# Re-engineering the Semiconductor Supply Chain with Intel CEO Lip Bu Tan

https://www.youtube.com/watch?v=asCgCv2XB4s
Translation: zh-CN

[00:00] Nine of the 10 companies I invest halfway, they change their business plan.
  我投资的10家公司中，有9家公司会中途改变他们的商业计划。

[00:04] Because the market has changed.
  因为市场已经改变了。

[00:05] So I like to have entrepreneurs as a team, not just one person.
  所以我喜欢团队中有企业家，而不仅仅是一个人。

[00:10] I always believe in when I was at Cadence and also at Intel, is first of all you crawl and then be humble, listen to the customer.
  我一直相信，在我还在Cadence和Intel的时候，首先是爬行，然后是谦虚，倾听客户。

[00:18] And then first step for me is to strengthen my balance sheets, focus on the products.
  然后我的第一步是加强我的资产负债表，专注于产品。

[00:23] And I really simplify the product, listen to the customer, and then drive the next generation leadership products.
  我真正地简化产品，倾听客户，然后推动下一代领导产品。

[00:29] And then right now the agentic AI and influence CPU become, you know, highly in demand.
  然后现在，你可以知道，代理AI和影响力CPU变得非常受欢迎。

[00:34] And so in some way I'm happy right now the demand is very high for my CPU.
  所以某种程度上我很开心，现在我的CPU需求量非常高。

[00:39] Certainly very happy that Jensen Huang, my old-time friend, he also put 5 billion uh in investing and support me.
  当然非常高兴我的老朋友黄仁勋也投资了50亿美元来支持我。

[00:45] 5 billion become 25 billion now.
  50亿现在变成了250亿。

[00:48] If you look at it 10 years from now, what will be the winning company?
  如果你展望未来10年，哪家公司会是赢家？

[00:51] The one that
  那家公司

[01:00] Hi listeners welcome back to no prior.
  大家好，欢迎回到 No Prior。

[01:03] Today Allad and I are here with Lehutan.
  今天，Allad 和我与 Lehutan 在一起。

[01:05] The legendary investor from Walden, then CEO of Cadence, now CEO of Intel.
  这位来自 Walden 的传奇投资者，曾是 Cadence 的首席执行官，现在是 Intel 的首席执行官。

[01:07] We talk about his plan to transform Intel.
  我们谈论了他重塑 Intel 的计划。

[01:13] Having the US government as a major shareholder.
  让美国政府成为主要股东。

[01:15] How to be an amazing semiconductors investor, and whether or not we can make chips in the United States.
  如何成为一名出色的半导体投资者，以及我们是否能在美国制造芯片。

[01:20] Welcome, Lipu.
  欢迎，Lipu。

[01:23] Lipu, it's great to see you.
  Lipu，很高兴见到你。

[01:25] We'll start with the obvious question.
  我们先从显而易见的问题开始。

[01:27] This is a really hard job to go be CEO of this incredibly important um American semis company.
  担任这家极其重要的美国半导体公司的首席执行官，这真是一项艰巨的任务。

[01:32] Why take the job at all?
  为什么要接受这份工作呢？

[01:33] It's a good question.
  这是个好问题。

[01:36] I'm 66 and people that well you should retire rather than take on this hottest job in the industry.
  我66岁了，人们觉得我应该退休，而不是接受这个行业里最热门的工作。

[01:41] And so a couple reason one is um this is iconic company and it's so important for the semiconductor ecosystem and also so important for United States.
  所以有几个原因，一是，这是一个标志性的公司，对半导体生态系统至关重要，对美国也至关重要。

[01:53] And so I decided you know do one more after cadence.
  所以我决定，你知道的，在 Cadence 之后再做一件事情。

[01:57] A lot has happened in this past year.
  在过去的一年里发生了很多事情。

[01:59] What has been the most surprising to you?
  什么让你感到最惊讶？

[02:02] Well the most surprising thing that I don't learn from my previous job or even training is one day early morning President Trump asking me to resign and conflict of interest and there's no exceptions and so I had to convince myself first of all you know I don't need this job I do it purely to save Intel and so take that personal issue out of the way then I figure out what what can I do to be helpful to Intel and so good news is I have a meeting you know Thursday morning and then Monday I have the meeting and then he listen to me like I have a chance to explain myself you know born in Malaysia grown up in Singapore went to MIT and I live in US and never live outside country and so something that I share and then somehow he listened very well and then uh he gave me the chance and so I'm delighted and now you have the chance to do the work um when you said you know the the job is to save Intel It's a really
  嗯，我从之前的工作或培训中学到的最令人惊讶的事情是，有一天一大早，特朗普总统要求我辞职，理由是利益冲突，而且没有例外，所以我不得不先说服自己，你知道，我不需要这份工作，我这样做纯粹是为了拯救英特尔，所以把这个个人问题放在一边，然后我想出我能做什么来帮助英特尔，所以好消息是，周四早上我有一个会议，然后周一我有一个会议，然后他听我说，就像我有一个机会向他解释我自己，你知道，出生在马来西亚，在新加坡长大，去了麻省理工学院，我住在合众国，从未在国外生活过，所以这是我分享的一些事情，然后不知何故他听得很好，然后呃，他给了我机会，所以我很高兴，现在你有机会做这项工作，嗯，当你说的，你知道，这份工作是拯救英特尔，这真的很

[03:03] Important company. What what does that look like to you?
  重要的公司。在你看来，那是什么样的？

[03:05] What does Intel winning or thriving look like?
  英特尔获胜或蓬勃发展是什么样的？

[03:07] Yeah, I just passed 14 months.
  是的，我刚过14个月。

[03:10] A lot of thing happened in this 14 months.
  这14个月里发生了很多事情。

[03:12] So, a couple of thing. One is to change the culture and then uh clearly want to drive more accountability and also in term of decision making had to be faster.
  所以，有几件事。一是改变文化，然后明确地想要提高问责制，在决策方面也必须更快。

[03:22] You know I'm so used to startup culture and you move fast in the speed of light and going to have that bureaucracy layer of layer of meeting and so something that I change the accountability listen to the customer and the customer delighted you know someone like liberal so humble willing to listen and then address some of the problem that they face and then try to delight the customer and also the other part from day one I decided all the engineering report to me I'm being an engineer by training I want to know what went wrong and what are the thing that I need to correct.
  你知道我太习惯初创公司的文化了，你们以闪电般的速度前进，然后就会有官僚主义一层一层的会议，所以我改变了问责制，倾听客户的声音，让客户满意，你知道，像自由主义者一样谦虚，愿意倾听，然后解决他们面临的一些问题，然后努力让客户满意，还有另一部分，从第一天起，我就决定让所有的工程报告都给我，我受过工程师的训练，我想知道哪里出了问题，以及我需要纠正哪些事情。

[03:56] Listen to the customer and delight the customer and then make sure that we have the right product.
  倾听客户的声音，让客户满意，然后确保我们有正确的产品。

[04:00] Simplify our product line and really
  简化我们的产品线，真正地

[04:05] Have the root map and the vision for the next five 10 years.
  拥有根基蓝图以及未来五到十年的愿景。

[04:08] What is your vision of where Intel should be in 10 years?
  您对英特尔十年后的发展愿景是什么？

[04:11] Yeah, I think couple of things. One I always believe in when I was at Cadence and also at Intel is first of all you crawl and then be humble. listen to customer and then secondly deciding to walk and then finally deciding to run in spring.
  是的，我认为有几件事。一是我在 Cadence 和英特尔时一直相信的，首先是爬行，然后保持谦逊，倾听客户，然后其次决定行走，最后决定奔跑。

