As an investor, I always look for where the bottleneck is — for example, interconnect became the bottleneck so I backed Cradle Semiconductor and Celestial AI in the optical side, and I look for AI/ML solutions that can drive better design and reduce cost and complexity in EDA.

(00:00:00) Nine of the 10 companies I invest, halfway they change their business plan because markets have changed. (00:00:06) So I like to have entrepreneurs as a team, not just one person. (00:00:09) I always believed in when I was at Cadence and also at Intel, is first of all you crawl and then be humble, listen to customers. (00:00:18) And then first step for me is to strengthen my balance sheets, focus on the products, and I really simplify the product, listen to the customer, and then drive the next generation leadership products. (00:00:29) And then right now, the agentic AI and influence, CPU become highly in demand. (00:00:35) And so in some way, I'm happy. (00:00:36) Right now, the demand is very high for my CPU. (00:00:39) Secondly, very happy that Jensen Huang, my all-time friend, he also put $5 billion in investing and support me. (00:00:45) His $5 billion become $25 billion now. (00:00:47) If you look at it, (00:00:49) Ten years from now, what will be the winning company? (00:00:52) The one that... (00:01:01) Hi listeners, welcome back to No Priors. (00:01:03) Today, Elad and I are here with Lip Bhutan, the legendary investor from Walden, then CEO of Cadence, now CEO of Intel. (00:01:10) We talk about his plan to transform Intel. (00:01:13) having the US government as a major shareholder, how to be an amazing semiconductors investor, and whether or not we can make chips in the United States. (00:01:21) Welcome, Lipu. (00:01:23) Lipu, it's great to see you. (00:01:24) We'll start with the obvious question. (00:01:26) This is a really hard job to go be CEO of this incredibly important American semis company. (00:01:32) Why take the job at all? (00:01:33) It's a good question. (00:01:34) I'm 66, and they put that, well, you should retire rather than take on this hardest job in the industry. (00:01:42) And so a couple of reasons. (00:01:43) One is this is an iconic company and it's so important for the semiconductor ecosystem and also so important for United States. (00:01:52) And so I decided, you know, do one more after Cadence. (00:01:58) A lot has happened in this past year. (00:02:00) What has been the most surprising to you? (00:02:02) Well, the most surprising that I don't learn from my previous job or even training is one day, early morning, President Trump asking me to resign. (00:02:12) and conflict of interest, and there's no exceptions. (00:02:16) And so I had to convince myself, first of all, I don't need this job. (00:02:22) I do it purely to save Intel. (00:02:24) And so take that personal issue out of the way. (00:02:27) Then I figure out, what can I do to be helpful to Intel? (00:02:31) And so good news is I have a meeting, you know, Thursday morning, and then Monday I have the meeting. (00:02:38) And then he listened to me. (00:02:40) I have a chance to explain myself. (00:02:42) I was born in Malaysia, grown up in Singapore, went to MIT, and I lived in the US. (00:02:48) I never lived outside the country. (00:02:50) And so something that I shared, and then somehow he listened very well, and then he gave me the chance. (00:02:56) And so I'm delighted. (00:02:58) And now you have the chance to do the work. (00:03:00) When you said the job is to save Intel, it's a really important company. (00:03:04) What does that look like to you? (00:03:05) What does Intel winning or thriving look like? (00:03:08) Yeah, I just passed 14 months. (00:03:10) A lot of things happened in this 14 months. (00:03:12) So a couple of things. (00:03:13) One is to change the culture. (00:03:15) And then clearly want to drive more accountability. (00:03:20) And also in terms of decision making, had to be faster. (00:03:23) I'm so used to startup culture and you move fast in the speed of light and going to have that bureaucracy layer of lay of meeting. (00:03:33) And so something that I changed accountability, listen to the customer and the customer delighted, you know, someone like a little so humble, willing to listen and then address some of the problem that they face and then try to delight the customer. (00:03:47) And also the other part from day one, I decided all the engineering report to me. (00:03:51) I'm being an engineer by training. (00:03:53) I want to know what went wrong and what are the things that I need to correct, listen to the customer and delight the customer, and then make sure that we have the right product, simplify our product line, and really have the roadmap and the vision for the next 5, 10 years. (00:04:08) What is your vision of where Intel should be in 10 years? (00:04:12) Yeah, I think a couple of things. (00:04:13) One, I always believed in when I was at Cadence and also at Intel is first of all, you crawl. (00:04:20) and then be humble, listen to customer. (00:04:23) And then secondly, you're starting to walk. (00:04:25) And then finally, you're starting to run in spring. (00:04:27) So that's kind of my culture of step by step doing it. (00:04:31) And then first step for me is to strengthen my balance