33.920 However, Huawei drew a collective gasp when it revealed its 0:00:34.040 --> 0:00:38.720 latest smartphone, the May sixty Pro, is powered by a 0:00:38.920 --> 0:00:43.400 Chinese homegrown processor. It became proof that China had narrowed 0:00:43.400 --> 0:00:47.520 the gap with the West, giving tech tensions new urgency. 0:00:47.960 --> 0:00:51.120 A Western sanctions too little, too late? And how does 0:00:51.120 --> 0:00:53.320 this compare to Apple's latest iPhone. 0:00:53.520 --> 0:00:57.200 Here to discuss all things tech related is Dan Hutchison, 0:00:57.440 --> 0:01:02.440 vice chair of Tech Insights. He joins us from San Jose, California. Dan, 0:01:02.560 --> 0:01:03.480 great to have you. 0:01:03.680 --> 0:01:06.400 Thanks John Tom for having me here. I'm really excited 0:01:06.400 --> 0:01:07.920 to be here on this podcast. 0:01:08.600 --> 0:01:11.959 Dan, your company did a pair down of Huawei's new smartphone, 0:01:12.080 --> 0:01:15.840 the Mate sixty. What surprised you the most about these findings? 0:01:16.640 --> 0:01:19.640 Well, the thing that was the most surprising. We kind 0:01:19.640 --> 0:01:22.520 of expected that it was going to be a seven 0:01:22.600 --> 0:01:25.720 nanimeter chip, and what surprised us the most was just 0:01:25.800 --> 0:01:29.680 the level of Chinese technology that was in there. You know, 0:01:29.760 --> 0:01:33.880 the motive one that on the APU was all Chinese technology. 0:01:34.240 --> 0:01:37.520 Okay, Dan, can you put that into Layman's terms? What 0:01:37.680 --> 0:01:40.680 is a seven nanimeter chip? And why is it so important? 0:01:41.040 --> 0:01:43.880 Yeah? Okay. The funny thing about it is is that, 0:01:43.959 --> 0:01:46.760 you know, explaining what a nanometer is really crazy. It's 0:01:46.800 --> 0:01:50.960 it's like a thousand a Salven's the thickness of human hair. 0:01:51.640 --> 0:01:54.720 But the smaller the better, right, faller about it, right? 0:01:54.760 --> 0:01:57.880 And so so what it means is is that the 0:01:57.920 --> 0:02:01.680 smaller the better means you can pack more transistors into 0:02:01.920 --> 0:02:05.360 a piece of silicon, so you get more performance your 0:02:05.800 --> 0:02:07.960 you know, your computer can do more. And so it's 0:02:07.960 --> 0:02:12.240 a it's a basic metric of technology, and we go 0:02:12.440 --> 0:02:14.920 down in every two years you try to double the 0:02:14.960 --> 0:02:17.600 density of what's on the chip. And that's the old 0:02:17.680 --> 0:02:18.239 More's law. 0:02:18.440 --> 0:02:21.920 So, Dan, prior to the May sixty being unveiled, the 0:02:22.040 --> 0:02:25.160 overarching sentiment was that China was further behind on this 0:02:25.280 --> 0:02:29.480 technology than it really was. You've been watching the chip 0:02:29.480 --> 0:02:32.639 industry for decades, what surprised you the most when you 0:02:32.800 --> 0:02:35.240 learned about the chip in the May sixty. 0:02:35.440 --> 0:02:37.560 The most interesting thing about the chip in the May 0:02:37.639 --> 0:02:40.480 sixty was that well, one was that it was not 0:02:40.600 --> 0:02:43.720 a five animeter, which the chip and the prior generation 0:02:43.840 --> 0:02:47.639 phone had a TSMC five animetor part in it, so 0:02:47.760 --> 0:02:50.400 it was kind of backed up, and that it was 0:02:50.400 --> 0:02:54.880 a smick chip. The big surprise was that it was 0:02:54.919 --> 0:02:56.800 a non EU of EA chip because they don't have 0:02:56.880 --> 0:03:00.440 access to those tools. And yet it was a pretty 0:03:00.480 --> 0:03:05.919 clean looking set of dimensions, the layout of the chip itself, 0:03:06.440 --> 0:03:09.239 it all looked really clean and it looked as easily 0:03:09.320 --> 0:03:12.440 as good as what TSMC did when it did it's 0:03:12.520 --> 0:03:14.920 seven nanimeter non and U of E chip. 0:03:15.160 --> 0:03:17.400 And the West was not expecting that was. 0:03:17.320 --> 0:03:20.520 It now the West there's a certain arrogance, I think 0:03:20.560 --> 0:03:22.960 in the West in which we think we can throw 0:03:23.000 --> 0:03:25.200 these rolls out there and that the other side will 0:03:25.200 --> 0:03:28.880 it just won't make any progress. And in fact, with 0:03:29.000 --> 0:03:31.960 any technology, if you know it could be done and 0:03:32.000 --> 0:03:35.320 you're throw enough talent at it, you can make progress. 