[04:26] So that's kind of my culture of step by step doing it and then first step for me is to strengthen my balance sheets.
  所以这就是我一步一步做的文化，对我来说第一步是加强我的资产负债表。

[04:34] And uh the balance sheet is really have horrible in some way.
  而且，资产负债表在某种程度上确实很糟糕。

[04:37] So I'm delighted you know US government become a big shareholder just I explained to President Trump TSMC when they started they have the D Taiwan government as a shareholder.
  所以我很高兴，你知道美国政府成为大股东，就像我向特朗普总统解释的那样，台积电刚开始时，他们有台湾政府作为股东。

[04:48] If you look at Japan, you look at Singapore, this is the infrastructure US government get to provide the support.
  如果你看看日本，看看新加坡，这就是美国政府提供支持的基础设施。

[04:54] Secondly, very happy that Jensen Hong, my oldtime friend uh he also put five billion uh in investing and support me and I'm glad I at least do some good work.
  其次，我很高兴我的老朋友黄仁勋也投资了五十亿美元来支持我，我很高兴我至少做了一些好事。

[05:04] His five
  他的五亿

[05:06] billion become 25 billion now and uh or more.
  十亿变成现在的250亿，或者更多。

[05:11] And then the other part is Soft Bank Master.
  然后另一部分是软银大师。

[05:14] I used to be at Soft Bank board and then he lend a hand to help me.
  我曾经在软银董事会，然后他伸出援手帮助我。

[05:16] So we strengthen the balance sheet and then uh focus on the products and I really simplify the product listen to the customer and then drive the next generation leadership products.
  所以我们加强了资产负债表，然后专注于产品，我真正简化了产品，倾听客户的声音，然后推动下一代领导产品。

[05:28] And then um in some way it's very lucky uh right now the agentic AI and influence CPU become you know highly in demand and so you know versus one to eight in training CPU to GPU.
  然后，在某种程度上，现在非常幸运，代理AI和影响CPU变得非常受欢迎，你知道，与训练CPU到GPU的1比8相比。

[05:41] Now I can see one to four maybe one to.
  现在我可以看到1比4，也许1比。

[05:47] And I'm delighted CPU become important.
  我很高兴CPU变得重要。

[05:50] I talked to some of the AI model and developer.
  我与一些AI模型和开发者进行了交谈。

[05:52] They said about in term of uh reinforced learning uh in term of the speed of orchestrating all the agents and turn up the CPU is actually is better.
  他们说在强化学习方面，在编排所有代理的速度方面，CPU实际上更好。

[06:03] And so in some way I'm happy right now the demand is very high for my
  所以，在某种程度上，我很高兴，现在我的需求非常高。

[06:07] CPU.
  中央处理器。

[06:10] So I think overall build on the product on the data center server side.
  所以，我认为总体上是在数据中心服务器方面构建产品。

[06:12] Then the other part is our foundry business and initially this is a capital intensive business and it's not easy and you really need to have couple of thing.
  然后另一部分是我们的代工业务，起初这是一个资本密集型业务，而且不容易，你真的需要有几样东西。

[06:22] You need to have all the right IP so that you can support the customer like for example if it is a mobile related you got to have low power uh IP set that you need to have without that you cannot serve them.
  你需要拥有所有正确的知识产权，这样你才能支持客户，例如，如果它与移动相关，你必须拥有低功耗的知识产权集，没有它你就无法为他们服务。

[06:34] It's a service business it's a trust business if people want to give you you know orders to have wafer to If the yield not good they will be toast in term of revenue miss.
  这是一项服务业务，这是一项信任业务，如果人们想给你下订单生产晶圆，如果良率不高，他们将在收入方面遭受损失。

[06:45] So with that I think it's very important to really focus on the yield the defect density the cycle time and then make sure that you really able to meet and serve the customer in high quality and reliable.
  所以，我认为非常重要的是要真正关注良率、缺陷密度、周期时间，然后确保你真的能够高质量、可靠地满足和为客户服务。

[07:01] And so those are the thing that I really focus on it and eventually you have to really move into a full stack.
  所以，这些是我真正关注的事情，最终你必须真正转向全栈。

[07:05] So not
  所以，不

[07:07] just a silicon you need to have a software and some of the customer asked me give me the whole rack.
  仅仅是硅，你需要有软件，有些客户问我，给我整个机架。

[07:12] So there's a system that you have to build and so I think those are the thing that I quietly building uh step by step and recruit some of the best talent I can find.
  所以你必须建立一个系统，我认为那些是我一直在悄悄地一步一步建立起来的事情，并招募我能找到的最优秀的人才。

[07:21] By the way all the rec recruitment I do it myself no search firm helping and so I think sometime it's good to have a roer deck that you know who to reach out to call for.
  顺便说一句，所有的招聘我都自己做，没有猎头公司帮忙，所以我想有时候有一个人脉网络会很好，这样你就知道该联系谁。

[07:30] Yeah.
  是的。

[07:32] I mean, you've been in the business for so long and you know, you've you've run a cadence for I think 12 years before this and so 13 years.
  我的意思是，你在这个行业已经很久了，你知道，在此之前你已经经营了大约12年，所以是13年。

[07:39] I'm sorry.
  抱歉。

[07:41] Yeah.
  是的。

[07:42] And then two more years as executive chairman.
  然后又担任了两年执行主席。

[07:45] So 15 years I signed up for three months.
  所以15年，我报名了三个月。

[07:49] So right now I be very careful the moment you said just do it for three months.
  所以现在我要非常小心，你一说就做三个月。

[07:51] It turned out to be 15 years.
  结果是15年。

[07:53] Yeah.
  是的。

[07:54] Well it seems like you have a lot of longevity ahead of you here as well.
  嗯，看起来你在这里也有很长的职业生涯。

[07:56] And so um the other big initiative that that has been sort of talked about is Terafab and working with Elon Musk on that.
  所以，嗯，另一个被谈论过的重大举措是Terafab以及与埃隆·马斯克在这方面的合作。

[08:00] Can you tell us a bit more about how that came together and your involvement and how you all are collaborating?
  你能多告诉我们一些关于它是如何形成的，你的参与以及你们是如何合作的吗？

[08:06] Yeah, good.
  是的，很好。

[08:06] I mean, Elon Mus I think we
  我的意思是，埃隆·马斯克，我想我们

[08:08] all agree is one of the best if not the best uh entrepreneur in this century.
  大家都认为他是本世纪最优秀的创业者之一，如果不是最优秀的话。

[08:15] He and I we share the same view that the semiconductor infrastructure actually is not catch up with the AI growth and uh in term of you need the capacity you need to have the productivity and you have the dry efficiency and so those are the thing that he and I we share that there's some something missing.
  我和他有同样的看法，即半导体基础设施实际上跟不上人工智能的增长，而且就你需要的产能、你需要具备的生产力以及你的干效率而言，这些都是我和他共同认为缺少的东西。

[08:34] and then secondly he just delighted to work with him and uh he is very I call it unconventional.
  其次，很高兴能与他共事，而且他非常，我称之为非传统。

[08:42] and he basically question every step and then why this traditional way of doing things and in some way it's very refreshing and I like that you know I like people have different opinion and let's work together find what is the best route and we both going to learn a lot together.
  他基本上质疑每一步，以及为什么用这种传统的方式做事，在某种程度上这非常令人耳目一新，我喜欢这样，你知道我喜欢人们有不同的观点，让我们一起合作找到最佳途径，我们都会一起学到很多东西。

[08:57] and then I think clearly he have a vision that his robots and his car you know he need a lot of silicon.
  然后我认为他显然有一个愿景，他的机器人和他的汽车，你知道他需要大量的硅。

