芯片禁令再升级，全球半导体产业格局将如何演变？
字幕摘录
| 时间 | 英文 | 中文 |
|---|---|---|
| 0:04 | Welcome everyone to webinar three in the series of Ted and Logan on high tech industries, | 欢迎大家参加Ted和Logan关于高科技产业系列的三集 |
| 0:09 | courtesy of the LSOM and genome and our great team, Ted and Logan. | 我们伟大的团队泰德和洛根 |
| 0:15 | Today we will be discussing the semiconductor industry that has been a lot in the news, | 今天我们将讨论 半导体工业 一直是很多的新闻, |
| 0:21 | but why don't I turn it to Ted and we will show the bio and then we'll get going with | 但是,为什么不我把它给Ted, 我们会展示的生物 然后我们去 |
| 0:28 | the topic. | 主题。 |
| 0:30 | Great. | 不错 |
| 0:31 | All right. | 好吧,我们走。 |
| 0:32 | Well, good morning. | 好吧,早上好,你好吗? |
| 0:33 | Good evening. | 边 |
| 0:34 | Good afternoon, everybody. | 午安,大家好,谢谢 |
展开字幕全文（652 条）
| 序号 | 英文 | 中文 |
|---|---|---|
| 1 | Welcome everyone to webinar three in the series of Ted and Logan on high tech industries, | 欢迎大家参加Ted和Logan关于高科技产业系列的三集 |
| 2 | courtesy of the LSOM and genome and our great team, Ted and Logan. | 我们伟大的团队泰德和洛根 |
| 3 | Today we will be discussing the semiconductor industry that has been a lot in the news, | 今天我们将讨论 半导体工业 一直是很多的新闻, |
| 4 | but why don't I turn it to Ted and we will show the bio and then we'll get going with | 但是,为什么不我把它给Ted, 我们会展示的生物 然后我们去 |
| 5 | the topic. | 主题。 |
| 6 | Great. | 不错 |
| 7 | All right. | 好吧,我们走。 |
| 8 | Well, good morning. | 好吧,早上好,你好吗? |
| 9 | Good evening. | 边 |
| 10 | Good afternoon, everybody. | 午安,大家好,谢谢 |
| 11 | As we usually say, I'm really happy to be joined by my colleague, Logan Bender. | 正如我们通常所说的,我很高兴能与我的同事洛根·班德一起工作。 |
| 12 | Logan, how are you? | 罗根,你好吗? |
| 13 | I'm well. | 我没事 |
| 14 | Thank you. | 谢谢 |
| 15 | It's nice to see our colleagues from around the world and who we are again for another | 很高兴能见到我们来自世界各地的同事 我们再次成为另一个 |
| 16 | 正如Jakob所说, | |
| 17 | It's really not an important industry, Logan. | 这真的不是一个重要的行业,洛根。 |
| 18 | Yeah, it's kind of a cottage industry niche, if you will. | 是啊,这是一种家庭工业的优势,如果你愿意。 |
| 19 | So, yeah, no, Logan, why do you think it is such an important industry at this point? | 所以,是的,不,洛根,你为什么认为 这是这样一个重要的行业在这个时候? |
| 20 | It's become ubiquitous. | 已成无边. |
| 21 | I mean, we have everything is powered by semiconductors in some way. | 半导体以某种方式提供动力 |
| 22 | I mean, you think about the fact that we carry the history of the human word on tiny machines | 我的意思是,你想想,事实上,我们 背负历史的人字 在微小的机器 |
| 23 | in our pocket are all enabled by these chips, these semiconductor chips, and not just with | 我们口袋里有这些芯片, 这些半导体芯片, 而不仅仅是用 |
| 24 | the AI boom. | AI的热潮。 |
| 25 | But for decades, this industry has been building. | 但几十年来,这个行业一直在建设中. |
| 26 | It's in the news more now. | 今更见新闻. |
| 27 | We have more semiconductor content in cars, in electronics, and everything that's around | 汽车 电子产品 周围的一切 |
| 28 | us that powers our lives, our digital lives. | 赋予我们生命 数字生活 |
| 29 | And it's in manufacturing. | 而在制造业. |
| 30 | It's everywhere. | 处处有. |
| 31 | So the ubiquity of the semis industry, there's a lot of opacity. | 故半种工业无所不在,多不透明. |
| 32 | It's opaque. | 其色不净. |
| 33 | The reason why is because it's so complicated, so complex. | 原因之所以如此复杂 |
| 34 | So we're going to break it down. | 故下释. |
| 35 | We're going to go to different parts of the value chain. | 我们要去价值链的不同部分。 |
| 36 | Where is the value captured? | 价值在哪里? |
| 37 | Is it transferred to the consumer ultimately? | 最终是否转移给消费者? |
| 38 | Is it captured by the people that manufacture the chips, such as the foundries, or is it | 是制造芯片的人,如铸币局抓的 还是 |
| 39 | GPU,IP设计师等等? | |
| 40 | So we're going to break open all that. | 所以,我们要打破这一切。 |
| 41 | We're going to talk about industry dynamics. | 谈行业动态. |
| 42 | It should be good. | 应该很好 |
| 43 | So take notes. | 所以记下来吧 |
| 44 | And back to you, Ted. | 回到你身边 泰德 |
| 45 | Let's get started. | 让我们开始吧。 |
| 46 | So if we go to the next slide, this is who we are. | 所以,如果我们去下一个幻灯片, 这就是我们是谁。 |
| 47 | Many of you who are joining were participants in the class that Logan and I taught for five | 你们中很多人都是洛根和我教的五门课的学员 |
| 48 | years, and Jakub supported so effectively, economics of high tech industries. | 和贾库布如此有效的支持 高科技产业的经济学。 |
| 49 | And what we're trying to do is reach out to people who took the course, Yale, SOM alumni, | 而我们试图做的是 接触的人 谁上课程,耶鲁,SOM校友, |
| 50 | Yale SOM students, people from GNAM, the Global Network for Advanced Management. | 耶鲁SOM的学生,来自GNAM的人,全球高级管理网络. |
| 51 | And then, of course, we want to say thanks at the outset to Yale Center Beijing, because | 当然,我们想首先感谢耶鲁中心北京,因为 |
| 52 | most of our participants tune in through the Yale Center Beijing WeChat. | 我们大部分的参与者都通过耶鲁中心 北京 WeChat 调校。 |
| 53 | Is that correct, Jakub? | 是这样吗,贾库布? |
| 54 | That is correct. | 没错 |
| 55 | OK. | 摆 |
| 56 | So and for those of you who are interested in updates from Logan and myself, check out | 还有你们这些对洛根和我最新消息感兴趣的人 请检查一下 |
| 57 | the website. | 网站。 |
| 58 | For those of you who would like to get a copy of the slides after the presentation, we'll | 对于你们中想在演讲后拿到幻灯片副本的人,我们将 |
| 59 | post those on the website. | 贴在网站上。 |
| 60 | And one thing that I do want to mention, Logan captured really why we're doing this and what | 还有一件事我想说,洛根 真正抓住了为什么我们这样做,什么 |
| 61 | we're doing. | 我们正在做。 |
| 62 | Some of the results that you're going to see today haven't been presented anywhere before, | 今天你将要看到的一些结果 还没有被介绍过 |
| 63 | and they are preliminary based on some data analysis. | 它们是初步的,基于一些数据分析。 |
| 64 | And I wanted to give a shout out to another member of our team, Harsic Agrawal, who has | 我想给我们团队的另一名成员 哈西克·阿格拉瓦尔大喊一声,他已经 |
| 65 | been working on collecting information about the market caps of various firms in the semiconductor | 正在收集半导体各公司市场上限的信息。 |
| 66 | industry. | 工业。 |
| 67 | All right. | 好吧,我们走。 |
| 68 | Well, let's go to the next one. | 好吧,让我们去下一个。 |
| 69 | So here are the questions and demand, supply. | 这就是问题和需求,供应。 |
| 70 | That's one and two. | 即为一二. |
| 71 | Who are the leading edge at each stage? | 每个阶段谁是领先者? |
| 72 | As Logan mentioned, which firms are capturing the most value? | 正如洛根提到的那样,哪些公司最有价值? |
| 73 | You might say, oh, it must be the rare earth suppliers, right? | 你可能会说,哦,一定是稀土供应商,对吧? |
| 74 | They're the ones that are in the news. | 他们是新闻里的人 |
| 75 | Well, that's not what we find. | 好吧,这不是我们发现的。 |
| 76 | Why do some firms succeed and others struggle? | 为什么有些企业要成功,另一些企业要挣扎? |
| 77 | And then related to that, do firms with high market shares have market power the way people | 与此相关 市场份额高的公司 是否有市场支配力 |
| 78 | normally think about it? | 通常考虑一下? |
| 79 | So some quick background on the industry, and Logan, I'll ask you your view on this. | 所以,一些关于这个行业的快速背景, 和洛根,我会问你 关于这一点的看法。 |
| 80 | But actually, you know, it's interesting. | 但实际上,你知道,这是有趣的。 |
| 81 | The industry has been around a very long time. | 这个行业已经很长一段时间了。 |
| 82 | If you go back to 1958, that's when Texas Instruments first integrated multiple electronic | 如果你回到1958年 那就是德克萨斯仪器公司首次整合多个电子设备的时候 |