0:03:36.280 --> 0:03:38.960 And then how does the chip performance compared to say, 0:03:39.000 --> 0:03:42.400 other smartphones? Such as maybe like the latest Samsung or 0:03:42.440 --> 0:03:45.640 the latest sell me, well, it's. 0:03:45.480 --> 0:03:47.480 Not going to be as good as a phone that 0:03:47.720 --> 0:03:52.280 has a current generation three N animeter chip. Like what 0:03:52.360 --> 0:03:56.720 we just we're just tearing down the Apple iPhone fifteen 0:03:56.560 --> 0:04:01.480 and we saw a micron hit the record for d 0:04:01.600 --> 0:04:05.760 ram density, which was around a twelve nanimeter twelve thirteen 0:04:05.800 --> 0:04:06.400 animeter n. 0:04:07.520 --> 0:04:10.680 In terms of specs technology, How many years is the 0:04:10.760 --> 0:04:14.119 Huawei Mate sixty pro behind the latest iPhone? 0:04:14.520 --> 0:04:18.080 It's about four years. And if you look at the 0:04:18.120 --> 0:04:20.479 next generation that'll come out in a year or so. 0:04:20.920 --> 0:04:23.039 You know, if we look at the two animeter, they're 0:04:23.080 --> 0:04:26.400 probably about five years behind the very leading edge of 0:04:27.200 --> 0:04:31.840 what's you know, being developed today. But we also expect 0:04:31.839 --> 0:04:35.839 that they're developing their own five nanimeter a part which 0:04:36.480 --> 0:04:38.640 we expect we'll be here in about two more years. 0:04:39.640 --> 0:04:42.720 The chips at power Huawei's latest phone are made by 0:04:42.760 --> 0:04:47.560 SNICK or Shanghai Manufacturing International Corporation. Do you think that 0:04:47.640 --> 0:04:50.240 SNIK is actually making money producing these chips? 0:04:51.120 --> 0:04:54.600 Well, there's two sides of me on that, Okay, okay. 0:04:55.480 --> 0:04:57.640 One side says that there's no way you would do 0:04:57.720 --> 0:05:01.240 a phone and try to produce its scale unless you 0:05:01.360 --> 0:05:04.720 knew the chips were priced right. And there's no way 0:05:04.800 --> 0:05:07.320 if you had a fab of that capability that you 0:05:07.480 --> 0:05:10.719 try to do this unless you believe you could yield. 0:05:11.560 --> 0:05:15.000 My sense is looking at the SEMs and some of 0:05:15.040 --> 0:05:17.799 our people that told me looking at the SEMs, they're 0:05:17.880 --> 0:05:20.520 really good and clean, and I would expect their yields 0:05:20.520 --> 0:05:24.080 to be somewhere in the sixty percent range. They need 0:05:24.080 --> 0:05:26.880 to get to eighty percent to be world class. But 0:05:26.920 --> 0:05:29.320 it's really a question of how well their yield management 0:05:29.400 --> 0:05:31.520 is working, and to know that, we really have to 0:05:31.600 --> 0:05:33.719 have access to their pro yields and look at the 0:05:33.800 --> 0:05:37.320 wafers that sort of thing, which that's not going to 0:05:37.400 --> 0:05:40.800 happen anytime soon. So that's the one side of me 0:05:40.880 --> 0:05:44.200 that says, you know, they wouldn't do this unless they 0:05:44.560 --> 0:05:47.000 really knew they could make money at it. The other 0:05:47.080 --> 0:05:51.080 side of me says, well, they're communist country and they 0:05:51.080 --> 0:05:53.000 could do what they want to in market still matter. 