[09:05] yeah could you actually explain what terapab is for people who aren't familiar with it and
  是的，你能为不熟悉的人解释一下terapab是什么吗？

[09:09] Terab, he decided he wants to build his own fab.
  特拉布决定他要建造自己的工厂。

[09:11] And then meanwhile, we are delighted to work with him and then make sure that we can work together and enable him to be faster and quicker to the production.
  同时，我们很高兴能与他合作，并确保我们能够协同工作，使他能够更快、更迅速地投入生产。

[09:20] And then using some of our technology and some of our process, and that's something that we both kind of collaborate together.
  然后利用我们的一些技术和一些流程，这是我们双方共同合作的事情。

[09:26] And he's a very good team that I work with weekly, and it's just refreshing to work with him.
  他是一个我每周都合作的非常好的团队，与他合作令人耳目一新。

[09:31] And he's talked about things like he wants you to be able to smoke inside the clean room and all these things that normally are the burger.
  他谈到了一些事情，比如他希望你能在洁净室里抽烟，以及所有这些通常是汉堡的事情。

[09:39] I think I don't go that far, and maybe some part of the clean room you can do that, but I think something that is open mind, and then we are also listen and see whether we can do that.
  我认为我不会走那么远，也许洁净室的某些部分你可以这样做，但我认为有些事情是思想开放的，然后我们也会倾听并看看我们是否能做到。

[09:49] Yeah, I mean, it's very exciting to see how you're morphing in the business here in the US in terms of um incrementally building out the foundry business in terms of collaborating with things like Terrafab.
  是的，我的意思是，看到你在美国的业务如何发展，通过逐步建立代工厂业务，与特拉法布这样的公司合作，这非常令人兴奋。

[09:59] If you think about the global AI and semiconductor supply chain, so say that you were to look at the changes that AI is driving on a macro basis, country by country.
  如果你考虑全球人工智能和半导体供应链，那么假设你要从宏观角度，按国家来看人工智能正在驱动的变化。

[10:07] And if I look at
  如果我看看

[10:09] certain countries, when I look at the layoffs that are claimed from AI, for example, um most of them I think are overstated right now.
  某些国家，当我看到声称来自人工智能的裁员时，例如，我认为目前大多数都被夸大了。

[10:16] You know, most of the layoffs are actually just overhiring during 2020 period.
  你知道，大多数裁员实际上是在2020年期间过度招聘。

[10:22] But the first things I see actually being cut are outsource firms where you'd rather cut external headcount versus internal.
  但首先被削减的是外包公司，你宁愿削减外部员工而不是内部员工。

[10:27] So you're cutting external customer support.
  所以你正在削减外部客户支持。

[10:29] you're cutting external IT and that has more of an impact I think for certain countries which have big BPOS the Philippines India etc and so they may be impacted in the short run by AI.
  你正在削减外部IT，我认为这对拥有大型BPO的某些国家，如菲律宾、印度等，影响更大，因此它们可能在短期内受到人工智能的影响。

[10:39] and then if you ask how do companies participate in the future in a positive way in AI you have to almost go country by country right places with cheap energy will do data centers places with the ability to train models will train models but it's probably only the US and one or two other places.
  然后，如果你问公司将来如何以积极的方式参与人工智能，你几乎必须逐个国家地考虑，能源便宜的地方将建立数据中心，有能力训练模型的地方将训练模型，但这可能只有美国和一两个其他地方。

[10:53] um how do you think about the the shift in global supply chain for the semiconductor industry should certain countries invest more like should Israel be doing more given Milanx and the Nvidia and Intel presence there and should they try to do more in semiconductors?
  嗯，你如何看待半导体行业的全球供应链转移？某些国家是否应该投入更多，比如以色列鉴于其在Milanx、Nvidia和Intel的存在，是否应该做得更多，他们是否应该在半导体领域做得更多？

[11:04] Should other should the Philippines move back to more of a manufacturing base like how do you think about that on a global basis?
  其他国家，菲律宾是否应该回归更多制造业基础，你如何看待全球范围内的这种情况？

[11:10] Yeah, good question.
  是的，好问题。

[11:12] So I think clearly the AI is changing the whole landscape.
  所以，我认为很明显，人工智能正在改变整个格局。

[11:15] And I think the impact will be bigger than internet and uh it's more profound also.
  而且我认为其影响将比互联网更大，而且也更深远。

[11:20] So I think the AI you know in initially is able to help you to do things more efficiently.
  所以，我认为人工智能，你知道的，最初能够帮助你更有效地做事。

[11:26] And then with a lot of agent helping you uh to do things that is now kind of mundane that you need to do but now they can give it to you faster.
  然后，有大量的代理帮助你，呃，做一些现在有点单调乏味但你必须做的事情，但现在它们可以更快地给你。

[11:35] So in some way I think it can drive a lot of efficiency even like semiconductor design.
  所以，在某种程度上，我认为它可以带来很多效率，甚至像半导体设计一样。

[11:41] How much you can drive the efficiency in term of timing uh how quickly can you come out and secondly the cost.
  在时间方面你能提高多少效率，呃，你能多快出来，其次是成本。

[11:48] And so I think those will be helping you to drive that.
  所以，我认为这些将有助于你推动这一点。

[11:52] And then I think couple of bottlenecks for the AI you know demand and growth.
  然后，我认为人工智能的需求和增长存在几个瓶颈。

[11:54] One is of course everybody knows power constraint.
  一个当然是每个人都知道的电力限制。

[11:59] Some country the power they just don't have that it get impacted.
  一些国家就是没有电力，它会受到影响。

[12:05] And then secondly a lot of people didn't realize the helium impact can be also quite significant for semiconductor.
  其次，很多人没有意识到氦气的影响对半导体来说也可能相当显著。

[12:10] And
  和

[12:12] then the thirdly is everybody know right
  然后第三点是大家都知道的，对吧

[12:15] now memory is a bigger shortage and
  现在内存是一个更大的短缺，而且

[12:17] everybody try to scramble for memory and
  每个人都试图争夺内存，而且

[12:20] then even though you have to build a fab
  然后即使你必须建造一个工厂

[12:21] to capacity increase it will take couple
  来增加产能，这需要几年时间

[12:24] of years to do that and same thing for
  来完成，同样的事情也适用于

[12:26] CPU GPU and all this will be highly
  CPU、GPU，所有这些都将是高度

[12:29] demanded and I think the also the
  需求的，而且我认为

[12:32] pricing also go up because we have to
  价格也会上涨，因为我们必须

[12:34] pass the price the cost to the customer.
  将成本转嫁给客户。

[12:37] So I think those will be the impact the
  所以我认为这些将是影响

[12:39] industry growth
  行业增长的

[12:40] and then uh I think overall I felt that
  然后，嗯，我认为总的来说，我感觉

[12:42] you know the company that most impacted
  你知道，受影响最大的公司

[12:45] is you are not embracing uh AI
  是你没有拥抱人工智能

[12:48] and because AI can help you to drive a
  因为人工智能可以帮助你驱动

[12:51] lot of efficiency across all the
  在企业的所有不同职能部门实现高效率

[12:53] different function of the enterprise. we
  我们应该拥抱它，并找到更好的方法来利用人工智能

[12:56] should embrace and also find way to
  来预测、设计，你知道工作负载的所有不同部分

[12:58] better use the AI for your prediction
  我认为这是巨大的

[13:01] for your design for your you know all
  一些人会说，反对台积电

[13:04] the different part of the workload and I
  反对英特尔代工厂具有竞争力的

[13:07] think that's tremendous
  简单论点

[13:08] a number of people would say the
  一些人会说，反对台积电

[13:09] simplistic argument against Terraab
  反对英特尔代工厂具有竞争力的

[13:12] against Intel foundry being competitive
  简单论点

[13:15] is really a question of you know there's there's all the factors internal to the building right you describe um IP and velocity of just how you're doing business.
  这确实是一个问题，你知道，有建筑内部的所有因素，对吧，你描述了知识产权和业务开展的速度。