| 83 | components on a single silicon chip. | 单硅芯片上的组件。 |
| 84 | And prior to that, IBM, for example, had already developed the mainframe computer. | 在此之前,例如IBM已经开发了主机计算机. |
| 85 | So this was an industry that had a role, a presence, for many, many decades. | 因此,这个行业在很多,几十年里 都扮演了角色,存在。 |
| 86 | But Logan, I don't know if you would agree, but I think the industry developed relatively | 但Logan,我不知道你是否同意, 但我认为这个行业 相对发展 |
| 87 | slowly until the last couple of decades, and then it became what you call an apex technology. | 慢慢地,直到过去几十年, 然后它变成了你所谓的顶级技术。 |
| 88 | But do you agree with that? | 但你同意吗? |
| 89 | I do. | 当然 |
| 90 | I mean, with the rise of computers in terms of a computer in every household, that was | 我的意思是,随着计算机的兴起 每个家庭的计算机,这就是 |
| 91 | one major inflection point. | 一个主要的切入点。 |
| 92 | Obviously, internet adoption drove business models that we couldn't even conceive of. | 显然,互联网收养驱动了商业模式 我们甚至无法想象。 |
| 93 | But, you know, we get wrapped up in how enabled we are in the digital world. | 但是,你知道,我们在数字世界里 被怎样的能力所包裹。 |
| 94 | But if we step back and we think to ourselves, OK, at the end of the day, a chip that's smaller | 但是如果我们退后,我们想着自己,好吧,在最后,一个芯片是较小的 |
| 95 | than our fingernail could easily exceed the output and productivity of 10,000 people, | 比我们的指甲容易超过一万人的产量和生产力, |
| 96 | what they would be able to achieve in a year, this tiny chip can achieve in mere seconds. | 这个小芯片能在短短几秒钟内实现 |
| 97 | And so the raw gain in productivity is unprecedented. | 因此,生产力的原始收益是前所未有的。 |
| 98 | And so there's going to be more of it. | 然则将多之. |
| 99 | And obviously, more firms are going to throw their hat in and compete in this industry. | 很明显,更多的企业 将会把帽子扔进这个行业 并进行竞争。 |
| 100 | And then you have these waves of innovation, such as what AI has done. | 然后你们有了这些创新的浪潮,比如AI所做的. |
| 101 | AI has also been around since the 70s and even the 60s. | 从70年代甚至60年代起,大赦国际也一直在身边。 |
| 102 | They were playing with different ideas about some general intelligence. | 他们正在玩 不同的想法 一些一般智能。 |
| 103 | But fast forward to today, we have AI has been democratized. | 但时至今日,我们的AI已经民主化了。 |
| 104 | It's been applied to almost every industry, and it's increasing the addressable market | 它被应用到几乎每个行业, 它正在增加可处理的市场 |
| 105 | and increasing the capabilities of many firms. | 提高许多公司的能力。 |
| 106 | And so those are powered by semiconductor chips, particularly GPUs with NVIDIA and AMD | 这些由半导体芯片供电 特别是带有NVIDIA和AMD的GPU |
| 107 | as the leaders in that industry. | 作为这个行业的领导者。 |
| 108 | And so we've seen this slow emergence of semis. | 因此,我们看到了这种缓慢的半衰期的出现。 |
| 109 | And then you have these dramatic inflection points, which is really exciting to see. | 然后你们有了这些戏剧性的插曲点, 这非常令人兴奋。 |
| 110 | So that's what we'll talk about today, Ted. | 所以这就是我们今天要谈的 Ted |
| 111 | Yeah. | 对 |
| 112 | So the thing that I would highlight, along with what Logan highlighted, | 所以,我要强调的东西, 与洛根强调, |
| 113 | which is the amazing increase in productivity. | 这是生产力的惊人增长。 |
| 114 | I mean, the chips with Moore's law have become more and more powerful. | 摩尔定律的芯片越来越强大了 |
| 115 | It's also, I think, worthwhile, and Logan, you mentioned the internet, | 我认为也是值得的 洛根 你提到了互联网 |
| 116 | to think about what complementary things have happened, complementary advances in technology. | 思考一下什么是互补的, 互补的技术进步。 |
| 117 | And as Logan mentioned, the internet, another very big one is the ability to communicate | 正如洛根提到,互联网, 另一个非常大的是交流能力 |
| 118 | through various technologies. | 通过各种技术。 |
| 119 | Now we're adding satellites. | 现在我们正在增加卫星。 |
| 120 | But that means that the communications abilities of devices and things and people | 但这意味着设备、物品和人的通讯能力 |
| 121 | who have access to powerful semis means that they can communicate with each other | 他们能接触到强大的半音 意味着他们可以互相交流 |
| 122 | and send information to each other. | 互相传递信息 |
| 123 | And probably if we go back to the history of telecommunications, | 也许如果我们回到电信的历史, |
| 124 | the move from 2G to 3G and then 4G was a major complement to what semiconductors could do | 从2G到3G,然后4G是半导体所能做的主要补充 |
| 125 | in smart devices. | 在智能设备。 |
| 126 | You could communicate through the modem chips, which end up being really an important | 你可以通过调制解调器芯片进行交流,这些芯片最终真正成为一个重要的 |
| 127 | type of chip. | 芯片类型。 |
| 128 | And now we're all connected. | 现在我们都连在一起了 |
| 129 | Things are connected, the internet of things. | 事情是连接的, 事物的互联网。 |
| 130 | So I think it's a combination of phenomenal increase in productivity, | 所以我觉得这是生产力惊人增长的结合 |
| 131 | as well as the development of complementary technologies. | 以及开发互补技术。 |
| 132 | All right, next slide. | 好吧,下一个幻灯片。 |
| 133 | Oh, wait a minute. | 单单 |
| 134 | No, I think that's right. | 不,我想这是正确的。 |
| 135 | I mean, obviously, semis are really an important component of the China-U.S. rivalry. | 我的意思是,很明显,半音其实是中美竞争的重要组成部分. |
| 136 | But let's turn to demand. | 但让我们开始要求。 |
| 137 | And when I see a chart like this, I always try to ask questions. | 当我看到这样的图表时,我总是想问问题。 |
| 138 | OK, government, 1%? | 好吧,政府,1%吗? |
| 139 | That can't be right. | 这不可能 |
| 140 | So this is the typical thing that Logan and I do. | 这就是我和Logan的典型做法。 |
| 141 | We see data and we try to go deeper. | 我们看到数据,我们试图更深入。 |
| 142 | This can't be right, that demand is only 1%. | 这不可能, 需求只有1%。 |
| 143 | Government must be on the receiving end of a lot of the other demand. | 政府必须接受许多其他要求。 |
| 144 | So there's a sequence of demand here. | 所以这里有一系列的需求。 |
| 145 | Another question I have is, where's AI on this chart? | 我还有另一个问题 这张图表上的AI在哪? |
| 146 | And Logan, I think you have an answer to that. | 还有洛根 我想你有答案了 |
| 147 | Well, yeah, I mean, now it's everywhere. | 嗯,是的,我的意思是,现在它到处都是。 |
| 148 | It's in all these slices that we see emergence of AI adoption in many of these companies. | 在所有这些片段里,我们看到许多这些公司出现了AI的收养. |
| 149 | Now, one important question is also, OK, well, how applicative is it, for example, | 现在,一个重要问题是,OK,嗯, 它是多么实用,例如, |
| 150 | Schneider Electric? | 施耐德电器公司? |
| 151 | AI对Schneider电动建筑的应用性 | |
| 152 | relative to an Apple or a Huawei or obviously Nvidia and AMD? | 相对于苹果还是华威 或者很明显的Nvidia和AMD? |
| 153 | It is, you definitely want to, when you listen to earnings calls, | 这是,你肯定想, 当你听收入电话, |
| 154 | you talk to management teams, you'll kind of get a sense of how much traction they're getting. | 你和管理团队交谈, 你会发现他们得到了多少动力。 |
| 155 | We'll look at that in capital expenditures and revenues that these firms, | 我们看看这些公司的资本支出和收入 |
| 156 | some of the firms are generating from their bets on AI and from their initiatives on AI. | 一些公司正在通过对大赦国际的赌注和对大赦国际的主动行动产生效果。 |
| 157 | But again, as ubiquitous, it's been infused in the ecosystem and it has been coming for years. | 但同样,作为无处不在的, 它被注入到生态系统中 并且它已经来了好几年。 |
| 158 | These many firms have been talking about AI for years. | 这些公司多年来一直在谈论AI. |
| 159 | But in the last couple of years, we've really seen the inflection. | 但在过去的几年里,我们真的看到了这种空虚。 |