sheets. (00:04:34) And the balance sheet is really horrible in some way. (00:04:39) So I'm delighted, you know, US government become a big shareholder. (00:04:43) Just I explained to President Trump, (00:04:45) TSMC, when they started, they have the Taiwan government as a shareholder. (00:04:49) If you look at Japan, you look at Singapore, this is the infrastructure US government get to provide the support. (00:04:56) Secondly, very happy that Jensen Huang, my all-time friend, he also put 5 billion in investing and support me. (00:05:04) And I'm glad I at least do some good work. (00:05:06) His 5 billion become 25 billion now or more. (00:05:10) And then the other part is SoftBank Masa. (00:05:13) I used to be at SoftBank board, and then he lent a hand to help me. (00:05:17) So we strengthened the balance sheet and then focused on the products. (00:05:22) And I really simplified the product, listened to the customer, and then drive the next generation leadership products. (00:05:28) And then in some ways, very lucky. (00:05:31) Right now, the authentic AI and influence CPO become highly in demand. (00:05:38) And so versus 1 to 8 in the training CPU to GPU, now I can see 1 to 4, maybe 1 to 1, and I'm delighted CPU become important. (00:05:50) I talked to some of the AI model and developer, and they said, well, in term of reinforced learning, in term of the speed of orchestrating all the agents, and turn out the CPU is actually better. (00:06:04) And so in some way, I'm happy. (00:06:06) Right now, the demand is very high for my CPU. (00:06:08) So I think overall, build on the product, on the data center server side. (00:06:13) Then the other part is our foundry business. (00:06:16) And initially, this is a capital intensive business and it's not easy. (00:06:20) And you really need to have a couple of things. (00:06:22) You need to have all the right IP so that you can support the customer. (00:06:26) Like for example, if it is a mobile related, you've got to have low power. (00:06:31) IP set that you need to have. (00:06:33) Without that, you cannot serve them. (00:06:35) It's a service business. (00:06:36) It's a trust business. (00:06:38) If people want to give you enough orders to have wafer to come, if the yield not good, they will be toast in terms of revenue miss. (00:06:47) So with that, I think it's very important to really focus on the yield, the defect density, the cycle time, and make sure that you're really able to meet and serve the customer in high quality (00:07:00) and reliable. (00:07:01) And so those are the things that I really focus on it. (00:07:04) And eventually you have to really move into a full stack. (00:07:07) So not just a silicon, you need to have a software. (00:07:11) And some of the customer ask me, give me the whole rack. (00:07:14) So there's a system that you have to build. (00:07:16) And so I think those are the things that I quietly building step by step and recruit some of the best talent I can find. (00:07:23) By the way, all the recruitment, I do it myself, no search firm helping. (00:07:28) And so I think sometimes it's good to have a Rolodex that who to reach out to call for. (00:07:33) Yeah, I mean, you've been in the business for so long, and you've run Cadence for, I think, 12 years before this. (00:07:38) And so. (00:07:39) 13 years. (00:07:40) 13 years, sorry, yeah, And then two more years as executive chairman, so 15 years. (00:07:44) I signed up for three months. (00:07:46) Three months, so right now I'm being very careful. (00:07:49) The moment you said, I just do it for three months, it turned out to be 15 years. (00:07:53) Yeah, it seems like you have a lot of longevity ahead of you here as well. (00:07:56) And so the other big initiative that has been sort of talked about is TerraFab and working with Elon Musk on that. (00:08:02) Can you tell us a bit more about how that came together and your involvement and how you all are collaborating? (00:08:07) Yeah, good. (00:08:07) I mean, Elon Musk, I think we all agree, is one of the best, if not the best, entrepreneur in this century. (00:08:15) He and I, we share the same view. (00:08:18) that semiconductor infrastructure actually is not catch up with the AI growth. (00:08:24) And in terms of you need the capacity, you need to have the productivity, and you have the dry efficiency. (00:08:30) And so those are the things that he and I, we share that there's something missing. (00:08:34) And then secondly, he's just delighted to work with him. (00:08:38) And he's very (00:08:40) I call it unconventional. (00:08:42) And he basically questioned every step and why this traditional way of doing things. (00:08:48) And in some way, it's very refreshing. (00:08:50) And I like that. (00:08:51) I like people have different opinion and let's work together, find what is the best route. (00:08:56) And we both going to learn a lot together. (00:08:58) And then I think clearly he have a vision that (00:09:01) his robots and his car, he need a lot of silicon. (00:09:06) Yeah. (00:09:06) Could you actually explain what TerraFab is for people who aren't familiar with it? (00:09:09) Yeah, TerraFab, he decided he wanted to build his own fab. (00:09:11) 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 and then using some of our technology and