[13:25] Then there are external factors and you know Allah's talking about a number of them but one of them is the the cost of labor and actually the manufacturing capacity you know in investing in the foundry business you obviously believe there's a version where you can manufacture domestically and Elon does too can you talk a little bit about that and you know how real that constraint is the labor constraint.
  然后还有外部因素，你知道阿拉提到了其中的一些，但其中之一是劳动力成本以及实际的制造能力，你知道，投资铸造业务，你显然相信有一种模式，你可以在国内制造，埃隆也相信，你能谈谈这一点吗？你知道这个限制有多真实，劳动力限制。

[13:47] right so I think you know the when I decided whether should double down on foundry or should I get out of the foundry Mhm.
  对，所以我想，你知道，当我决定是应该加大在铸造方面的投入还是退出铸造行业时，嗯。

[13:55] And I there's a lot of voices, a lot of voices in the marketplace as you can tell.
  而且我，有很多声音，市场上的声音很多，你可以听出来。

[14:00] It's very expensive.
  这非常昂贵。

[14:01] It's very expensive not going to work.
  这非常昂贵，行不通。

[14:03] But I finally decided this is very important for United State and also very important for the industry. Mhm.
  但我最终决定，这对美国非常重要，对这个行业也非常重要。嗯。

[14:09] And I'll give you the idea that you know this uh we all live through this challenges of supply chain and uh it's
  我给你一个想法，你知道，这个呃，我们都经历过供应链的这些挑战，而且呃，它

[14:17] Very important for any of the big company in semiconductor and really have to think about the supply chains and you have to have a robust and resilient supply chain.
  对于任何大型半导体公司来说，这一点都非常重要，而且确实需要考虑供应链，并且您必须拥有一个强大而有弹性的供应链。

[14:28] You cannot just depend on one or two player in different geographic goal and so I think you know the more and more people going to realize making in United States is critical and then the most advanced process like for example we have the you know 14A is like 1.4 4 nanometer and we already starting to plan for 1 nanometer and 0.7 nanometer.
  您不能仅仅依赖不同地理区域的一两个参与者，所以我想，您知道，越来越多的人会意识到在美国制造至关重要，然后是最先进的工艺，例如，我们有，您知道，14A，就像1.4 4纳米，我们已经开始计划1纳米和0.7纳米。

[14:48] It's getting smaller and smaller.
  它变得越来越小。

[14:52] So in a way it's much like our hair so thin.
  所以从某种意义上说，它很像我们的头发一样薄。

[14:54] So it's a lot of complexity.
  所以它有很多复杂性。

[14:57] It's not that easy to do and every step if you make a mistake that you just go down uh you know go down the drain.
  这并不容易做到，而且每一步，如果你犯了错误，你就会，你知道，付诸东流。

[15:04] So in some way you have to be really precise and in that manufacturing.
  所以从某种意义上说，你必须在制造中非常精确。

[15:08] So in some way this is become more and more going to be the bottleneck.
  所以从某种意义上说，这正变得越来越成为瓶颈。

[15:13] Mhm.
  嗯。

[15:14] So we felt that we you know we have a lot of respect for TSMC.
  所以我们觉得，你知道，我们非常尊重台积电。

[15:16] We are great
  我们很棒

[15:18] partner and then the more important we both need to have more capacity to serve the customer.
  合作伙伴，然后更重要的是，我们需要有更多的能力来服务客户。

[15:25] and then uh so I think we decided bite the bullet longer term I think is critical and then that's where I can create more value for the industry.
  然后，嗯，所以我想我们决定咬紧牙关，从长远来看，我认为这是至关重要的，在那里我可以为行业创造更多价值。

[15:34] People have been talking for a long time about eventually um hitting a point of resolution where you can't really uh miniaturaturize things further like the line with width just gets too small to uh be able to uh keep going.
  人们一直在谈论最终达到一个分辨率点，在那里你无法真正地进一步缩小尺寸，比如线宽变得太小而无法继续。

[15:44] Uh when do you think we actually hit that limit?
  嗯，你认为我们什么时候会真正达到那个极限？

[15:49] Good question.
  好问题。

[15:52] So I think I can see you know right now we have 18A and then now to going the production of 14A I can see 10 and seven and so I think that path I think we can get there but going to be more and more expensive and more difficult to do.
  所以我想我可以看到，你知道，现在我们有18A，然后现在正在生产14A，我可以看到10和7，所以我想这条路，我认为我们可以做到，但会越来越昂贵，越来越困难。

[16:05] and that's why we need partners we cannot just do it oursel alone partner with the subscript vendor partner with equipment vendors so that make sure that we can really drive those yield and performance.
  这就是为什么我们需要合作伙伴，我们不能单打独斗，与订阅供应商合作，与设备供应商合作，这样才能确保我们真正提高产量和性能。

[16:16] and then the other part also really become the bottleneck is packaging the
  然后另一部分也确实成为瓶颈的是封装，

[16:21] advanced packaging and so we all know about co by TSMC.
  先进封装，所以我们都知道台积电的共晶。

[16:26] now we have a really good one called e-ip that is a really next generation.
  现在我们有一个非常好的叫做e-ip，这是一个真正的下一代。

[16:30] I had to make sure that it become able to do in the production yield that meet the customer requirement.
  我必须确保它能够达到满足客户要求的生产良率。

[16:37] and now see starting to run out of steam like you describe.
  现在它开始像你描述的那样后继乏力了。

[16:40] so right now I also look at some new material.
  所以现在我也在看一些新材料。

[16:42] so become going back to the material size or the chemical table.
  所以回到材料尺寸或化学元素表。

[16:47] so guardian nitrite, silicon carbide and Indian fastfite.
  所以氮化镓、碳化硅和砷化镓。

[16:49] So I invest in all three.
  所以我在这三者上都进行了投资。

[16:54] And then looking at some of this new material, how can we really drive that?
  然后看看这些新材料，我们如何真正推动它？

[16:58] And then in term of packaging, I starting to invest into glass.
  然后就封装而言，我开始投资玻璃。

[17:02] Glass is a very good heat insulator.
  玻璃是一种非常好的绝缘体。

[17:05] So we I invest a venture site called 3DGS.
  所以我们投资了一个名为3DGS的风险投资公司。

[17:08] Then I realized that Intel we have like 1,000 pattern on the module.
  然后我意识到英特尔在模块上有大约1000个图案。

[17:11] So how the you know subscript and the module put it together and then we just announced a big program with Indian government to manufacturing in in India.
  那么，你知道的子脚本和模块如何组合在一起，然后我们刚刚宣布了一个与印度政府合作的大型计划，在印度进行制造。

[17:22] Plus in US in New Mexico.
  在美国新墨西哥州也有。

[17:25] So I think this advanced packaging very important.
  所以我觉得这种先进的封装非常重要。

[17:28] I also starting to look at artificial diamond.
  我也开始关注人造金刚石。

[17:31] And that's another very good you know insulator.
  这是另一个非常好的绝缘体。

[17:33] So I also invest into uh you know diamond foundry.
  所以我还投资了金刚石铸造厂。

[17:38] And that's something is the next generation to look at.
  这是下一代值得关注的产品。

[17:39] So new material, new subscript material and new you know uh design methodology to drive that.
  所以新材料，新的子材料以及新的设计方法来推动它。