0:05:53.920 --> 0:05:56.640 I remember a friend of mine who had been in 0:05:56.720 --> 0:06:01.159 a Russian fab told me in the eighty because he's 0:06:01.200 --> 0:06:03.320 in a country that they could go back and forth 0:06:03.520 --> 0:06:06.080 cel to the US, sell to Russia. At the time 0:06:06.640 --> 0:06:09.240 and he'd been in a Russian fab and he was 0:06:09.279 --> 0:06:12.520 watching an operator smoke cigarettes and flick the ashes that 0:06:12.680 --> 0:06:15.039 she was moving away first, and he says, aren't you 0:06:15.120 --> 0:06:18.360 worried about yield? And the fab manager looked at us 0:06:18.400 --> 0:06:24.039 as yield that's a capitalist concept, and a government like 0:06:24.080 --> 0:06:27.039 the size of China, which is the second largest economy 0:06:27.040 --> 0:06:29.800 in the world, can do what they want to and 0:06:29.839 --> 0:06:32.200 there's no way for anyone on the outside to really 0:06:32.240 --> 0:06:33.240 know what's happening there. 0:06:33.279 --> 0:06:36.640 When it comes to profitability, yep, and semiconductors has been 0:06:36.720 --> 0:06:39.320 chosen as a critical industry for China. 0:06:39.640 --> 0:06:43.960 So absolutely absolutely, and so you know, you have those 0:06:44.040 --> 0:06:47.000 two sides of it. And so the fact is is 0:06:47.040 --> 0:06:50.719 that the profitability question may not matter. I mean, either 0:06:50.760 --> 0:06:53.080 they are or they don't care. In that sense, it 0:06:53.120 --> 0:06:56.720 doesn't matter. And eventually, the other thing with semi auctors 0:06:56.800 --> 0:06:58.480 is the more you make, the better you get at 0:06:58.480 --> 0:07:01.760 the higher yields go. So even if they're not profitable now, 0:07:02.120 --> 0:07:04.480 they will be at some point as long as they 0:07:04.520 --> 0:07:06.000 have that government support. 0:07:06.520 --> 0:07:08.240 So basically, if you keep on throwing money at the 0:07:08.240 --> 0:07:09.960 problem will eventually sort itself out. 0:07:10.200 --> 0:07:11.280 Yeah. 0:07:10.720 --> 0:07:15.720 Yeah, And what does Huawei's progress tell you about the 0:07:15.760 --> 0:07:20.160 effectiveness of the sanctions and their prospects for success going forward. 0:07:20.960 --> 0:07:22.880 Well, I think unlike some people that say, oh, the 0:07:22.920 --> 0:07:25.400 sanctions aren't working, I looked at it and saw all 0:07:25.440 --> 0:07:29.360 the sanctions are doing what originally everybody thought they were 0:07:29.400 --> 0:07:31.800 supposed to do, which was not to put Huawei out 0:07:31.800 --> 0:07:36.040 of business, but to keep them a couple of generations behind, 0:07:36.160 --> 0:07:40.000 you know, keep them about for you know, two generations 0:07:40.080 --> 0:07:40.880 is four years. 0:07:41.320 --> 0:07:44.120 And then what can always do to narrow the gap 0:07:44.160 --> 0:07:44.800 going forward? 0:07:45.520 --> 0:07:48.920 Well, they don't have access to EUV tools, so you 0:07:48.960 --> 0:07:49.800 know extreme. 0:07:49.520 --> 0:07:52.840 Moment, the equipment made by ASML right been. 0:07:52.760 --> 0:07:57.160 Made by SML, and they don't have access to the 0:07:57.200 --> 0:08:00.400 emergent tools that they bought. Plus there's an older jenneration 0:08:00.600 --> 0:08:03.960 of emersion tool that's still available to them, and the 0:08:04.000 --> 0:08:07.840 immersion tool actually allows them to continue to move forward 0:08:07.840 --> 0:08:10.760 as long as they do multi patterning. And the problem 0:08:10.840 --> 0:08:15.120 with multi patterning is is that it does add inaccuracy 0:08:15.880 --> 0:08:20.280 in the overlay and in the actual dimensions of the 0:08:20.320 --> 0:08:23.640 ship itself. So the result of it will be that 0:08:23.800 --> 0:08:28.880 you get a lower yielding device, it doesn't perform as fast, 0:08:29.520 --> 0:08:31.920 and it consumes more power than what you would like 0:08:31.960 --> 0:08:34.680 it to than what you have with the EUV tool, 0:08:34.800 --> 0:08:36.240 but you can still pull it off. 0:08:37.320 --> 0:08:40.120 Then I also saw your report on the MAT sixty 0:08:40.200 --> 0:08:42.720 ten down and that your company also found some memory 0:08:42.760 --> 0:08:46.600 trips made by sk heinez inside these phones. They weren't 0:08:46.600 --> 0:08:48.600 supposed to be inside these phones, were they. 0:08:48.800 --> 0:08:52.920 No, they weren't. Let me correctly. Was not my report personally. 