some of our process. (00:09:24) And that's something that we both kind of collaborate together. (00:09:27) And he's a very good team that I work with weekly. (00:09:30) And it's just refreshing to work with him. (00:09:33) 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. (00:09:38) The burger. (00:09:38) Yeah. (00:09:38) I think I don't go that far. (00:09:41) Maybe some part of the clean room, you can do that. (00:09:44) But I think something that is open mind and then we also listen and see what we can do that. (00:09:50) Yeah, I mean, it's very exciting to see how you're morphing the business here in the US in terms of (00:09:54) incrementally building out the foundry business in terms of collaborating with things like TerraFab. (00:09:58) 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. (00:10:09) And if I look at certain countries, when I look at the layoffs that are claimed from AI, for example, (00:10:14) Most of them, I think, are overstated right now. (00:10:16) Most of the layoffs are actually just overhiring during 2020, COVID period. (00:10:20) COVID period. (00:10:21) But the first things I see actually being cut are outsourced firms where you'd rather cut external headcount versus internal. (00:10:27) So you're cutting external customer support. (00:10:30) You're cutting external IT. (00:10:31) And that has more of an impact, I think, for certain countries which have big BPOs, the Philippines, India, et cetera. (00:10:36) And so they may be impacted in the short run by AI. (00:10:38) And then if you ask, how do companies participate in the future in a positive way in AI, (00:10:43) You have to almost go country by country, right? (00:10:44) Places with cheap energy will do data centers. (00:10:47) Places with the ability to train models will train models, but it's probably only the US and one or two other places. (00:10:53) How do you think about the shift in global supply chain for the semiconductor industry? (00:10:57) Should certain countries invest more? (00:10:58) Like, should Israel be doing more given Melanix and Nvidia and Intel presence there? (00:11:03) And should they try to do more in semiconductors? (00:11:06) Should the Philippines move back to more of a manufacturing base? (00:11:08) Like, how do you think about that on a global basis? (00:11:10) Yeah, good question. (00:11:12) So I think clearly the AI is changing the whole landscape. (00:11:15) And I think the impact will be bigger than internet. (00:11:19) And it's more profound also. (00:11:21) So I think the AI initially is able to help you to do things more efficiently. (00:11:27) And then with a lot of agents helping you to do things that is now kind of mundane that you need to do, but now they can give it to you faster. (00:11:37) So in some way, I think it can drive a lot of efficiency. (00:11:41) even the semiconductor design, how much you can drive the efficiency in terms of timing, how quickly can you come out, and secondly, the cost. (00:11:50) And so I think those will be helping you to drive that. (00:11:53) And then I think a couple of bottlenecks for the AI demand and growth. (00:11:58) One is, of course, everybody knows power consumption. (00:12:01) Some countries, the power, they just don't have that. (00:12:04) It gets impacted. (00:12:04) And then secondly, a lot of people didn't realize the helium impact can be also quite significant for semiconductor. (00:12:13) And then thirdly, is everybody know right now memory is a bigger shortage. (00:12:17) And everybody try to scramble for memory. (00:12:20) And then even though you have to build a fab to capacity increase, it will take a couple of years to do that. (00:12:25) And same thing for CPU, GPU. (00:12:28) and all this will be highly demanded. (00:12:30) And I think also the pricing also go up because we have to pass the price, the cost to the customer. (00:12:37) So I think those will be the impact the industry growth. (00:12:41) And then I think overall I felt that the company (00:12:44) that most impacted is you're not embracing AI. (00:12:49) And because AI can help you to drive a lot of efficiency across all the different function of the enterprise, we should embrace and also find a way to better use the AI for your prediction, for your design, for your, you know, all the different part of the workload. (00:13:07) And I think that's tremendous. (00:13:08) A number of people would say the simplistic argument against (00:13:12) TeraFab against Intel Foundry being competitive is really a question of, there's all the factors in terms of the building, right? (00:13:21) You describe IP and velocity of just how you're doing business. (00:13:26) Then there's external factors and, you know, Elad's talking about a number of them, but one of them is the cost of labor and actually the manufacturing capacity. (00:13:36) You know, in investing in the Foundry business, you obviously believe there's a version where you can manufacture domestically and (00:13:42) does too. (00:13:43) Can you talk a little bit about that and how real that constraint is, the labor constraint? (00:13:48) Right. (00:13:48) So I think, when I decided whether I should double down on Foundry or should I get out of Foundry, and I... (00:13:56) And there's a lot of voices. (00:13:58) A lot