| 160 | And will it drive the next leg of growth in the tech space, Ted? | 它能带动科技空间的下一段成长吗 Ted? |
| 161 | So if you look at automotive there and PC, computer, industrial, those are major drivers. | 所以,如果你看看汽车 那里和PC,计算机,工业, 这些是主要的驱动器。 |
| 162 | If you go to the next slide, just a quick flash on, they're everywhere. | 如果你去下一张幻灯片, 只是一个快速闪光, 他们到处都是。 |
| 163 | Automotive, PCs, industrial, et cetera. | 汽车,个人电脑,工业等. |
| 164 | Let's go to the next. | 来至下座. |
| 165 | Well, actually, it's just pause there on cars. | 其实,这只是暂停在汽车上。 |
| 166 | How many semis are in a car? | 车里有几个半人? |
| 167 | Well, it's hard to know, but of course it varies. | 嗯,这很难知道, 但当然不同。 |
| 168 | But a good starting point would be for a new EV would be around 3,000 semiconductors. | 但一个良好的起点是 新的EV是大约3000个半导体。 |
| 169 | Let's go to the next slide. | 下一张幻灯片 |
| 170 | Logan, this is a phenomenon that maybe you could break down, | 洛根,这是一个现象 也许你可以打破, |
| 171 | but on the left-hand side of the vertical axis on this graph, | 但在本图的垂直轴左侧, |
| 172 | is the percentage of revenues. | 是收入的百分比。 |
| 173 | And what's being depicted is how much, | 被描绘的就是多少 |
| 174 | what percentage of revenues are being spent by major US tech companies. | 美国主要技术公司在收入中所占的百分比是多少。 |
| 175 | And it may be a bit of an eye test there, but it's a huge percentage. | 这可能是一个有点 眼睛测试,但它是一个巨大的百分比。 |
| 176 | So Logan, maybe you could put this into historical context. | 所以洛根,也许你可以把这个 历史背景。 |
| 177 | Yeah, we've been tracking this for a couple of years. | 是的,我们已经追踪了几年了 |
| 178 | But about 18 months ago or so, and on the next slide we can see as well, | 但大约18个月前, 在下一张幻灯片上,我们也可以看到, |
| 179 | let's stick with this one for now. | 我们先继续这个 |
| 180 | We saw investment, heavy investment in data center, | 我们看到投资, 大量投资数据中心, |
| 181 | but well above and beyond, head and shoulders beyond what we've seen before. | 但远超前方 超越我们以前所见 |
| 182 | The striking thing was that we saw subsequently every quarter they spent more. | 令人惊讶的是,我们后来看到每个季度 他们花费更多。 |
| 183 | And these are leading companies. | 这些是领先的公司。 |
| 184 | Obviously, they employ some of the most brilliant minds in the world. | 显然,他们雇用了世界上最聪明的人 |
| 185 | These are what we call the hyperscalers or the big cloud giants. | 这些就是我们所谓的超标或大云巨人。 |
| 186 | You have Meta, Microsoft, Google, Amazon. | 你有Meta 微软 谷歌 亚马逊 |
| 187 | There's some others as well that are spending almost on the level. | 还有一些也几乎花在水平上。 |
| 188 | But it's not just the scale of investment, but it's the velocity of investment. | 但是,这不仅仅是投资的规模,而是投资的速度. |
| 189 | That they're spending more and more and more every quarter | 他们每个季度都花得越来越多 |
| 190 | into the hundreds of billions of dollars without stopping, without pause. | 进入千亿亿无间断无间断无间断. |
| 191 | Which tells you that these operating teams, these executive teams, | 这告诉你,这些操作团队, 这些执行团队, |
| 192 | that these are some of the largest companies on earth, | 这些是世界上最大的公司, |
| 193 | are making a huge bet and taking tremendous operational firepower | 正在做一个巨大的赌注 和巨大的作战火力 |
| 194 | and putting it towards the idea that the data centers and AI are going to power the future | 并把它引向一个想法 数据中心和AI 将驱动未来 |
| 195 | and they can't miss out. | 他们不能错过。 |
| 196 | One point as an analyst, I was a semiconductor analyst. | 有一点作为分析师,我是半导体分析师。 |
| 197 | I was an analyst covering global technology stocks from China all the way to Europe to US. | 我是一个分析家,负责从中国到欧洲到美国的全球技术储备。 |
| 198 | And when I look at a chart like this, | 当我看到这样的图表时, |
| 199 | and the reason why we put this chart in some of our decks | 以及我们把这张图放进甲板的原因 |
| 200 | is some companies you see they seem to spend defensively, | 你所看到的是一些公司 他们似乎在防御性地花费, |
| 201 | meaning that there's an inflection in the industry. | 意思是这个行业有漏洞 |
| 202 | They feel like they have to get behind it. | 他们觉得必须躲在后面 |
| 203 | Maybe they feel like they're behind, so they spend to catch up. | 也许他们觉得自己落后了,所以花时间追赶. |
| 204 | Whereas other companies may spend proactively making an operational bet saying, | 虽然其他公司可能主动作出业务赌注, |
| 205 | hey, listen, we're going to go out and achieve this growth initiative | 嘿,听着,我们要出去 实现这个成长计划 |
| 206 | and not try to chase the market. | 而不是试图追求市场。 |
| 207 | And so I think different firms in the subset, for example, Amazon, | 所以,我认为不同的公司 在子集, 例如亚马逊, |
| 208 | they have their AWS ecosystem. | 它们有它们的AWS生态系统。 |
| 209 | They have a phenomenal ecosystem. | 它们有一个惊人的生态系统。 |
| 210 | It's very sticky, meaning it's very difficult for users to leave. | 其粘度很高,意味着用户很难离开. |
| 211 | The company that I am with now, | 我现在所在的公司 |
| 212 | we have Amazon, we have AWS. | 我们有亚马逊,我们有AWS。 |
| 213 | It would be very difficult for us to transition to another cloud or have dual cloud. | 我们很难过渡到另一个云层,或者有双重云层。 |
| 214 | And so Amazon has a captive audience they can sell incremental product into. | 因此,亚马逊有一个 被囚禁的受众 他们可以销售增量的产品。 |
| 215 | This is really driven growth. | 这是真正的驱动增长。 |
| 216 | And so I think they're spending proactively. | 所以我认为他们正在积极花费。 |
| 217 | Whereas Meta, some would argue, is chasing. | 而Meta,有些人会争论,在追逐。 |
| 218 | They're saying, okay, our advertising business, maybe it isn't so good anymore. | 他们说,好吧,我们的广告业务, 也许它已经不是那么好。 |
| 219 | And so we're going to invest really heavily in a defensive way | 因此,我们将投入大量 防御方式 |
| 220 | to try to catch up to the market or establish some kind of a foothold | 试图赶上市场 或建立某种立足点 |
| 221 | to cement ourselves as a leader in artificial intelligence. | 作为人工智能的领导者 |
| 222 | And you can see it reflected in how much they're spending. | 你可以从他们花费的多少中看出这一点。 |
| 223 | And imagine a firm as big as Meta spending $4 of every 10 that it makes | 想象一个像梅塔那样大的公司 每10美元花4美元 |
| 224 | for advancing these technologies and these initiatives. | * 推动这些技术和举措。 |
| 225 | It's amazing. | 不可思议. |
| 226 | And I've never seen anything like it. | 亦不曾见. |
| 227 | And it's been its historic amount of spend, Ted. | 这是它的历史开支 Ted |
| 228 | So three things, if I can remember these three things so early in the morning. | 因此,三件事,如果我能记得这三件事 这么早。 |
| 229 | One is obviously here we're focusing on US firms. | 一个显然在这里,我们正在专注于美国公司。 |
| 230 | But we could show an analogous set of data on Chinese firms. | 但我们可以展示一套关于中国企业的类似数据。 |
| 231 | Chinese firms are Tencent, Alibaba, Baidu are doing the same thing. | 中国企业是Tencent,阿里巴巴,白都也在做同样的事情. |
| 232 | But here we have data on the US. | 但这里有美国的数据 |
| 233 | But this is a phenomenon. | 但这是一个现象。 |
| 234 | I think it's mainly the US and China. | 我认为主要是美国和中国。 |
| 235 | Second point is, as Logan mentioned, the velocity and the implications of that. | 第二点是,正如洛根提到的那样,速度及其影响。 |
| 236 | So now we look at this chart and we say, oh my gosh, | 现在我们看这个图表 然后说,哦,我的天哪, |
| 237 | look at the percentage of revenues being spent. | 研究所花收入的百分比。 |
| 238 | It used to be that these investments were funded internally. | 过去,这些投资是内部供资的。 |
| 239 | But when you get to 40% of investments by these firms, | 但当你得到40%的投资 这些公司, |