0:08:53.360 --> 0:08:56.240 There were a lot of really good people working at 0:08:56.280 --> 0:08:58.520 this and you know, and I've been kind of following 0:08:59.040 --> 0:09:01.920 what they've been writing about really closely and looking at 0:09:01.920 --> 0:09:03.960 their SEMs and the pictures they've been coming at. So 0:09:04.200 --> 0:09:07.800 I should kudos to them for what they did. The 0:09:07.920 --> 0:09:14.040 sk Heinex chips, the memories generations back, they weren't current notes, 0:09:14.120 --> 0:09:17.079 so it doesn't look like they actually acquired them directly 0:09:17.120 --> 0:09:21.520 from Heinex, and according to the sanctions, Heinex can't sell 0:09:21.840 --> 0:09:27.599 to Huawei. There's a myriad of ways that because of 0:09:27.600 --> 0:09:31.120 the commodities, those chips can work their way through the 0:09:31.160 --> 0:09:34.920 supply chain, and Huawei could have picked them up without 0:09:34.960 --> 0:09:37.960 them being directly coming, especially since there's a lot of 0:09:38.000 --> 0:09:42.080 memory right now. There's a huge clut of memory coming 0:09:42.120 --> 0:09:45.520 because the memory producers overproduced during the shortage. They kind 0:09:45.559 --> 0:09:48.000 of drank the kool aid about how great this shortage 0:09:48.040 --> 0:09:49.160 was and it was going to last. 0:09:49.800 --> 0:09:53.600 You're listening to Asia Centric from Bloomberg Intelligence. By the way, 0:09:53.679 --> 0:09:56.440 If you like what you hear, and we hope you do, 0:09:56.920 --> 0:10:01.800 please rate us on Apple podcasts, st Google podcasts, or 0:10:01.840 --> 0:10:04.560 wherever you may be listening to us. Of course, more 0:10:04.600 --> 0:10:08.360 stars are better. Your feedback matters, and we love hearing 0:10:08.360 --> 0:10:14.439 from our listeners. Dan Hutchison. One of Apple's defining qualities 0:10:14.440 --> 0:10:17.720 has been its ability to innovate, to come up with 0:10:17.840 --> 0:10:22.320 a new idea and just dazzle the world and power 0:10:22.400 --> 0:10:25.560 its sales to new heights. Is there a perception do 0:10:25.600 --> 0:10:28.880 you think that Apple is starting to come up against 0:10:28.920 --> 0:10:32.640 a bit of a wall in its ability to generate 0:10:32.679 --> 0:10:35.920 that wow factor with new products? And if yes, how 0:10:36.000 --> 0:10:38.520 much are technological limitations at play here? 0:10:39.200 --> 0:10:42.320 Oh? Well, technological limitations are becoming far more at play 0:10:44.000 --> 0:10:47.920 the thing about Apple, And realize that this is probably 0:10:47.960 --> 0:10:52.120 more in my opinion, but you know, you can't discount 0:10:52.200 --> 0:10:56.240 the immense ability of stee jobs. And typically what I've 0:10:56.280 --> 0:10:59.400 seen in the companies I followed is a good strategy 0:11:00.080 --> 0:11:03.560 to last for about five ten years tops before it 0:11:03.640 --> 0:11:06.360 kind of taps out when it comes to the market. 0:11:06.720 --> 0:11:10.760 And Steve had this amazing ability to look at technology 0:11:10.840 --> 0:11:13.840 that was out there piece it together into something that 0:11:14.080 --> 0:11:18.880 was really amazing and fantastic. And I don't think Apple 0:11:18.960 --> 0:11:23.120 struggled to duplicate that. He just had an innate sense 0:11:23.240 --> 0:11:27.680 of what was really great. And those kind of people 0:11:27.760 --> 0:11:30.640 are really hard to find, you know historically, you know, 0:11:30.679 --> 0:11:34.000 when you look at companies that make themselves and Apple's 0:11:34.200 --> 0:11:37.440 struggling to reinvent itself, which the whole thing with the 0:11:37.480 --> 0:11:41.720 App Store and the Apple TV that's coming along, but 0:11:41.800 --> 0:11:44.520 it's also something that there's plenty of other competitors out 0:11:44.520 --> 0:11:45.600 there doing the same thing. 