of voices in the marketplace, as you can tell. (00:14:00) It's very expensive. (00:14:02) It's not going to work. (00:14:03) But I finally decided this is very important for United States. (00:14:07) and also very important for the industry. (00:14:10) And I'll give you the idea that we all live through these challenges of supply chain. (00:14:17) And it's very important for any of the big company in semiconductor and really have to think about the supply chains. (00:14:24) And you have to have a robust and resilient supply chain. (00:14:27) You cannot just depend on one or two players in different geographic growth. (00:14:32) And so I think (00:14:34) more and more people are going to realize making it in the United States is critical. (00:14:38) And then the most advanced process, like for example, we have the 14A is like 1.4 nanometer. (00:14:46) And we're already starting to plan for 1 nanometer and 0.7 nanometer. (00:14:51) It's getting smaller and smaller. (00:14:53) So in a way, it's much like our hair, so thin. (00:14:57) So it's a lot of complexity. (00:14:58) It's not that easy to do. (00:15:00) And every step, if you make a mistake that you just go down, go down the drain. (00:15:05) So in some way, you have to be really precise in that manufacturing. (00:15:10) So in some way, this has become more and more going to be the bottleneck. (00:15:14) So we felt that we have a lot of respect for TSMC. (00:15:18) We're a great partner. (00:15:19) And then the more important, we both need to have more capacity to serve the customer. (00:15:25) And so I think we decided by the bullet, longer term, (00:15:30) I think it's critical, and that's where I can create more value for the industry. (00:15:35) People have been talking for a long time about eventually hitting a point of resolution where you can't really miniaturize things further. (00:15:42) Like the line width just gets too small to be able to keep going. (00:15:47) When do you think we actually hit that limit? (00:15:49) Good question. (00:15:50) So I think I can see, you know, right now we have 18A, and then now we're going to production of 14A. (00:15:57) I can see 10 and seven. (00:15:59) And so I think that path, I think we can get there, but it couldn't be more and more expensive and more difficult to do. (00:16:06) And that's why we need partners. (00:16:07) We cannot just do it ourself alone, partner with a subscript vendor, partner with equipment vendors, so that make sure that we can really drive those yield and performance. (00:16:16) And then the other part also really become the bottleneck is packaging, the advanced packaging. (00:16:22) And so we all know about Corewort by TSMC, (00:16:26) Now we have a really good one called EMIP-T that is really next generation. (00:16:32) I had to make sure that you become able to do in the production yield that meet the customer requirement. (00:16:38) And now see more starting to run out of steam like you described. (00:16:42) So right now I also look at some new materials. (00:16:44) So become going back to the material size or the chemical table. (00:16:48) So gallium nitride, silicon carbide, and indium phosphide. (00:16:53) So I invest in all three. (00:16:55) And then looking at some of this new material, how can we really drive that? (00:16:59) And then in terms of packaging, I starting to invest into glass. (00:17:03) Glass is a very good heat insulator. (00:17:06) So I even invested a venture site called 3DGS. (00:17:09) Then I realized that Intel, we have like 1,000 patent on the module. (00:17:14) So how the subscript and the module put it together. (00:17:18) And we just announced a big program with Indian government to manufacturing in India plus in US and New Mexico. (00:17:25) So I think this advanced packaging very important. (00:17:28) I also starting to look at artificial diamond. (00:17:32) And that's another very good insulator. (00:17:35) So I also invest into, diamond foundry. (00:17:38) And that's something is a next generation to look at. (00:17:41) So new material, new subscript material, and new, design methodology to drive that. (00:17:49) So one thing good about being an engineer, you're always hitting the wall, then you find a way to either jump over the wall or you walk around the wall and then to get to the better result. (00:18:00) And that's what I'm being (00:18:02) have been long time as an investor and building semiconductor from the EDA tool to design to manufacturing. (00:18:10) It's kind of nice to have that experience. (00:18:13) Now I can help find a way to make a small contribution to the industry. (00:18:16) Yeah, and it's very exciting. (00:18:17) 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 (00:18:23) Once you hit 7 angstroms or whatever the limitation is, you start to run into it. (00:18:27) Yeah, you need to find new materials or find other workarounds. (00:18:30) And then the interesting question is, and we've been talking about this for a long time. (00:18:34) I remember 20 years ago, people were talking about how we'd eventually hit a point where we ran out of space on this, is do you run into some sort of asymptote that actually normalizes performance across different foundries or not? (00:18:45) Yeah, good question. (00:18:46) In terms of like