| 240 | they're placing huge capital structure bets on AI. | 他们把巨大的资本结构 赌在AI。 |
| 241 | And I think we both agree this is really changing the nature of competition | 我觉得我们都同意 这真的改变了竞争的性质 |
| 242 | among these hyperscalers. | 这些超规模的 |
| 243 | It's a kind of bet that to me is so large, | 这是种赌注,对我来说是如此之大, |
| 244 | I can't believe that all of these firms are going to win. | 我不敢相信所有这些公司都会赢 |
| 245 | And as Logan pointed out, and this is the third point, | 正如洛根指出 这是第三点 |
| 246 | these are not sort of just homogeneous investments. | 这些投资并不仅仅是同质投资。 |
| 247 | The firms have different strategies. | 公司有不同的策略。 |
| 248 | They have different complementary resources. | 它们有不同的补充资源。 |
| 249 | Meta may be more defensive. | 梅塔也许更防守 |
| 250 | Amazon is in a position where they have the complementary customer base, | 亚马逊的客户基础是互补的 |
| 251 | AWS, cloud computing, etc. | AWS,云计算等. |
| 252 | Where their bets have, if you will, more of a natural work playground | 他们的赌注,如果你愿意, 更多的自然工作游乐场 |
| 253 | to which they can apply these investments and look for returns. | 它们可以运用这些投资并寻找回报。 |
| 254 | But let's shift from percentage investments to absolute. | 但让我们从百分比投资转向绝对投资. |
| 255 | And this chart, again, it's a subset of the major US companies, | 这个图表也是美国各大公司的子集 |
| 256 | but these dollar amounts are, I think, in my experience, historic. | 但是,我认为,根据我的经验,这些美元数额是历史性的。 |
| 257 | I don't know what comments you would offer about these, | 我不知道你会怎么评价这些 |
| 258 | but look at those dollar amounts. | 但看看这些美元的数量。 |
| 259 | And this is just, if I recall, is this just for a particular quarter? | 这仅仅是,如果我记得, 这只是一个特定的季度? |
| 260 | Yeah, this is quarterly. | 是的,这是季度。 |
| 261 | And $52 billion, I mean, just that amount of money alone in absolute terms | 520亿美元,我的意思是, 光是这笔钱的绝对值 |
| 262 | exceeds the value of what, one third of the S&P 500 | 超过什么值, 三分之一的 S&P 500 |
| 263 | in terms of companies that are below the $50 billion in market cap. | 就低于500亿美元市场上限的公司而言。 |
| 264 | I mean, it's incredible. | 我的意思是,这是不可思议的。 |
| 265 | They can't be making enough money right now to justify these investments. | 他们现在不能赚到足够的钱来证明这些投资是合理的。 |
| 266 | That's right. | 没错 |
| 267 | This is why we call them bets. | 这就是为什么我们叫他们下注。 |
| 268 | Ah, yes. | 对 |
| 269 | Okay. | 摆 |
| 270 | And then next slide, it's not obviously just the firms that we saw before | 接下来的幻灯片 不仅仅是我们之前看到的公司 |
| 271 | or even the MAG-7 in the US. | 甚至是美国的MAG -7 |
| 272 | A recent example is Oracle this week. | 最近的一个例子是本周的甲骨文。 |
| 273 | Their stock popped because of their investments in so-called intelligent agents. | 他们的股票因为投资 所谓的智能代理。 |
| 274 | And it's interesting. | 则趣之. |
| 275 | We had put on the website our list of potentially top 10 kind of success stories in AI, | 我们在网站上公布了AI十大成功故事 |
| 276 | and we included Oracle. | 而我们包括甲骨文。 |
| 277 | So we could say, good to us. | 所以我们可以说,对我们很好。 |
| 278 | We saw Oracle outside the MAG-7, and right before this webinar, we get this event. | 我们在MAG -7门外看到甲骨文 就在这个网络前 我们得到了这个事件 |
| 279 | I'm a little, though, suspicious of this jump. | 我有点怀疑这个跳跃 |
| 280 | I'm hearing that these contracts actually are not very well distributed across | 我听说这些合同实际上 分布得并不理想 |
| 281 | a large number of companies seeking Oracle's software and technology. | 大量公司寻求甲骨文的软件和技术. |
| 282 | And instead, it's concentrated in contracts with OpenAI. | 相反,它集中于与OpenAI的合同. |
| 283 | And whenever you see something like that, I think you should question, | 当你看到类似的东西, 我想你应该质疑, |
| 284 | well, how solid is that? | 那么,有多扎实? |
| 285 | What happens if OpenAI can't meet the terms of the contract? | 如果OpenAI不能满足合同条款怎么办? |
| 286 | But obviously, it means something. | 但显然,它意味着什么。 |
| 287 | I just don't think it. | 我只是不认为。 |
| 288 | My guess is I'd be suspicious of the 40% jump. | 我猜我会怀疑40%的跳跃 |
| 289 | But yeah, and then, of course, outside of the US, Logan, maybe you could comment on that. | 但是,当然,在美国之外,洛根, 也许你可以评论一下。 |
| 290 | We don't have access to data like this. | 我们无法访问这样的数据。 |
| 291 | The Wall Street Journal is a great source on US companies, but not so much for Chinese companies. | "华尔街日报"是美国公司的一个绝佳来源,但对于中国公司来说却不那么重要. |
| 292 | But Logan, we have every reason to believe that Chinese companies are doing the same thing. | 但洛根,我们有充分的理由相信中国公司也在做同样的事情. |
| 293 | Oh, yes. | 对 |
| 294 | They have been spending the Chinese companies that are leaders in the space. | 他们一直在花中国公司 在太空领先。 |
| 295 | And even e-commerce or a diverse company conglomerate like Alibaba, | 即使是电子商务 或像阿里巴巴这样的多元化公司集团 |
| 296 | they have a phenomenal cloud business. | 他们有一个惊人的云层业务。 |
| 297 | I rely on my former colleagues in Hong Kong that cover these from the investment banks to catch up. | 我靠在香港的前任同事,从投资银行那里掩盖这些事情来赶上. |
| 298 | When Ted mentioned Oracle here, this is part of our class and our program. | 当Ted在这里提到Oracle时 这是我们的班级和计划的一部分 |
| 299 | We bring in current events that are the latest market news | 我们带来了最新市场新闻 |
| 300 | that relate directly to the themes that we're covering. | 直接涉及我们所报道的主题 |
| 301 | And we'll unpack that and our views on that in real time. | 我们会及时整理出来 整理我们的观点 |
| 302 | And so we try to enable the students or our audience who are speaking to, such as this morning, | 因此,我们努力使学生 或我们的观众 谁说话,如今天上午, |
| 303 | what does the Oracle move mean for Oracle? | 甲骨文对甲骨文意味着什么? |
| 304 | What does it mean for the competition or the people that are competing with Oracle? | 这对竞争者或与甲骨文竞争的人意味着什么? |
| 305 | And what does that mean for the ecosystem and its future impact? | 这对生态系统及其未来影响意味着什么? |
| 306 | Oracle had the largest earnings quarter since, I think, the early 1990s, I think 1992. | Oracle是1990年代初以来收入最高的季度,我想是1992年。 |
| 307 | Part of that probably was short covering. | 这部分可能只是短篇报道。 |
| 308 | But that aside, Oracle now shows tremendous promise. | 但是,除了这个之外,甲骨文现在显示出巨大的希望。 |
| 309 | Are they going to be a leader? | 他们要当领袖吗? |
| 310 | Are they going to be as competitive as some of the giants that we saw before? | 他们会像我们之前看到的巨人一样有竞争力吗? |
| 311 | We've seen, and Ted, maybe you want to comment on this before we move on, | 我们已经看到,和泰德, 也许你想 评论这一点之前,我们继续, |
| 312 | lane convergence of Google used to do search. | 用于搜索的 Google 的车道汇合。 |
| 313 | Facebook曾销售广告. | |
| 314 | Meta used to sell advertisements. | Meta曾销售广告. |
| 315 | Xiao Mei came out as a cell phone maker. | 小梅作为手机制作人出道. |
| 316 | Now Xiao Mei does automotive. | 现在小美做汽车。 |
| 317 | Meta does data center and other applications. | Meta负责数据中心和其他应用. |
| 318 | Oracle's doing cloud agents to the tune of hundreds of billions of dollars. | 甲骨文作云剂以调千亿. |
| 319 | We've seen convergence in competition, hundreds of billions of dollars of spend. | 我们看到竞争趋同,花费了数千亿美元。 |
| 320 | What an interesting time to watch these companies compete and what their agendas are. | 看着这些公司竞争和它们的议程是怎样的有趣时刻。 |
| 321 | Yeah, I don't have anything to add. | 是的,我没有什么要补充的。 |