0:11:46.280 --> 0:11:48.520 Once Apple has changed the world, what can to do 0:11:48.640 --> 0:11:49.720 for an encore. 0:11:49.640 --> 0:11:54.720 Yeah, that's a very difficult thing. And if we talk 0:11:54.760 --> 0:11:57.560 about this, how Apple led with the iPhone, with the 0:11:57.600 --> 0:12:01.360 Mac and everything, but it midst the cloud and it missed, 0:12:01.720 --> 0:12:04.480 you know, the impact of what the cloud was doing, 0:12:04.480 --> 0:12:08.199 in which we saw other companies like Google and Facebook 0:12:08.320 --> 0:12:13.600 and Amazon come along. So really, sometimes the innovation comes 0:12:13.640 --> 0:12:17.840 from other places, and that's really where we look forward to. 0:12:18.080 --> 0:12:21.400 And I think in the whole artificial intelligence arena, that's 0:12:21.679 --> 0:12:26.080 a whole spectacular area for massive innovation. It's also an 0:12:26.120 --> 0:12:29.720 area where China has a very strong leadership position. 0:12:30.440 --> 0:12:33.880 And talking about innovation, Apple just release its new iPhone 0:12:33.920 --> 0:12:37.480 fifteen and the buzz seems to be the adoption of 0:12:37.520 --> 0:12:41.520 a USB C port. Is this really the major breakthrough 0:12:41.520 --> 0:12:43.520 for this phone or is there more to it than that? 0:12:44.640 --> 0:12:48.840 No, I think that was that's like expected. Everybody has 0:12:48.880 --> 0:12:51.640 been telling them, you need to come up with their comments, 0:12:51.679 --> 0:12:53.960 so you know, but they wanted to sell more cards 0:12:54.040 --> 0:12:57.480 and make their devices more unique, and you know, that 0:12:57.640 --> 0:13:01.320 was always the downside of Steve Jobs, right, was is 0:13:01.720 --> 0:13:03.920 he wanted everything to be done his way and he 0:13:03.960 --> 0:13:07.160 wanted absolute control over it. And so I just don't 0:13:07.200 --> 0:13:12.000 see any value in fact. For me is a photographer. 0:13:12.320 --> 0:13:15.080 The more exciting part about the iPhone fifteen is the 0:13:15.080 --> 0:13:17.600 new cameras that come with it and the ability that 0:13:17.640 --> 0:13:22.319 they've done. And the Apples phones have really become driven 0:13:22.400 --> 0:13:25.640 by what their cameras could do, and I'm surprised that 0:13:25.640 --> 0:13:27.240 that didn't generate more excitement. 0:13:27.800 --> 0:13:32.280 What about the Age seventeen Chip Dan Hudgson, You've just 0:13:32.400 --> 0:13:36.120 mentioned that consumers seem to gravitate more toward the gadgets 0:13:36.120 --> 0:13:39.040 and the gizmos, the cameras. Is the A seventeen chip, 0:13:39.160 --> 0:13:40.920 Is there anything really game changing in there? 0:13:41.480 --> 0:13:43.880 Well, the big game changer for it's going to be 0:13:43.920 --> 0:13:47.280 that you'll get longer battery life and you will get 0:13:47.360 --> 0:13:50.920 more performance. But I also think it's probably somewhere near 0:13:50.960 --> 0:13:55.000 where the PC hit where we're beginning to see less 0:13:55.000 --> 0:13:58.559 and less performance. So in terms of you know, does 0:13:58.559 --> 0:14:01.400 your call go through all that the same when you 0:14:01.559 --> 0:14:06.600 really see the performance is things like in the artificial 0:14:06.760 --> 0:14:11.600 intelligence application to the image processing for the camera itself, 0:14:12.040 --> 0:14:15.520 and so you get much better stabilization, much lower light, 0:14:15.960 --> 0:14:19.160 much more realistic images, and things like that. The camera 0:14:19.200 --> 0:14:21.200 has become the big reason to upgrade. 