[17:48] So one thing good about being an engineers you always hitting the wall then you find way to either jump over the wall or you work around the wall and then uh to get to the better result.
  所以作为工程师的好处是，你总是会遇到障碍，然后你会找到翻越障碍的方法，或者绕过障碍，然后取得更好的结果。

[18:01] And that's what I being uh have been long time as a investor and building semiconductor from the EDA tool to design to manufacturing.
  这就是我作为投资者，从EDA工具到设计再到制造，长期以来一直在做的事情。

[18:10] It's kind of nice to have that experience now I can help find a way to make a small contribution to the industry.
  拥有这样的经验现在感觉很好，我可以帮助为这个行业做出一点贡献。

[18:16] Yeah, I it's very exciting and one of the reasons I'm asking about it as well is to your point there's always some things that you can vent around but there are also physical limits where
  是的，这非常令人兴奋，我问这个问题的原因之一也是因为你的观点，总有一些事情你可以变通，但也有物理极限，在那里

[18:23] Once you hit seven, inst, whatever the limitation is, you start to run into.
  一旦你达到七个，安装，无论限制是什么，你都会开始遇到。

[18:27] You need to find new materials or find other workarounds.
  你需要找到新材料或找到其他解决方案。

[18:31] And then the interesting question is, uh, and we've been talking about this for a long time.
  然后有趣的问题是，呃，我们已经谈论了很长时间了。

[18:33] I remember 20 years ago people were talking about how we'd eventually hit this, hit a point where we ran out of space on this.
  我记得20年前人们就说过，我们最终会达到这一点，达到一个我们用完空间的地步。

[18:38] Is do you run into some sort of asymptote that actually normalizes performance across different foundries or not?
  你是否会遇到某种渐近线，实际上可以使不同代工厂的性能正常化，还是不会？

[18:44] Yeah, good question.
  是的，好问题。

[18:46] In term of like most law is a know double you know.
  就摩尔定律而言，你知道，翻倍你知道。

[18:49] Yeah.
  是的。

[18:50] And then the power and the cost and then you can double the performance but you cannot double down on the cost and and area.
  然后是功耗和成本，然后你可以将性能翻倍，但你不能在成本和面积上加倍。

[19:00] So those are the thing you have to give, give way unless you find some new way of material, new way of design and then become material science.
  所以这些是你必须放弃的东西，除非你找到一些新的材料方法，新的设计方法，然后成为材料科学。

[19:07] I starting to hire more people in material science.
  我开始在材料科学领域雇佣更多的人。

[19:11] So that is kind of innovation in our area.
  所以这是我们领域的一种创新。

[19:13] How can we do that?
  我们该怎么做？

[19:16] And I still remember 18 years ago and I, I still investing in semiconductor and actually most of the VC firm some of
  我仍然记得18年前，我仍然投资于半导体，实际上大多数风险投资公司中的一些

[19:24] them were very nice tier one venture firm a good friend of mine and initially the partners meeting the whole partners in the room then after I talking about semiconductor half make excuse to run out of the room then eventually the other half they said available do you have any software service.
  他们是非常好的顶级风险投资公司，我的一位好朋友，以及最初的合伙人会议，房间里的所有合伙人，然后在我谈论完半导体后，他们找借口离开了房间，然后最终另一半说，你们有任何软件服务吗？

[19:44] so then they left with only who sympathetically listen to me.
  所以他们只剩下同情地听我说话的人了。

[19:48] So it's kind of the history have changed and now semiconductor if you look at it Jensen is a 5.3 trillion market cap company and then Broadcom and TSMC is two trillion market cap company and Lisa my good friend at AMD is almost 800 billion and I'm close to 600 billion.
  所以历史似乎已经改变了，现在半导体，如果你看看英特尔，它是一家市值 5.3 万亿美元的公司，然后博通和台积电是市值 2 万亿美元的公司，我最好的朋友丽莎在 AMD 公司，市值接近 8000 亿美元，而我的公司接近 6000 亿美元。

[20:09] So in some way it's kind of semiconductor become hot again and it become essential because 15 years 20 years ago when I invest in semiconductor no VC want to join me except you know some of the big corporation like Samsung you know ARM and soft bank and others and investing.
  所以某种程度上，半导体又变得热门起来，而且变得至关重要，因为 15 年前，20 年前，当我投资半导体时，没有风险投资愿意加入我，除了你知道的一些大公司，比如三星，你知道的 ARM 和软银以及其他公司，并且在投资。

[20:26] with me and then now I starting to see a

[20:29] lot of VC like to come investing in semi

[20:31] so I'm very happy

[20:33] >> given the enormous interest in investing

[20:36] in this area that used to be considered

[20:38] too

[20:39] Right. Yes.

[20:39] >> Um what do you think I mean you've been

[20:41] a venture investor with Walden for a

[20:43] very long time as well as an operator.

[20:45] Uh you know the the general fears I'm

[20:49] just going to list a a bunch of them. Um

[20:51] the the general fears have been uh it's

[20:54] very capital intensive. Um and you

[20:56] should tell me what I'm missing. It's

[20:57] very unpredictable in terms of you know

[20:59] shipping a design that works missing

[21:01] tape out. Um and uh you need to

[21:05] understand the workload very well. I

[21:07] think there's a there's another which is

[21:09] just like it's it's very high risk for

[21:10] the customer. Yes. To switch, right? I

[21:13] think you know we've been involved in

[21:14] companies together where you know

[21:16] there's a design win and then there's

[21:17] still the question of like scaling order

[21:19] volume. Um uh and then there's a

[21:22] cyclicality. Yes.

[21:24] >> Right. Of you know you you build hard

[21:26] manufacturing capacity and demand may

[21:28] may change or not in any any given year.

[21:31] Um what is your view on how a bunch of

[21:35] you know what makes it hard as an

[21:38] industry uh and then the the secular

[21:41] demand growth from a bunch of different

[21:43] areas right so you have the um

[21:46] recognition of how important the a more

[21:48] diverse supply chain is and then you

[21:50] have this like explosive demand growth

[21:52] on the AI side how do you you're still

[21:54] an investor and then you're making the

[21:56] biggest bet ever like go be CEO how do

[21:58] you like think about these different

[22:00] risk and advise others about where to

[22:02] invest in this supply chain. I realize

[22:04] that's a very large question, but just

[22:05] given your your history with it, I I

[22:07] think there's a there's a lot of like

[22:09] yolo action of like there's a memory

[22:11] shortage by memory stocks um as well as

[22:14] you know just an unwillingness to take

[22:17] on things that have a 10-year timeline

[22:19] like material science.