| 322 | I think this is a really fascinating time when we're going to see divergence in strategies | 我认为这是一个非常迷人的时代, 我们将看到在策略上的分歧 |
| 323 | and success rates across these firms. | 这些公司的成功率。 |
| 324 | Let's go to the supply side, the next slide. | 让我们去供给方,下一个幻灯片。 |
| 325 | If you look at this chart, and I'll let people digest it a little bit, | 如果你看看这个图表, 我会让大家稍稍消化一下, |
| 326 | but I direct your attention to number one on the left-hand side, raw materials. | 但我请大家注意左边的第一,原料 |
| 327 | Then right below that are end purchasers. | 那下面是最终购买者 |
| 328 | What do we mean by end purchasers? | 最终购买者是什么意思? |
| 329 | We basically mean buyers of finished semiconductors. | 我们基本上是指半导体成品的买家。 |
| 330 | Nvidia doesn't make them, so they are actually a buyer, | Nvidia不制造它们 所以它们其实是买家 |
| 331 | 但特斯拉购买了完成的半导体, Tencent does, Microsoft, Samsung等. | |
| 332 | But what this chart is intended to tell us is that there are a lot of stages of production | 但这张图表想告诉我们的是 生产过程有很多阶段 |
| 333 | between the initial stage of raw materials and the final purchasers number nine. | 在原材料初始阶段与最后购买者第9之间。 |
| 334 | They include the wafer production, which is... | 包括编织,即... |
| 335 | I actually visited an Intel foundry a long time ago and I saw the wafer production. | 很久以前,我去过一个英特尔铸造厂 我看到了蜡烛的制作 |
| 336 | That's the disks that are developed and then imprinted on those are the designs that generate | 这些磁盘是开发出来的,然后印在那些上面的 是生成这些磁盘的设计 |
| 337 | actually yield semiconductors. | 实际产生半导体。 |
| 338 | But we've got design, we've got equipment. | 但我们有设计,有设备. |
| 339 | Number five is a very important step. | 第五是非常重要的一步。 |
| 340 | That's the actual production of the chips themselves. | 这就是芯片本身的实际生产. |
| 341 | Then there are steps involved in manufacturing and assembly and testing before they get to the | 接下来还有制造、装配和测试步骤 |
| 342 | end purchasers. | 最终购买者。 |
| 343 | We have a complicated supply chain. | 我们有一个复杂的供应链。 |
| 344 | That's point number one. | 即第一义. |
| 345 | Let's go to the next slide. | 下一张幻灯片 |
| 346 | What we've tried to do, and this is highly subjective, everybody, | 我们试图做的, 这是高度主观的,大家, |
| 347 | we acknowledge this, but we've tried to identify some leading firms on the supply side. | 我们承认这一点, 但我们试图确定一些主要公司 在供应方面。 |
| 348 | And we've also conducted case studies on how these individual firms are doing. | 我们还对这些公司的情况进行了个案研究。 |
| 349 | But here are leading firms at the raw material stage. | 但这里是主要企业 在原材料阶段。 |
| 350 | Some of these names you recognize. | 你认识的一些名字 |
| 351 | 然后在设计阶段,ARM和Synopsys和Qualcomm等. | |
| 352 | Let's go to our subjective identification of leading firms at each stage. | 我们先从主观上确定每个阶段的主要企业. |
| 353 | This is the same chart populated by what we would just call out as the top | 这张图和上面的图一样 |
| 354 | supplier at each stage of production. | 每一生产阶段的供应商。 |
| 355 | Again, subjective, but raw materials. | 再说一遍,主观的,但原料的。 |
| 356 | Shinetsu, wafer production. | 光绪间,瓦片制作. |
| 357 | We've got global wafers, design. | 我们有全球饼,设计。 |
| 358 | 我们有Qualcomm设备,ASML。 | |
| 359 | By the way, we'll come back to ASML. | 顺便说一句 我们会回到ASML的 |
| 360 | But then that key foundry step, TSMC, end purchasers, companies like NVIDIA. | 但后来那个关键的铸造步骤,TSMC,最终购买者,NVIDIA等公司. |
| 361 | But ASML is really an important step. | 但ASML确实是一个重要的步骤. |
| 362 | Logan, do you want to say anything about ASML? | 罗根 你想谈谈ASML吗? |
| 363 | Really briefly. | 真的很短暂。 |
| 364 | So these are the companies such as ASML. | 这些是ASML等公司。 |
| 365 | There's not very many of them. | 其数不多. |
| 366 | They do what's called lithography, meaning they make the machines that make the chips. | 他们做所谓的绘画, 意味着他们制造制造芯片的机器。 |
| 367 | They etch the chips with lasers. | 他们用激光打芯片 |
| 368 | So you have a mask and they shoot lasers through the mask and they create the chip designs. | 所以,你有一个面具, 他们射激光 通过面具, 他们创造了芯片设计。 |
| 369 | And so these are very expensive machines, hundreds of millions of dollars for one unit. | 所以这些机器非常昂贵, 一个单位需要数亿美元。 |
| 370 | It's highly specialized. | 高度专业化. |
| 371 | The IP is tremendous that goes into this. | 这里面的IP是巨大的。 |
| 372 | And so the ASML is a very important company. | 因此ASML是一个非常重要的公司. |
| 373 | So some of the value is bottlenecked into ASML. | 因此部分值被瓶颈为ASML. |
| 374 | And a couple of other competitors, ASML being the leading one, Ted. | 还有另外几个竞争对手,ASML是领先者,Ted. |
| 375 | And probably the only major firm outside US or China. | 可能是美国或中国之外唯一的大公司 |
| 376 | I'm probably going to regret that statement, but they really are unique. | 我可能要后悔那份声明, 但他们确实是独一无二的。 |
| 377 | All right, so on the supply side, what would we say | 好吧,所以在供给方面,我们会说什么 |
| 378 | are the most important things you should know about the supply? | 你该知道最重要的事情吗? |
| 379 | Obviously, we've already established that there are multiple stages. | 显然,我们已经确定 有很多阶段。 |
| 380 | Each stage involves a lot of lags in terms of the development of new products and new processes. | 每个阶段在开发新产品和新工艺方面都涉及许多滞后. |
| 381 | A lot of the firms are specialized. | 很多公司都是专业的 |
| 382 | It's hard to be good at each of these stages. | 这些阶段的每个阶段都很难做到好. |
| 383 | And another observation, which I think is important not only for today, | 还有另一个观点 我认为不仅对今天很重要 |
| 384 | but our likely next topic for our next webinar, | 但我们下个婚礼的下一个话题, |
| 385 | the top firms are not located in the same country. | 顶级企业不在同一国家。 |
| 386 | 所以想一想TSMC和Qualcomm。 | |
| 387 | TSMC在台湾. | |
| 388 | Qualcomm在美国. | |
| 389 | Well, that's going to complicate the supply curve. | 这会使供给曲线复杂化 |
| 390 | And then one more point, which is, | 还有一点,那就是, |
| 391 | it's global, but not only is it global, there's a mismatch | 是全球性的,但不仅是全球性的,还有不匹配的 |
| 392 | between the producing and consuming countries. | 在生产国和消费国之间。 |
| 393 | And that's illustrated on the next slide. | 下一张幻灯片就说明了这一点。 |
| 394 | Hard to read, but producing is the blue bar and consuming is the orange bar. | 很难读,但生产是蓝色的条,消耗是橙色的条。 |
| 395 | And you can see how the bars for individual countries and regions don't match up. | 你可以看到,个别国家和地区的酒吧是如何不匹配的。 |
| 396 | So countries are net purchasers or net suppliers along the way. | 因此,各国沿途都是净购买者或净供应商。 |
| 397 | And I'll just tack on that TSMC headquartered in Taiwan is the biggest fab in the world | 我就去调查总部在台湾的TSMC 是世界上最大的潮流 |
| 398 | that produces the most. | 生产量最大 |
| 399 | Most of us know this, but it's tremendously important for the entire ecosystem of semis. | 我们大多数人都知道这一点, 但是这对整个半个生态系统来说都极为重要。 |
| 400 | And we'll see in terms of value capture. | 且从价值捕捉上看. |
| 401 | So let's remember value capture, which firms are capturing the value | 所以,让我们记住价值捕捉, 哪家公司正在捕捉价值 |
| 402 | in this inflection, in this era of semiconductor and AI inflection. | 在这个半导体和人工智能的时代 |
| 403 | Who's capturing the value? | 谁在捕捉价值? |
| 404 | Think about that as we go through these slides. | 当我们通过这些幻灯片时考虑一下。 |
| 405 | Taiwan is tremendously important because they have the biggest foundries, | 台湾很重要 因为他们有最大的铸币局 |
| 406 | the biggest fabs in the world, Ted. | 世界上最大的潮流,泰德。 |