0:14:21.800 --> 0:14:24.520 Then there's been some excitement that the new iPhone with 0:14:24.600 --> 0:14:27.520 the A seventeen chip could be a breakthrough for gaming. 0:14:28.320 --> 0:14:30.800 I know that you know some versions of Resident Evil 0:14:30.840 --> 0:14:33.120 have arrived on the iPhone. Do you think this is 0:14:33.160 --> 0:14:34.080 a major game changer? 0:14:34.960 --> 0:14:38.360 Well, you know, if you look at kids and everything, 0:14:38.440 --> 0:14:41.920 they're all using their phones and their iPads for gaming, 0:14:42.040 --> 0:14:46.280 and they're not using laptops or computers. And this is 0:14:46.320 --> 0:14:48.320 the big advantage of this chip. I come back, I 0:14:48.560 --> 0:14:50.800 mentioned the performance and you see it in the AI 0:14:50.880 --> 0:14:54.280 processing for images. You're also going to see it in 0:14:54.480 --> 0:14:58.920 the speed and reactiveness and gaming, which at the end 0:14:58.960 --> 0:15:03.120 of the day, in gaming, that's those slight incremental school 0:15:03.280 --> 0:15:06.000 games give you an advantage in where you're playing a 0:15:06.080 --> 0:15:08.040 driving game or you're on the battlefield. 0:15:08.280 --> 0:15:10.840 So what's next for the iPhone in terms of spekes, 0:15:11.000 --> 0:15:13.320 in terms of the next breakthrough technology? 0:15:13.960 --> 0:15:15.520 To be honest, I think we're going to see more 0:15:15.520 --> 0:15:16.040 of the same. 0:15:16.680 --> 0:15:18.520 The improvements in the camera, you know, and. 0:15:18.440 --> 0:15:20.200 It's improvements in the camera. You're going to get a 0:15:20.200 --> 0:15:23.560 faster chip that works better. The gaming will improve, and 0:15:23.680 --> 0:15:27.880 their whole goal is simply to give people enough incremental 0:15:28.000 --> 0:15:30.880 advantage to buy a new phone and to stay in 0:15:30.920 --> 0:15:34.160 their world garden, because that's becoming more and more. Where 0:15:34.640 --> 0:15:38.800 the money for Apple comes is the access to the 0:15:38.840 --> 0:15:42.840 TV and also just the fantastic usability of its products 0:15:42.880 --> 0:15:46.080 compared to other products that are out there in the world. 0:15:46.480 --> 0:15:48.400 They've had me for ten years in counting. 0:15:48.880 --> 0:15:51.920 Yeah, they've had me since the first iPhone. 0:15:51.680 --> 0:15:54.680 Dan Hudgson, what do you think Steve Jobs would say 0:15:54.720 --> 0:15:58.760 if he could see the Apple of today, the iPhone 0:15:58.800 --> 0:16:03.880 of today, and and the ecosystem. Any thoughts on what 0:16:03.960 --> 0:16:06.760 he would think if he could see all that today. 0:16:07.960 --> 0:16:11.600 I think he'd probably asked the question of what have 0:16:11.640 --> 0:16:15.760 you guys been doing? How did you miss Ai? And 0:16:16.520 --> 0:16:18.840 at the end of the day, they need to come 0:16:18.920 --> 0:16:22.880 up with another product, And they're kind of bifurcated, you know, 0:16:22.880 --> 0:16:26.240 because they've got hardware products that they're known for and 0:16:26.280 --> 0:16:29.600 the usability of those hardware products are known for. But 0:16:29.640 --> 0:16:33.040 they haven't gone they haven't improved on those. The focus 0:16:33.080 --> 0:16:37.640 has all been on the app store and the applications, 0:16:37.680 --> 0:16:40.320 the TV trying to get into the movie business, the 0:16:40.360 --> 0:16:44.040 content business, compete with Netflix, those sorts of things that 0:16:44.160 --> 0:16:46.560 really just keep you looking at their screens. 0:16:47.080 --> 0:16:50.000 What about the virtual reality headsets? It made a lot 0:16:50.040 --> 0:16:51.760 of news last year. Do you think that could be 0:16:51.760 --> 0:16:53.680 a major driver for Apple going forward? 