[22:20] >> Good. You have quite a broad range of

[22:23] questions. Let me try to uh explain

[22:26] that. So first of all I think uh you

[22:29] know the venture capital startup is in

[22:31] my blood and I really enjoy it and uh so

[22:34] I think uh this is not tied to brag

[22:37] about it and so there's some good exit

[22:40] you know I still have 159 IPO

[22:44] 126 uh you know M&A and that's include

[22:47] semiconductor just break down the

[22:49] semiconductor I invest over the years uh

[22:52] 200 and 38% is in US so what are

[22:56] actually look at some mechanisms

[22:57] >> just to just to be clear that's

[22:59] incredible right

[23:00] >> thank you thank you it's just enjoy

[23:01] building it and uh but more important I

[23:04] look at is first of all on the

[23:06] investment side I always look at where

[23:08] is the bottleneck what are you trying to

[23:10] solve and for example I invest in

[23:12] company called uh cradle semiconductor

[23:15] australa lab is this interconnect become

[23:18] the bottleneck so I decide to back and

[23:20] also I back uh celestial AI you know

[23:24] optical sign and then because speed

[23:26] become more important in the

[23:27] interconnect in the cluster. So I think

[23:30] optical become very important. Look at

[23:32] Jensen he invests in almost every

[23:33] company is photonic related

[23:36] >> and then the other part I looking at is

[23:37] you know okay what are the uh solution

[23:40] that need like for example we talk about

[23:43] design and then the complexity and also

[23:46] the cost. Can you find some using AI

[23:49] machine learning to drive better design

[23:52] and better solution? So a couple of new

[23:55] startup actually go into the EDA related

[23:58] area to drive performance improvement. I

[24:00] think it's a goal mine to do that. And

[24:03] then the other part you look at the new

[24:04] material and we talk about you know this

[24:07] u you know Indian fastfi that's why I

[24:09] invest in infi and then well bought it

[24:12] and then uh then you invest into some of

[24:14] the new material that gallium nitrite

[24:17] and then silicon carbide and then some

[24:19] of the company starting to being

[24:20] acquired include one of them you know

[24:23] doing power management and ADI just

[24:25] bought empower and so again this IVR

[24:29] that's a very very good area in power

[24:31] management become bottleneck in term of

[24:33] converting from 40 volt down to one volt

[24:36] and then those in term of that

[24:37] conversion you lost a lot of power and

[24:39] how you do drive the power improvement

[24:41] so I think power thermal those become

[24:44] the bottleneck so I think I always look

[24:46] at from what is the problem we try to

[24:49] solve is it real is customer crying for

[24:52] it and then I starting to invest the

[24:54] next thing is look at it's very

[24:56] important from day one you'll have to

[24:58] target the first customer

[25:00] >> and usually I like the customer is hyper

[25:02] the skill. They have the skill. If they

[25:04] like what you have, they're willing to

[25:06] pay million of dollars next few years

[25:09] and even giving some warrant is worth it

[25:12] because you have a big one customer you

[25:15] can scale. So I always look at some of

[25:17] the formula how do you do that and then

[25:20] where do you get the talent and then

[25:22] yeah you know sometime it's very

[25:24] important to find the talent. That's why

[25:25] I'm very interested in US and then

[25:28] Silicon Valley and then some Austin and

[25:31] then the other part is Israel a lot of

[25:33] talent so I back quite a few quite a

[25:36] significant amount my investment in

[25:37] Israel and then because they have very

[25:40] disruptive innovative entrepreneur and

[25:42] they work really hard even in this

[25:45] wartime they still have conference call

[25:47] and sometime they say okay there's a

[25:49] there's a warning I have to go to

[25:52] underground and then the internet may

[25:54] not be good maybe we does use voice in

[25:56] some way it's kind of fun the kind of

[25:58] resilient entrepreneurship I really

[26:00] enjoy so I think all in all I felt that

[26:02] there's a lot of opportunity and

[26:04] especially in the AI and then right now

[26:06] beside the agentic AI now you're looking

[26:08] at physical AI next a mix big frontier

[26:11] >> and then you had to really look at the

[26:13] full stack that's why I'm still involved

[26:15] with a lot of this frontier model that

[26:17] we very familiar and some of the

[26:19] investment I back because I really like

[26:21] open-source frontier uh you know

[26:24] technology technology for physical AI. I

[26:26] think that's a goal, mate.

[26:27] >> You mentioned the opportunity to uh make

[26:31] certain parts of the design and test of

[26:34] um of chips uh faster, cheaper, more

[26:37] creative with AI um given your cadence

[26:40] experience like where do you what do you

[26:41] think is most fertile? Is there anything

[26:43] you think is already working? Yeah, I

[26:45] think you know uh for almost 15 years

[26:48] with Cadence and I'm so happy one of my

[26:50] highlight is able to find my successor

[26:53] on the road and I train him and he

[26:56] becomes super great CEO and then he

[26:59] really embracing the AI you know driving

[27:02] the agentic AI to drive more efficient

[27:04] uh but there good part I think synopsis

[27:07] sashin also tried to do that and they

[27:09] have investment from you know Nvidia 2

[27:12] billion I think helping him to do a lot

[27:14] and he acquire answers to move into the

[27:17] whole system uh design. So I think all

[27:19] in all they all do the best thing they

[27:21] can but also some opportunity for

[27:24] startup to do some of the more

[27:26] disruptive and then eventually they can

[27:28] I go public or being acquired by both of

[27:30] them or seaman to acquire them. So I

[27:32] think there's opportunity for all depend

[27:35] on what the entrepreneur vision and then

[27:38] as long as I always have philosophy if

[27:41] entrepreneur want to sell the company

[27:43] and this quicker way for exit you don't

[27:46] have a lock up you don't have to worry

[27:47] about quarterto quarter earning and then

[27:50] some entrepreneur they from day one they

[27:52] want to go IPO you know for being a VC I

[27:55] think three of you with three of us we

[27:56] all VC we support the entrepreneur their

[27:59] dream and then help them to fulfill

[28:01] their dream Yeah, if you look at the

[28:03] different areas that you mentioned in

[28:05] terms of future either product

[28:06] development or impact of AI on the

[28:09] semiconductor industry, there's

[28:10] companies like Periodic doing materials,

[28:12] there's to your point folks working on

[28:14] the EDA side and uh design and other

[28:17] aspects and sort of throughout the chain

[28:18] there's manufacturing. Um do you think

[28:21] that either Intel or future

[28:23] semiconductor company 10 years from now

[28:25] looks radically different from today

[28:26] given AI and if so how?

[28:28] >> Yeah, I think so. I think first of all

[28:30] uh back to Sara your question about

[28:33] capital intensive and a little bit

[28:35] unpredictable and cyclical. So you have

[28:38] to kind of put that into factor into

[28:40] your decision making investment you know

[28:42] I usually like to go in very early put a

[28:45] team together it's kind of fun to do

[28:46] that I think you you also do that and

[28:49] then secondly you try to find the right

[28:51] investor that can co-partner with you uh

[28:54] it's not just the whatever the brain

[28:57] firm I usually go for the individual and

[29:00] then whoever the individual that really

[29:02] knowledgeable in this space you can the

[29:04] most important to find a partner to

[29:07] difficult time and good time. A lot of

[29:09] the time people are very enjoyable

[29:11] working with you is a good time. When

[29:13] the company trouble they just walk away

[29:15] I like to have partner that really work

[29:17] through a lot of successful company they

[29:19] have multiple time almost bank club that

[29:22] eventually take off. So I think it's

[29:24] both important to find a partner willing

[29:26] to do that and then the other part is

[29:28] look at what are the strategic investor

[29:31] that can help you either in

[29:32] manufacturing or memory connectivity or

[29:35] various way to add value to the company

[29:38] and also have couple of friend they are

[29:40] in the growth stage and also in the

[29:42] hedge fund and I really enjoy them

[29:44] because they have a different

[29:45] perspective they know about the public

[29:47] market he can guide the company

[29:49] entrepreneur where not to go and so

[29:52] those can be very helpful. So I think

[29:54] all you know I think is just fun to do

[29:56] that and then just realize is a

[29:59] engineering for startup is like problem

[30:01] solving each step of the way you have to

[30:04] find people to help you to solve the

[30:06] problem and then if you trigger that

[30:07] then great next frontier to work on

[30:10] >> and then frankly speaking I look back

[30:13] nine of the 10 company I invest halfway

[30:16] they change their business plan

[30:18] >> because market have changed. So I like

[30:20] to have entrepreneur as team not just

[30:22] one person. Secondly open mind. Yeah.

[30:25] Willing to listen and listen you know

[30:28] getting coaching from us

[30:29] >> and then eventually they formulate their

[30:32] own plan. It's not just do what I want.