| 407 | Yeah, so, okay, so given that the supply chains are complicated and global | 是啊,所以,好吧,所以考虑到 供应链是复杂和全球性的 |
| 408 | and firms specialize, a question arises. | 和公司专门, 一个问题。 |
| 409 | Well, how do firms actually interact with each other? | 公司如何实际互动? |
| 410 | What is the nature of their transactions? | 他们的交易性质是什么? |
| 411 | And for those of you who are alumni of 960, you've seen this chart before. | 而对于你们这些960届校友来说,你们以前也看过这个图表. |
| 412 | But in different market settings, | 但在不同的市场环境下, |
| 413 | you have a range of ways that firms transact with each other. | 你们有各种公司互相交易的方式 |
| 414 | And then the left-hand part of this continuum, | 然后这个连续体的左手部分, |
| 415 | there's the dark blue bar that says spot market. | 有暗蓝色的酒吧 说点市场。 |
| 416 | Well, the spot market for semiconductors is pretty much irrelevant. | 半导体的现货市场是无关紧要的 |
| 417 | There's too much specific investment and the processes are long, | 有太多的具体投资 过程很长, |
| 418 | the lead times are long, there's co-development. | 准备时间很长,有共同开发。 |
| 419 | So what happens is firms, when they do enter into potential transactions, | 所以,会发生的是公司, 当它们确实进入潜在的交易, |
| 420 | they focus more on contracts and sometimes exclusive contracts. | 它们更侧重于合同,有时还侧重于排他性合同。 |
| 421 | The right-hand side of this continuum is much more important than the left side. | 这个连续体的右侧比左侧重要得多. |
| 422 | So firms choose between partial vertical integration, | 因此,公司选择 部分垂直整合, |
| 423 | long-term contracts, single source contracts that have durations. | 长期合同,有期限的单一来源合同。 |
| 424 | So that's where the action is and I think it's important to mention that | 所以这就是行动所在 我认为必须提及 |
| 425 | because that sets up how the answer to the question. | 因为这样就能回答问题了 |
| 426 | Well, given that that's how firms contract, that's also how firms win and lose. | 嗯,考虑到这是公司合同的方式, 这也是公司赢输的方式。 |
| 427 | Okay, I'm looking, I'm trying to monitor some of these interesting questions. | 好吧,我看,我试图监视 这些有趣的问题。 |
| 428 | We'll try to get to these later. | 我们等会再谈 |
| 429 | Okay, given how firms contract, | 好吧,鉴于如何 公司签约, |
| 430 | we wanted to get into the question of which firms at what stages of production, | 我们想问一下哪个企业处于生产阶段 |
| 431 | again, which firms at which stages of production are capturing the largest | 生产阶段最大的企业 |
| 432 | percentage increases in their market cap. | 其市场上限百分比增加。 |
| 433 | And the time period we selected is 2015 to 2025. | 我们选择的时间段是2015年到2025年。 |
| 434 | And this is based on our subjective selection of leading firms with the caveat | 这是基于我们主观选择的 领先公司,但警告 |
| 435 | that for us, for the analysis to proceed, these have to be publicly traded. | 对我们来说,为了继续分析,这些必须公开交易。 |
| 436 | So for those reasons, I would say these are preliminary results. | 因此,出于这些原因,我要说,这些都是初步结果。 |
| 437 | But here, Logan, when we look at which ones pop, I mean, what are your comments | 但在这里,洛根,当我们看看哪一个流行, 我的意思是,你有什么评论 |
| 438 | on these increases in market cap by stage? | 市场上限的提高? |
| 439 | I'm not so surprised by these gains in these various stages. | 我对这些不同阶段的进展并不感到惊讶。 |
| 440 | And my comment would be that when you look at everything from rare earths, | 我的评论是 当你从稀土中观察一切 |
| 441 | raw materials, all the way to the IP designers or something like an Nvidia, | 原料,一路IP设计师 或者像Nvidia之类的东西, |
| 442 | which designs GPUs, they do not manufacture in-house, they contract that out, | 设计GPU,他们不生产 内部,他们承包, |
| 443 | but they're the designers. | 但他们是设计师 |
| 444 | When you go further towards the edge of the leading technology, of the bleeding edge | 当你更接近领先技术的边缘 流血的边缘 |
| 445 | or the leading edge, however you want to put it, you seem to see the most value transfer, | 或领先的边缘,无论你想把它, 你似乎看到最价值转移, |
| 446 | the most value capture towards the leading edge of that tech stack. | 最高值捕捉到 领先的科技堆。 |
| 447 | That's expressed again here. | 再说一遍 |
| 448 | You have the most gains in the IP heavy companies such as Nvidia. | 在Nvidia等IP重型公司中 你收益最大 |
| 449 | Nvidia is up several thousand percent in the last several years. | Nvidia在过去几年中上升了几千 %。 |
| 450 | And it's surprising how much they've gained, but it's not surprising that they've | 令人惊讶的是,他们得到了多少, 但这并不奇怪,他们 |
| 451 | had a huge spike in their stock price and their valuation, | 他们的股票价格和估价都大幅上涨, |
| 452 | TED9 or Classco and devaluation in great detail, how it relates to both public | TED9 或 Classco 和 贬值非常详细, 它如何联系两个公共 |
| 453 | and private markets. | 和私人市场。 |
| 454 | For this chart specifically, it doesn't surprise me that it's farther towards | 对于这个图表来说,我并不感到惊讶的是,它更接近 |
| 455 | the higher end of the tech stack that we see the value capture. | 高端的技术堆栈,我们看到值捕获。 |
| 456 | So you mentioned Nvidia. | 所以你提到了恩维迪亚 |
| 457 | Nvidia is one of the end purchasers. | 恩维迪亚是最终购买者之一. |
| 458 | So that's a big driver in that historic increase. | 因此,这是一个巨大的驱动力 在历史性的增加。 |
| 459 | Are you surprised, Logan, by raw materials being on the low end of the increase? | 你感到惊讶吗,洛根? 原料在低端增加? |
| 460 | Not as much because there's not as much IP that's involved in that. | 没有那么多 因为参与其中的IP不多 |
| 461 | I'm not saying that the machines that mine the rare earths are not sophisticated or the supply | 我不是说开采稀土的机器 不是精密的,也不是供应的 |
| 462 | chains and logistics are not sophisticated by any means, but as you have more specialization | 连锁店和物流不以任何手段先进,但随着你的专业化程度提高 |
| 463 | in human capital, labor, research scientists, et cetera, you're going to tend to see, at least | 在人力资本,劳动,研究科学家等等, 你会倾向于看到,至少 |
| 464 | in high tech, you're going to tend to see more value capture there. | 在高科技中,你会看到更多的价值捕捉。 |
| 465 | Let's relate this to one really important question, though. | 让我们把这个联系到 一个非常重要的问题,虽然。 |
| 466 | And I posed this earlier on. | 之前我还做了这个 |
| 467 | When you think about a technology such as AI powered by semiconductors, is it ultimately | 当你想到一个技术,如由半导体驱动的 AI,它是最终的 |
| 468 | the consumer? | 消费者? |
| 469 | Do I, with my chat GPT and my tools to augment my job is what I do every day. | 难道我,用我的聊天GPT 和我的工具 增加我的工作是我每天做的。 |
| 470 | In my personal time, do I yield more gain than an Nvidia or an OpenAI? | 在我的个人时间里,我能得到的比恩维迪亚或OpenAI更多的收益吗? |
| 471 | Is OpenAI giving away so much value to me that I, the consumer, may ultimately, | OpenAI给了我这么多价值 我,消费者,最终可能, |
| 472 | one could argue, that I yield more value over time. | 人们可以争辩,我逐渐产生更多的价值。 |
| 473 | And so does the value transfer? | 价值转移呢? |
| 474 | An analogy that's maybe a little bit clumsy, but I use is over time with food production gains, | 一个类比,也许有点笨拙, 但我使用的时间 与粮食生产收益, |
| 475 | you've seen the calorie value transfer towards the consumer and not necessarily towards the | 你见过卡路里价值 转移给消费者,而不一定对 |
| 476 | manufacturer. | 制造商。 |
| 477 | Many food manufacturers are low margin businesses. | 许多食品制造商是低利润企业。 |
| 478 | In technology, are we going to see, with the democratization of these interesting | 在科技方面,我们是否会看到, 随着这些有趣的民主化 |
| 479 | technologies, are we going to see that value ultimately transfer towards the consumer? | 技术,我们会看到这种价值最终转移到消费者身上吗? |