0:16:54.200 --> 0:16:56.880 I seriously doubt that the headset can be a major 0:16:56.960 --> 0:17:01.960 driver for anyone and been a real doubter on the headset. 0:17:02.240 --> 0:17:04.960 And here's my thinking about the headset. Well, one is 0:17:05.400 --> 0:17:07.640 you're not going to carry it with you. In fact, 0:17:07.640 --> 0:17:10.760 most people won't carry a laptop anymore. They'll just carry 0:17:10.800 --> 0:17:13.560 their phone and maybe an iPad. But now that the 0:17:13.600 --> 0:17:16.840 iPad has gotten as heavy as a laptop with a keyboard. 0:17:16.880 --> 0:17:19.719 You wind up choosing one or the other, right if 0:17:19.800 --> 0:17:21.639 you have to. But most of the time you just 0:17:21.680 --> 0:17:25.000 take your phone because it's such a universal device, right. 0:17:25.640 --> 0:17:29.399 And the other issue you have is the usability of 0:17:29.440 --> 0:17:33.280 those devices. The market is driven by the ability of 0:17:33.359 --> 0:17:37.960 people to either have screen time for it or to 0:17:38.320 --> 0:17:41.040 take it with them. The reason why the auto market 0:17:41.240 --> 0:17:44.280 is so huge and why it drove the last century 0:17:44.600 --> 0:17:47.680 it is because it has about a five percent utilization. 0:17:48.560 --> 0:17:50.960 You buy a car and you drive it to work, 0:17:51.000 --> 0:17:53.159 you park it all day, you drive it home, you 0:17:53.240 --> 0:17:56.439 park it all night. And so if everybody has to 0:17:56.520 --> 0:17:59.760 have one and they don't use it as much, that's there. 0:18:00.160 --> 0:18:03.199 Maybe that's the case with the headsets and with the 0:18:03.359 --> 0:18:06.280 virtual reality, but you know, you kind of have to 0:18:06.280 --> 0:18:08.720 come back to how do they get people to move 0:18:09.520 --> 0:18:14.080 towards that over just using their phone or something else 0:18:14.160 --> 0:18:15.119 for their reality. 0:18:16.320 --> 0:18:18.920 So, Dan, you mentioned that Apple appears to have missed 0:18:18.920 --> 0:18:21.800 the step in terms of cloud. They also seem to 0:18:21.800 --> 0:18:25.600 have missed AI revolution. Which companies are leading the innovation? 0:18:25.840 --> 0:18:29.800 Do you think, well, in terms of the cloud, I mean, 0:18:29.840 --> 0:18:33.760 clearly you've got Microsoft and Amazon leading the whole cloud 0:18:33.760 --> 0:18:38.879 space and when it comes to AI, how can you 0:18:38.960 --> 0:18:41.920 say the word AI and not mentioned in videos name 0:18:42.119 --> 0:18:45.520 yes and what they've been able to do. And then 0:18:45.720 --> 0:18:49.479 you've also got Intel and A and D trying to 0:18:49.520 --> 0:18:51.919 come back and dig their way out of the PC 0:18:52.160 --> 0:18:56.000 trough if they've been in with their AI chips, but 0:18:56.040 --> 0:18:58.600 they've got a long ways to go because while they 0:18:58.640 --> 0:19:03.680 were focusing on AI, Wang kind of snuck in and 0:19:03.720 --> 0:19:04.719 stole the party. 0:19:05.119 --> 0:19:07.560 And what about some of the Asian companies. Is TSMC 0:19:07.800 --> 0:19:09.760 still leading in terms of the logic chips and do 0:19:09.800 --> 0:19:11.719 you think you see this continue going forward? 0:19:12.000 --> 0:19:15.159 Absolutely? The thing when you get in the lead of 0:19:15.200 --> 0:19:19.320 the semiconductor industry, you have so many advantages because of 0:19:19.359 --> 0:19:22.840 the scale you have. That said, it's really up to 0:19:22.880 --> 0:19:26.040 the engineers to continue to make the innovations. And it's 0:19:26.080 --> 0:19:29.399 so easy to trip up and to fall behind. Intel 0:19:29.480 --> 0:19:32.840 did this in the nineteen eighties, TSMC did it in 0:19:32.920 --> 0:19:36.600 the late nineties two thousands, they started to slip behind, 0:19:37.160 --> 0:19:41.280 and really they had slipped systematically behind