Moore's law, it's a double and then the power and the cost. (00:18:52) And then you can double the performance, but you cannot double down on the cost and area. (00:19:00) So those are the things you have to give way unless you find some new way of material, new way of the material science. (00:19:07) And then become material science, I starting to hire more people in the material science. (00:19:12) So that is kind of innovation in our area. (00:19:15) How can we do that? (00:19:16) And I still remember 18 years ago, and I still investing in semiconductor. (00:19:22) And actually most of the VC firm, some of them are very nice tier one venture firm, a good friend of mine. (00:19:28) And I initially the partners meeting, the whole partners in the room. (00:19:32) Then after I'm talking about semiconductor, half make excuse to run out of the room. (00:19:38) Then eventually the other half, they said are valuable. (00:19:42) Do you have any software service? (00:19:44) So then they even left with only two sympathetically listen to me. (00:19:48) So it's kind of the history have changed. (00:19:51) And now semiconductor, if you look at it, Jensen is a 5.3 trillion market cap company. (00:19:57) And then Broadcom and TSMC is 2 trillion market cap company. (00:20:02) And Lisa Su, my good friend at AMD, is almost 800 billion and I'm close to 600 billion. (00:20:09) So in some way, it's kind of semiconductor become hot again. (00:20:13) and it become essential because 15 years, 20 years ago, when I invest in semiconductor, no VC want to join me except me now some of the big corporation like Samsung. (00:20:23) ARM and SoftBank and others and investing with me. (00:20:28) And now I'm starting to see a lot of VC like to come investing in Semi, so I'm very happy. (00:20:33) Given the enormous interest in investing in this area that used to be considered too hard, right? (00:20:39) Yes. (00:20:40) What do you think, I mean, you've been a venture investor with Walden for a very long time, as well as an operator. (00:20:47) the general fears, I'm just going to list a bunch of them. (00:20:51) The general fears have been, it's very capital intensive, and you should tell me what I'm missing. (00:20:57) It's very unpredictable in terms of, you know, shipping a design that works, missing tape out, and you need to understand the workload very well. (00:21:07) I think there's another, which is just like, it's very high risk for the customer to switch, right? (00:21:13) I think, we've been involved in companies together where, there's a design win, and then there's still the question of like scaling order volume. (00:21:20) And then there's a cyclicality, right? (00:21:24) Of, you build hard manufacturing capacity and demand may change or not in any given year. (00:21:33) What is your view on how a bunch of, what makes it hard as an industry and then the secular demand growth from a bunch of different areas, right? (00:21:44) So you have the recognition of how important a more diverse supply chain is, and then you have this like explosive demand growth on the AI side. (00:21:53) How do you (00:21:54) You're still an investor, and then you're making the biggest bet ever, like go be CEO. (00:21:58) How do you think about these different risks and advise others about where to invest in this supply chain? (00:22:04) I realize that's a very large question, but just given your history with it, I think there's a lot of like YOLO action of like there's a memory shortage by memory stocks, as well as just an unwillingness to take on things that have a 10-year timeline, like material science. (00:22:21) Good. (00:22:21) You have quite a broad range of questions. (00:22:24) Let me try to explain that. (00:22:27) So first of all, I think the venture capital startup is in my blood and I really enjoy it. (00:22:34) And so I think this is not tied to brag about it. (00:22:38) And so there's some good exit. (00:22:40) I still have 159 IPO, 126 M&A, and that's includes semiconductor. (00:22:49) Just break down to semiconductor, I invest over the years 200. (00:22:54) and 38% is in US. (00:22:56) So what I usually look at semiconductor. (00:22:58) Just to be clear, that's incredible. (00:23:00) Oh, thank you. (00:23:01) Thank you. (00:23:01) You just enjoy building it. (00:23:03) But more important, I look at this, first of all, on the investment side, I always look at where is the bottleneck. (00:23:10) What are you trying to solve? (00:23:11) For example, I invest in a company called Cradle Semiconductor, Australia Lab. (00:23:17) Is this interconnect become the bottleneck? (00:23:20) So I decided to back and also back Celestial AI in the optical side. (00:23:25) And then because speed become more important in the interconnect in the cluster. (00:23:29) So I think optical become very important. (00:23:32) Look at Jensen. (00:23:32) He invest in almost every company is photonic related. (00:23:36) And then the other part I'm looking at is, you know, okay, what are the solution that need? (00:23:42) Like, for example, we talk about design and then the complexity and also the cost. (00:23:48) Can you find some using AI machine learning to drive better design and better solution? (00:23:54) So a couple of new startups actually go into the EDA-related areas.