| 480 | Or will it be towards the firm or it'll be some other part of the industry, Ted? | 还是向着律所走 还是从其他行业走 泰德 |
| 481 | So I agree, Logan. | 所以我同意,罗根。 |
| 482 | Now, these are not weighted by value. | 现在,这些不是按价值加权的。 |
| 483 | So we're just focusing on the percentages. | 所以我们只是专注于百分比。 |
| 484 | But again, raw materials, modest increase compared to the others. | 但同样,原料,与其他原料相比略有增加. |
| 485 | That's not what you would expect if you just read the news | 如果你只看新闻,你就不会指望这样 |
| 486 | about how raw materials are being used as leverage, etc. | 关于原材料如何被用作杠杆等等。 |
| 487 | But Logan's mega point is right. | 但洛根的特大点是正确的。 |
| 488 | Especially if you weight this by the amount of market cap at the end purchaser stage. | 特别是如果在最终购买者阶段,你用市场上限的金额来加权。 |
| 489 | Very simply, that's where most of the value is being captured, because that 652% | 很简单,这是大部分值 被捕获,因为652% |
| 490 | is weighted by massive market caps for the firms at that juncture in the stage of production. | 在生产阶段的这一关头,企业的市场上限是巨大的。 |
| 491 | And then Logan is right. | 洛根是对的 |
| 492 | How much of that is then being transferred to end users, consumers? | 然后,其中有多少被转移到终端用户、消费者手中? |
| 493 | So end purchasers are at the interface between chips and products that we use. | 因此,终端购买者处于芯片和我们使用的产品之间的接口。 |
| 494 | And that's the next stage of analysis. | 这是下一阶段的分析。 |
| 495 | But we're getting at least to that stage. | 但我们至少要进入那个阶段。 |
| 496 | Let's go to the next slide. | 下一张幻灯片 |
| 497 | And this is an example. | 这是一个例子。 |
| 498 | And this is suggestive of the proposition that even for Apple, | 这暗示了一个提议 即使对苹果来说 |
| 499 | we've got this interesting situation where maybe they have, for their low-end phones, | 我们有一个有趣的情况 也许他们有, 对于他们的低端手机, |
| 500 | maybe they have $150 worth of semiconductors. | 也许它们有价值150美元的半导体 |
| 501 | They are able to complement that with other features in the iPhone | 他们可以用iPhone中的其他功能来补充 |
| 502 | and sell that iPhone for $700 plus. | 卖掉那台iPhone七百块 |
| 503 | So this is a case where I think we would say we see the value being captured by end purchasers. | 因此,这个案例,我想我们会说,我们看到最终购买者捕捉到的价值。 |
| 504 | If you're the leading supplier of individual components here, | 如果你是这里单个部件的主要供应商 |
| 505 | you're not charging so much to capture a bigger share of the final price. | 你没有收这么多钱来占最后价格的更大份额 |
| 506 | I know we've got a lot of interesting questions. | 我知道我们有很多有趣的问题 |
| 507 | Let's see if we can cover a few more things, and then we'll open it up to questions. | 试问能否再盖几件事情,再开解问. |
| 508 | Individual firms that are capturing the most value. | 获得最大价值的个别公司。 |
| 509 | If you just focus on the left-hand side, those names are ranked by market cap, | 如果你只专注于左侧, 这些名字按市场上限排序, |
| 510 | with Intel and Samsung being the highest at $163 billion and a company like SK Hynex at $18. | 英特尔和三星是最高的1,630亿美元,还有像SK Hynex这样的公司是18美元. |
| 511 | Well, the change in the rankings is dramatic. | 排名的变化是惊人的。 |
| 512 | Firms that were big market cap back 10 years ago, it's not necessarily going to carry forward 10 | 早在十年前就已经是大市场封顶的公司 不一定能继续发展下去 |
| 513 | years later. | 数年后. |
| 514 | And you've got a really interesting phenomenon with Qualcomm doing well, TSMC that | 你有一个非常有趣的现象 Qualcomm做得很好,TSMC |
| 515 | mentioned Intel falling way off. | 提到英特尔坠落 的方式。 |
| 516 | So as they say in the information disclosures, | 正如他们在信息披露中所说的, |
| 517 | past success is not a predictor of future success. | 过去的成功并不是对未来成功的预测。 |
| 518 | And if we go to the next slide, we've just got some of those names highlighted. | 如果我们去下一张幻灯片, 我们刚刚得到了一些这些名字的突出。 |
| 519 | TSMC, a ninefold increase in value. | TSMC,价值增加了9倍. |
| 520 | Now, why is that? | 为什么? |
| 521 | What's going on? | 这是怎么回事? |
| 522 | What explains these divergences? | 何以解释这些分歧? |
| 523 | And we've identified three hypotheses for that. | 我们为此确定了三个假设。 |
| 524 | One is | 一个是 |
| 525 | compete and win. | 竞争和胜利。 |
| 526 | This is a highly rivalrous industry. | 这是一个高度竞争的行业。 |
| 527 | Firms compete continuously, | 企业不断竞争, |
| 528 | and they can win and they can lose. | 他们可以赢,他们可以输。 |
| 529 | And they compete on the development of new products. | 它们在开发新产品上竞争。 |
| 530 | For those of you who likes global football, it's a game where the players have to | 对于你们中喜欢全球足球的人来说,这是一场球员必须参加的比赛. |
| 531 | anticipate where the ball is going and how to position themselves and how to react to that. | 预测球的去向 如何定位 如何反应 |
| 532 | It's extraordinarily dynamic. | 这是异常动态的。 |
| 533 | And then some firms are better than that than others. | 然后,一些公司比其他公司要好。 |
| 534 | Some firms have good foresight, some firms may have good foresight, but then they're unlucky | 有些公司很有远见,有些公司可能很有远见,但后来却不走运 |
| 535 | and things don't develop the way they expected. | 事情没有发展 他们的期望。 |
| 536 | And then the second hypothesis is firms that get | 第二种假设是 公司得到 |
| 537 | ahead stay ahead because of so-called competitive moats. | 因为所谓的竞争性护城河,所以先走 |
| 538 | And the third hypothesis is, well, | 第三个假设是,嗯, |
| 539 | the firms with high market shares must be doing something bad. | 市场份额高的公司一定在做坏事 |
| 540 | And if we go to the next slide, | 如果我们去下一个幻灯片, |
| 541 | when we look at the case studies so far, for example, dial in on memory chips and look at | 当我们看到目前为止的案例研究, 例如,拨打内存芯片和看 |
| 542 | what's happened to Samsung versus SK Hynex and Micron, when we look at foundries and we can look | 三星对SK Hynex和Micron的遭遇 当我们看着铸币局,我们可以看看 |
| 543 | at what's happened to TSMC versus Intel, when we look at the success of ASML, we strongly favor | TSMC对英特尔的遭遇 当我们看ASML的成功 我们强烈赞成 |
| 544 | hypothesis one. | 假设一。 |
| 545 | If the moats were so big, we would see persistence and the incumbents who have strong | 如果护城河这么大,我们会看到坚韧不拔和坚固的 |
| 546 | value persisting, but we don't see that. | 价值持续,但我们看不出来。 |
| 547 | So we're not saying that incumbents don't have some | 所以,我们不是说 现任者没有一些 |
| 548 | advantages, but the nature of competition generates a lot of winners and losers. | 优势,但竞争的性质产生大量的赢家和输家. |
| 549 | And who | 谁 |
| 550 | are the winners 10 years from now, probably will be different from the ones who are winning now. | 10年后的获胜者可能与现在的获胜者不同 |
| 551 | That's important, yes. | 这很重要,是的。 |
| 552 | All right, Logan, I'm talking too much, but do we need to go through this? | 好吧,洛根,我说得太多了, 但我们需要经历这一切吗? |
| 553 | I mean, a little bit, | 我的意思是,一点点, |
| 554 | maybe the steps in reaching a contract? | 或许是达成合同的步骤? |
| 555 | We have a few more slides to cover and we'll still do a Q&A. | 我们还有几张幻灯片要盖 我们还会做一个QQA |
| 556 | So I think let's give us a little bit | 所以,我想让我们 一点点 |
| 557 | of attention and keep our momentum up. | 并保持我们的势头。 |
| 558 | There's some important points to cover ahead. | 前面还有一些重要的点要覆盖. |
| 559 | Yeah, I mean, just the key point here, everybody who's tuning in, is these contract processes have | 是啊,我的意思是,这里的关键点, 每个人谁的调整, 这些合同程序有 |
| 560 | multiple stages and they're long. | 多个阶段,它们很长。 |