until twenty eight 0:19:41.280 --> 0:19:43.800 and animeter and that was when they started to catch 0:19:43.840 --> 0:19:48.160 back up again. So it's really a question of can 0:19:48.240 --> 0:19:52.040 they continue to innovate, and that may become a limit 0:19:52.119 --> 0:19:55.840 for them because the labor shortage of technical talent in 0:19:55.880 --> 0:19:59.320 Taiwan is becoming really short. Because they've grown so much, 0:19:59.359 --> 0:20:02.760 they've become so big, it's harder for them to hire and. 0:20:02.600 --> 0:20:05.800 I Samsung, like Tronics, dominates in memoratips, but they've also 0:20:06.000 --> 0:20:09.080 made a lot of investments into the logic space. Does 0:20:09.119 --> 0:20:11.879 Samsung all oppose any threat to TSMC? 0:20:13.080 --> 0:20:17.840 If we had the TSMC executives here talking honestly, they 0:20:17.880 --> 0:20:23.280 would probably tell you that their biggest worry is Samsung yep. 0:20:23.640 --> 0:20:26.720 And just because of the raw resources and their ability 0:20:26.800 --> 0:20:30.600 to throw money at it and their ability to use 0:20:30.640 --> 0:20:33.439 the scale from memory and then apply that to logic, 0:20:34.240 --> 0:20:38.120 and that's really the big issue for them. The hard 0:20:38.160 --> 0:20:43.360 part for Samsung is they're really good at making memory, 0:20:43.480 --> 0:20:47.159 and memory is a very different animal to logic in 0:20:47.280 --> 0:20:49.199 terms of what you're building and how you build it 0:20:49.200 --> 0:20:53.480 because memory is a pure commodity and logic is not 0:20:53.560 --> 0:20:57.800 a commodity. All of those designs go to a particular 0:20:58.040 --> 0:21:02.520 end system and they're totally dependent on where their customer, 0:21:02.600 --> 0:21:04.919 how what the volume they want to buy. And if 0:21:04.960 --> 0:21:07.200 they say, oh, we don't want those wafers and they're 0:21:07.200 --> 0:21:11.520 halfway into production. You're kind of host where in memory, 0:21:11.680 --> 0:21:15.000 whatever you make you can only sell. So it makes 0:21:15.040 --> 0:21:18.159 it a much more difficult market in terms of managing 0:21:18.200 --> 0:21:22.120 your fabs and managing the losses they trace around the fabs. 0:21:22.440 --> 0:21:26.040 And that's something that TSMC has been a real master of. 0:21:26.480 --> 0:21:28.840 And the old model for TSMC, why they were so 0:21:28.960 --> 0:21:33.040 successful was the intersections in type, you know, and y 0:21:33.119 --> 0:21:37.560 TSMC came out winning versus in memory. The intersections. Everybody 0:21:37.640 --> 0:21:40.000 stops at the stop light and waits. That's the way 0:21:40.040 --> 0:21:44.119 a memory fab works, and in a boundary fab it 0:21:44.200 --> 0:21:46.800 has to work like an intersection in type, where the 0:21:46.840 --> 0:21:49.359 only rule is you don't hit the fender of the 0:21:49.400 --> 0:21:50.040 car next to you. 0:21:50.600 --> 0:21:51.520 Organized chaos. 0:21:51.960 --> 0:21:56.960 Yeah, Dan, you've been covering the chip industry since the 0:21:57.040 --> 0:22:00.840 late nineteen seventies nineteen seventy eight. I think what in 0:22:00.880 --> 0:22:05.040 those forty years of covering this industry surprises you the 0:22:05.080 --> 0:22:09.760 most about the way the industry has evolved, technology, the 0:22:09.800 --> 0:22:13.800 overarching economic and political implications. What stands out in your 0:22:13.880 --> 0:22:16.840 mind when you look back over your career and gaze 0:22:16.960 --> 0:22:18.359 over that landscape. 0:22:18.520 --> 0:22:21.880 Well, I would have to say more recently, Number one 0:22:22.160 --> 0:22:25.920 is that Washington finally figured out that chips were important 0:22:26.520 --> 0:22:28.760 and that all of those smart things. 0:22:28.760 --> 0:22:30.040 They were a little late to that game. 0:22:30.119 --> 0:22:31.880