[30:34] It's more they figure out the best thing

[30:36] is you give them enough feedback they

[30:39] draw their own conclusion that you

[30:42] exactly what you like and or different

[30:44] that you can embrace is the right

[30:45] decision. That's kind of fun of doing

[30:48] startup. you know they can much faster.

[30:50] >> So back your question if you look at it

[30:52] >> 10 years from now what will be the

[30:55] winning company this is just my personal

[30:57] view

[30:58] >> the one that articulate and laser focus

[31:01] on one niche area

[31:03] >> and also find the right partner and also

[31:06] able to scale the company and so in some

[31:10] way and back to my point about full

[31:12] stack. So in the way you need to have a

[31:14] full stack solution

[31:16] >> and uh so it can be big company they re

[31:19] you know transform themsel to be looking

[31:22] at big platform like Jensen I admire him

[31:25] >> you know he focus on kuda he focus on

[31:27] elibu I want to be a platform company

[31:30] and he did it and so in some way you can

[31:33] do that or startup company like entropic

[31:36] openai they find a way to do it in a

[31:39] more elegant way they change the game

[31:41] and and then it start up move fast you

[31:44] know speed of light you can really

[31:45] become a dominant player and hopefully

[31:48] Intel can play the role because we have

[31:50] the XPU and we have the advanced

[31:53] packaging and we have foundry if you put

[31:56] that all together can build some of the

[31:58] purposebuilt silicon for different

[32:00] workload I think that's where I'm going

[32:03] >> yeah that makes a lot of sense and I

[32:04] guess part of the question I was I was

[32:05] wondering is where you're going and the

[32:07] other part is does it fundamentally

[32:08] change how you work because when I look

[32:10] in the software world I I think there's

[32:12] a very big shift happening right now in

[32:13] terms of who you hire, in terms of who

[32:15] you think you want on board, in terms of

[32:17] people managing multiple agents. And so,

[32:19] you know, many people now that I know

[32:22] are hiring people more in their 30s,

[32:23] 40s, 50s because they're used to

[32:24] managing teams. And I think that

[32:26] transfers directly over to managing

[32:27] agents in terms of understanding the

[32:29] complexity of what to set up and the QA

[32:30] and everything else. And I wonder in the

[32:32] context of the physical world or in the

[32:35] context of a fab how you think about

[32:37] shifts in terms of either team structure

[32:40] or capabilities or how AI layers on. And

[32:43] so I just wasn't sure if there's if it's

[32:45] a natural slow evolution or if there's

[32:48] areas where there's a radical shift

[32:49] where it's like oh from materials now we

[32:50] should just use these 3A models plus

[32:52] some chemistry or whatever it is. So

[32:54] that's why I was a little bit curious

[32:55] about

[32:56] >> how you think about the future world

[32:57] there.

[32:57] >> Good question. I think you know the as a

[33:00] back to that crawl walk and run. So I

[33:03] think crow you basically try to I

[33:05] recruit some of the best talent in the

[33:07] semiconductor industry and then now I

[33:10] starting to look at what are the

[33:12] software talent I need to bring on board

[33:14] and in order to build a full stack and

[33:16] now I starting to look at you know my

[33:18] average age of my team in the 40 late 40

[33:22] 50

[33:23] >> I need to bring in some new talent and

[33:25] then so they're understanding the

[33:27] workload understanding the frontier

[33:29] model open source uh that is important

[33:32] So I found out that my son become my

[33:34] teacher now.

[33:35] >> So every time he invite me to go to his

[33:37] house, we're playing to grandkids. I

[33:39] starting to tap on him on all the AI

[33:41] machine learning, he's more plug-in than

[33:43] me. So I learned a lot and then try to

[33:46] understand investing and then bring some

[33:49] of the talent to come in. So we are

[33:51] changing Intel used to be a very old

[33:53] legacy spreadsheet company. Now I'm

[33:57] transform it to become AI and not uh AI

[34:01] enable uh using some of our design and

[34:04] also across all the engine uh all the

[34:06] organization embracing AI and then so

[34:09] they become uh less less uh depend on

[34:12] the spreadsheet and label to do that and

[34:15] you're going to combine the two talent

[34:17] plus the best AI tool that I can use not

[34:21] only for my organization not only for my

[34:24] sales and then now exciting to look at

[34:26] not just marketing and now the design

[34:29] and then to embrace that.