| 561 | And when that's all done, they have a duration and then the whole | 当这一切完成, 他们有一个期限,然后整个 |
| 562 | process starts again with renewal. | 进程从更新开始。 |
| 563 | So if we go to the next slide, I've asked myself the question, | 所以,如果我们去下一个幻灯片, 我已经问自己的问题, |
| 564 | well, okay, compete and win. | 好吧,好,竞争和胜利。 |
| 565 | Where is it that firms can win or fail? | 企业能赢能败能败,何在? |
| 566 | And it's multiple stages, | 这是多个阶段, |
| 567 | qualification, bidding stage, the collaboration between customer and supplier is really important. | 资格、投标阶段、客户和供应商之间的合作非常重要。 |
| 568 | Each side is making specific investments and getting ready for new products. | 各方都在进行具体投资,为新产品做好准备。 |
| 569 | That may go well, it may not go well. | 这也许进展顺利,但可能不顺利。 |
| 570 | And then a supplier who's hoping to get a single source | 然后一个供应商 谁希望得到一个单一的来源 |
| 571 | contract may fail at that. | 合同可能就此失效。 |
| 572 | And even if they do, they may fail at execution, which means production | 即使他们这样做,他们可能失败 执行,这意味着生产 |
| 573 | and delivery. | 和交付。 |
| 574 | So there are a lot of points where firms can do better and do not so well. | 所以有很多观点 公司可以做得更好 也不好。 |
| 575 | Next slide. | 下一个幻灯片。 |
| 576 | So I mentioned this, incumbency, we believe has advantages. | 所以我提到这一点, 在职,我们认为有优势。 |
| 577 | For those of you who took the class, | 你们中那些上过课的人, |
| 578 | I mean, we emphasized learning by doing, switching costs, the role of IP. | 我的意思是,我们强调学习 通过做,转换成本, IP的作用。 |
| 579 | So all of those things are | 所以所有这些事情 |
| 580 | advantages for incumbents. | 对任职者的好处。 |
| 581 | But the key point here is number four. | 但这里的关键是第四点。 |
| 582 | Incumbency does not guarantee | 任职不能保证 |
| 583 | future success. | 未来成功。 |
| 584 | I mean, Apple changed suppliers for the modem chips several times. | 苹果公司多次更换调制解调器芯片供应商. |
| 585 | Samsung lost | 三星输了 |
| 586 | their edge with memory chips. | 他们的边缘与记忆芯片。 |
| 587 | Intel was such an important iconic firm with foundries and they've | 英特尔是这么重要的标志性公司 有铸币局,他们已经 |
| 588 | lost their edge to TSMC. | 失去他们的边缘 到TSMC。 |
| 589 | So do we think incumbency has some advantages? | 那么,我们是否认为在职有一些优势? |
| 590 | Sure, but not far from a | 当然,但离这里不远 |
| 591 | guarantee of success. | 保证成功。 |
| 592 | And Ted, just to tack onto this point and what you were mentioning earlier | 还有Ted 刚才你提到的事 |
| 593 | about the leaders of the future 10 years from now, likely not the leaders today. | 未来十年的领袖们 可能不是今天的领袖们 |
| 594 | An example we use | 我们用的例子 |
| 595 | is if you look at the market cap of the top 20 tech names in the dot com era versus say 2010 versus | 如果你看看20个技术名的市场封顶 相对于2010年 |
| 596 | even 2015 and 2025 now, you'll see a lot of different names that have come to the floor | 即使是2015年和2025年,你也会看到很多不同的名字 来到地上 |
| 597 | and have faded. | 并逐渐消失。 |
| 598 | And so tech moves with tremendous velocity, unlike some other industries that are | 因此技术的移动速度非常快 不像其他工业 |
| 599 | maybe more legacy. | 也许更多的遗产。 |
| 600 | And one of the reasons why it's so fascinating to us is that we see these | 我们之所以如此着迷的原因之一 就是我们看到这些 |
| 601 | huge changes in valuation. | 估值发生了巨大变化。 |
| 602 | And which is why Ted goes into, it's not a guarantee of future success, | 这就是为什么Ted去, 它不能保证未来的成功, |
| 603 | just because you're there and you feel like your firm is at the top has no bearing on 20 years from | 只是因为你在那里,你觉得你的公司 在顶端 与20年后没有关系 |
| 604 | now, you still being the leader in your respective space because of the adoption cycles, the | 现在,你仍然是你各自空间的领导者 因为收养周期, |
| 605 | innovation, the pace of innovation, and how much investment, let's not forget this industry, this | 创新,创新的速度, 和多少投资, 让我们不要忘记这个行业,这个 |
| 606 | broadly in TMT, not just semis, but broadly in technology, consumes more investment than almost | 一般来说,在TMT中,不仅半成品,而且在技术中,消耗的投资比几乎要多。 |
| 607 | anything else on earth. | 地球上的任何东西。 |
| 608 | And so you have a tremendous interest from all governments and countries and | 所以,你对政府和国家的极大兴趣 |
| 609 | firms, et cetera, pushing these innovations forward into the marketplace, Ted. | 公司,等等,把这些创新推向市场,泰德。 |
| 610 | So perfect | 这么完美 |
| 611 | comments there. | 评论。 |
| 612 | Let's cover two more slides and then we'll turn to questions. | 我们再盖两张幻灯片 然后我们再谈谈问题 |
| 613 | Given what Logan | 考虑到什么洛根 |
| 614 | said, one of the questions we posed at the outset was, well, what do market shares mean? | 说,我们一开始提出的问题之一是,市场份额意味着什么? |
| 615 | And first of all, there are snapshots in time. | 首先,及时出现快照。 |
| 616 | Another thing to obviously recognize is that | 另一件显然需要承认的事是: |
| 617 | they basically aggregate successes in gaining individual contracts. | 他们基本上在获得个人合同方面取得了成功。 |
| 618 | And given what we said | 鉴于我们说过的话 |
| 619 | earlier, they're sensitive to changes in so-called ex ante before the contracts are written, | 之前,他们很敏感 在合同写成之前 所谓的事前变更 |
| 620 | competition. | 竞争。 |
| 621 | I think of these market shares as outcomes of rivalry, which firms won, which firms | 我认为这些市场份额是竞争的结果 公司赢得了竞争 |
| 622 | lost. | 输了 |
| 623 | It's that simple. | 就这么简单 |
| 624 | They're outcomes at any point in time. | 它们在任何时候都是结果. |
| 625 | And then when we see sustained | 然后当我们看到持续 |
| 626 | high market shares, that's going to be due to several things. | 高的市场份额, 这将是由于一些事情。 |
| 627 | Continued success with | 继续取得成功 |
| 628 | winning the contracts, successful execution. | 赢得合同,成功执行。 |
| 629 | There's a lot of foresight. | 诸有前观. |
| 630 | Where are the new | 何处是新 |
| 631 | products going? | 产品在走? |
| 632 | And do you invest properly in capabilities to supply those new products? | 你是否对供应这些新产品的能力进行了适当的投资? |
| 633 | So I and Logan, I think we would summarize this on the next slide, which is firms that win | 所以我和罗根,我想我们会在下一张幻灯片上总结一下, 这是公司赢得 |
| 634 | have foresight, they make good investment decisions, and they execute. | 他们有远见,他们做出良好的投资决定,并且执行。 |
| 635 | And when you're | 当你是 |
| 636 | analyzing firms and you're saying, well, how are they doing? | 分析公司,你说, 他们怎么样? |
| 637 | You want to look at their market shares, | 你想看看他们的市场份额, |
| 638 | right? | 对吧? |
| 639 | But you also want to realize, given the dynamics of this industry, you want to look at, | 但你也想意识到,考虑到这个行业的动态, 你想看看, |
| 640 | well, what are they saying about the future? | 那么,他们在说什么 关于未来? |
| 641 | Are they getting that right? | 他们听懂了吗? |
| 642 | Are they making | 他们正在制作 |
| 643 | the right bets? | 正确的赌注? |
| 644 | Are they making the right investments and capabilities? | 他们是否做出了正确的投资和能力? |
| 645 | And are they | 他们是 |
| 646 | going to be able to execute? | 能够执行吗? |
| 647 | So this is a phrase that actually has been around in the management | 因此,这是一个短语 实际上一直围绕在管理层 |
| 648 | literature for decades. | 几十年的文学。 |
| 649 | But we think it's probably should resurface. | 但是我们认为它应该重新浮出水面。 |
| 650 | And it's a good way to capture. | 且道妙法摄取. |
| 651 | It's a framework for analyzing how individual companies are doing. | 这是一个分析个别公司如何工作的框架. |
| 652 | And with that, I think let's pause. | 这样的话,我们暂停一下 |
该视频共有字幕 652 条。解锁更多字幕为会员功能，请移动到 价格