[34:31] >> I I think a lot of investors um you know

[34:35] at at least for me the last few years

[34:36] since I started a firm it's been very

[34:38] educational thinking about the different

[34:40] capital sources for more capital

[34:42] intensive companies. Uh I did a lot of

[34:44] software before and um and so your need

[34:47] to have smart friends with a very

[34:49] different stance in balance sheet was

[34:51] less if you're like ah I need $150

[34:53] million before this thing gets to you

[34:55] know some critical mass um and and so

[34:58] you've lived that for a very long time

[35:01] and then you have the unique experience

[35:02] of um working with the government as a

[35:05] large stakeholder. How do you think uh

[35:09] this sort of industrial policy it's led

[35:11] to huge successes like TSMC right the

[35:15] most important companies in the world um

[35:17] it's also been a bit frowned upon in

[35:20] American business culture for a long

[35:21] time like how do you think that should

[35:23] change now or where is it relevant

[35:26] >> good question so I think you know

[35:28] clearly you know for capital intensive

[35:31] uh business and uh infrastructure play

[35:34] uh you need to access to the capital and

[35:37] then in some way I think for our early

[35:39] day venture capital investment you know

[35:42] now starting become very capital

[35:44] intensive

[35:45] >> yes

[35:45] >> and some of the venture firm willing to

[35:47] put 1 billion into some company

[35:50] >> is very unheard of in the VC business

[35:52] now it's happening

[35:53] >> and so in some way you just have to be

[35:56] you know I like this kind of bell curve

[35:58] either you go in very early and then

[36:00] because you're starting to do the series

[36:02] A is over one billion valuations

[36:04] >> and so you had to go and pre- money uh

[36:07] preede to go into that kind of a 20 30

[36:10] billion valuation is very rare right

[36:13] now. So you just have to do that pick

[36:15] the right one and then the other part is

[36:17] able to find capital to scale and that's

[36:21] why some of this mutual fund they also

[36:23] like to move into the pre-market uh

[36:25] early states to join join me to

[36:27] investing I delight them because they

[36:29] are very less sensitive of whether I had

[36:32] to own 20% of the company there's not

[36:34] too many 20% to give so you have to find

[36:37] the right investor to come in and then

[36:39] in term of the capital intensive like AI

[36:42] you know factory and also the foundry

[36:45] and then you really need to tap either

[36:47] government funding or some sovereign

[36:50] fund and also some very big capital uh

[36:53] you know there are some big fund they're

[36:54] doing that and they really the fund

[36:57] they've organized is basically support

[36:59] the infrastructure and we like to tap

[37:01] into some of them and then to make sure

[37:03] that they can scale our operation so I

[37:06] think in overall government sovereign

[37:08] fund become very important and also as a

[37:12] public company. I also purposfully want

[37:14] to focus on some of the investor

[37:16] >> that are more long-term growth oriented

[37:19] and so that they can help me to grow the

[37:21] business and then rather than short-term

[37:24] asking capital location you know where

[37:27] do you going to you know buy back your

[37:28] shares those are good question but

[37:30] meanwhile I also had to build the

[37:32] business

[37:32] >> and so I think it's kind of that balance

[37:34] is important

[37:35] >> do you think there is something that

[37:37] investors like most misunderstand about

[37:40] Intel at this moment

[37:42] quite a few thing. First of all, I think

[37:44] um you know as a back to this uh crawl,

[37:48] run and walk

[37:49] >> last four month I crawl and then but

[37:52] people starting to recognize that

[37:53] potential of it and so the other part is

[37:56] very important. We need to really get

[37:58] the best product out either PC client we

[38:01] still have a market share but we really

[38:04] need to really build more perform better

[38:07] performance. So that's why I'm quietly

[38:09] building up the CPU architect, GPU

[38:11] architect and the software architect so

[38:14] that we can leaprog just like I look at

[38:17] Intel I want to be a multiple of startup

[38:20] culture so that we move fast and we can

[38:22] leaprog using better technology and then

[38:25] the other part is beside the product

[38:27] there are some new energy coming in like

[38:30] aentic AI the physical AI that's a lot

[38:33] of area that we can invest market is

[38:36] huge that's on the product side and the

[38:38] foundry side we are very distant from

[38:41] TSMC and then in term of their

[38:43] performance so there uh we have to be

[38:45] humble looking at building the building

[38:48] block like I mentioned earlier the IP

[38:50] the yield the defect density and the

[38:53] cycle time to make it more efficient and

[38:56] more reliable is a trust business people

[38:59] want to trust you before they give you

[39:01] the wafer to count on you so those are

[39:04] the thing will take longer time but I

[39:06] think by 2030

[39:09] 2032 31 32 I think I was starting to

[39:13] surface up people may not understand how

[39:15] big potential I can be in term of

[39:18] product you know the PC client that's

[39:20] our bread and butter and we moved up to

[39:22] the edge and move into the physical AI

[39:25] and a gentic AI and because not right

[39:28] now in the past you basically provide

[39:31] the server provide the PC for human

[39:35] >> now you're starting to have Another

[39:36] different dimension is millions of agent

[39:40] they need to access to the compute they

[39:42] access into the the software stack. So I

[39:45] think that part I think we have a chance

[39:47] to really play the game is not over yet.

[39:49] We can play on the in the injected AI

[39:52] and also the physical AI. So that's kind

[39:54] of where I'm going and the AI is just a

[39:57] beginning. you know you have the

[39:59] training that Jensen own and the in the

[40:01] edge and also you know in term of

[40:03] agentic AI with agents and also physic

[40:06] AI I think is the jumbo everybody have a

[40:09] chance so I think that's part that I

[40:11] want to go for it and so I think

[40:12] hopefully uh the investor will know even

[40:15] though in 14 month you know we make six

[40:18] time return to the shareholder they just

[40:20] a beginning we still have a lot of room

[40:22] to go

[40:23] >> there's venture returns from here

[40:24] >> yeah so you know I always look for 10x

[40:27] you know being adventure at heart. You

[40:29] want to look for 10x and know at cadence

[40:32] when I step down as a CEO I think we

[40:35] make about close to 76 time you know

[40:39] starting from interim CEO $242

[40:43] and then when I retire executive

[40:45] chairman about 85 time return to the

[40:47] shareholder so it hard to do that the

[40:50] Intel because the base is bigger so I

[40:52] kind of say okay let's do it at 10x you

[40:55] know and then and five year 10 years if

[40:58] we can do 10x X I think is a good return

[41:00] uh being a venture capital at hug that's

[41:03] kind of my goal.

[41:04] >> So there's a Godspeed on this very very

[41:07] large mission from this um from this

[41:09] huge base already. Um there's an

[41:11] embedded belief in what you described

[41:13] about where the workload is right where

[41:15] I think some would say like we're just

[41:18] going to be build bigger and bigger data

[41:20] centers and a gigawatt is the beginning

[41:22] and then uh uh but the the

[41:25] centralization and the efficiency from

[41:27] running even the inference compute in a

[41:29] centralized way is the is the dominant

[41:32] way versus thinking about the edge

[41:35] thinking about the client. um do you

[41:37] think that there's like an equilibrium

[41:39] state that you believe in of where the

[41:41] compute is or or is it just we will find

[41:44] out from the workload? How do you think

[41:45] about that?

[41:46] >> Yeah, I think you know that's a very

[41:47] good question. You know the right now

[41:50] there's a massive buildup in term of the

[41:52] AI you know the I think it's the right

[41:55] thing to do. I don't see there anything

[41:57] to slow it down uh because the workload

[42:00] is increasing a lot and then I think the

[42:02] question mark is how

[42:03] >> we are supply constraint.

[42:04] >> We're supply constraint. So I think

[42:06] anything slow down is the supply

[42:08] constraint. But I think the other part

[42:10] is uh I always look at all this

[42:13] infrastructure build up at the end you

[42:15] have to look at what is the solution

[42:18] what is the application you want to

[42:19] drive and I'm more focused on

[42:21] application. So if you can identify the

[42:24] application that is humongous or add up

[42:27] a few application to become meaningful

[42:29] and you focus on that it's not everybody

[42:32] build going to be winning

[42:33] >> and so some going to be winning big time

[42:35] and some going to lose over time or go

[42:38] sideway so you know just like internet

[42:41] you can see some of them turn out to be

[42:43] very big like Amazon like the Netflix

[42:46] and then some of them is kind of go

[42:48] sideway and disappeared or being

[42:49] acquired and so I think to me it's the

[42:51] same approach

[42:52] Then they really focus on what

[42:54] application they try to serve and that

[42:57] application how big is that and whether

[42:59] it's sustainable or not or is very

[43:01] crowded. So if it's too crowded you know

[43:04] maybe one or two may survive the other

[43:06] maybe just consolidate. So I think this

[43:08] industry go through that big growth and

[43:11] then then starting to consolidate maybe

[43:13] eventually one or two become the real

[43:16] winner. So I think that's kind of a

[43:18] we've watched the movie before so it's

[43:21] not surprise to me but focus on

[43:23] application like Netflix is application

[43:26] >> you know Amazon is a real application

[43:29] that to me they're winning

[43:31] >> but you're assuming that some of these

[43:32] applications they will be better served

[43:34] by client or edge compute than the than

[43:37] than only by the data

[43:39] >> exactly

[43:39] >> okay

[43:40] >> exactly

[43:40] >> yeah I mean I I will say as a I'm an

[43:42] investor in a number of companies that

[43:44] uh you know they're they're doing

[43:46] robotics they're doing defense and so

[43:48] the compute on the device is a very

[43:50] important choice in terms of our and

[43:53] what we assume around it like let's say

[43:55] a robot in the home eventually like what

[43:57] you assume is in the home and in

[43:58] connectivity around it um determines

[44:00] what you're able to do right um and I I

[44:02] think that that's been kind of it was

[44:04] kind of forgotten for a little bit in

[44:05] the in the SAS era

[44:07] >> yes yes I think I more my investment

[44:10] thesis is find a problem that is really

[44:13] need to solve

[44:14] >> and secondly who will be the player that

[44:16] you can partner with. And then thirdly,

[44:19] look at the application. How big is that

[44:21] application? Is that sustainable? And if

[44:24] it's really big, you believe in it,

[44:26] double, triple down.

[44:28] >> But you're including betting on

[44:29] applications that have not yet been

[44:31] broadly deployed. Okay,

[44:33] >> it's amazing.

[44:34] >> Well, thank you so much for joining us

[44:35] today. It was a pleasure. Thank you so

[44:37] much.

[44:37] >> Thanks, Leu.

[44:38] >> Thank you.

[44:41] >> Find us on Twitter at no prior pod.

[44:44] Subscribe to our YouTube channel if you

[44:46] want to see our faces. Follow the show

[44:48] on Apple Podcasts, Spotify, or wherever

[44:50] you listen. That way, you get a new

[44:52] episode every week. And sign up for

[44:53] emails or find transcripts for every

[44:55] episode at no-briers.com.