TSMC: Humanoid Robots Will Look After the Elderly

0:00This BBC podcast is supported by ads outside the UK.

0:29CFO. LinkedIn has a word for that. Bull spend. Now you can invest in what looks good to your CFO. LinkedIn ads generates the highest ROAS of all major ad networks. You'll reach the right buyers because you can target by company, industry, job title, and more. So cut the bull spend. Advertise on LinkedIn. The network that works for you. Spend$250 on your first campaign on LinkedIn ads and get a 250 credit for the next one. Just go to linkedin.com slash broadcast. That's linkedin.com slash broadcast. Terms and conditions apply.

1:29Wendell Huang:And as ever, we're asking you to decide whether they're good, bad or just another billionaire. That's good, bad billionaire from the BBC World Service. Listen now wherever you get your BBC podcasts.

1:57Hello and welcome to Big Boss Interview. I'm Sean Farrington and I've got BBC's Asia business correspondent Surinjana Tiwari with me because this episode is focusing on a company not everybody might have heard of, but they are at the forefront of technological developments in the world. Surinjana, TSMC, Taiwan Semiconductor Manufacturing Company. I mean, a bit of a giveaway what they do, but why are they so important? Yeah, they're really important to the devices that we use every day. Essentially, TSMC makes advanced computer chips for companies like Apple and NVIDIA. And so they pretty much power everything from your smartphones to your apps and AI systems like ChatGPT.

2:43And so the company is really sitting right at the center of the AI boom right now. But it's also a real sticking point in geopolitical tensions between the US and China, especially because it supplies so many chips to some of America's biggest tech companies. but it almost does almost all of its manufacturing in Taiwan, which is an island that China claims as its own. And I actually travelled to the company's headquarters in Taiwan and sat down with its chief financial officer, Wendell Huang. There was a lot to get into, the AI boom, whether it's all a bubble, the geopolitics of chips and what it means for the price of things that we buy.

3:22And so that's what the conversation was all about. But you got to have a bit of a look round as well, did you? We don't often get the chance to do that on Big Boss Interview. What did you see? That's right. And the moment you step off the plane in Taiwan, you really get a sense of how dominant chip making and the industry is there. And we travelled to the headquarters of TSMC, which is in a place called Sinshu. It's a city just outside the capital, Taipei. And there's this huge science park. And as far as the eye can see, all you can see is fabrication plants or fabs. That's where those chips are made.

4:00And if they're not fully operational ones, they're ones that are being built, which just gives you a sense of how much this industry is continuing to expand to keep up with the demand. And even when we did this interview, we actually had all of our phones and laptops taken off us. And that's because TSMC is so good at making chips that it just doesn't want any information about the company or its processes. to leak. And the other really interesting thing is that lots of companies in Taiwan are actually trying to get into the chip making business. They might be making something else like badminton rackets or sports equipment, but they all want a piece of the chip industry because it's so successful there.

4:43Well, I'm looking forward to hearing this. Here is Wendell Huang, Chief Financial Officer at the giant of a business, TSMC, speaking to Surinjana Tawar.

4:57Thank you so much for talking to us. First, very simply, given the spend that your customers are making in AI infrastructure, can TSMC meet the demand? We're doing everything we can, wherever we can and however we can. We're expanding not just in Taiwan, but also in the US, in Japan and in Germany, trying to expand the capacity and meet the customer's demand. We're trying to catch up with the demand. So far, still trying. Do you have any sort of idea of how much you might fall short in the coming year or so? Well, the customers ask us to breathe so much, but all we can do is try to grow as fast as possible, Not just with the new fab, we can also increase our productivity so that we can generate more output from the existing fab and try to get closer to the demand.

6:02How do you increase your productivity? Is it capital spend or is it the use of AI in your processes? That's part of it. The other thing is our operations people are very good at increasing the productivities, generally more often using the same tools so that we are able to provide more wafers to the customers. And then we also are trying to build new fabs. But building new fabs will take two to three years to build the fab and take a year or two to ramp it up. So it takes some time. Obviously, you're getting orders, But is there any chance that this whole AI system, this move to the technology, is a bubble?

6:53Our conviction in this AI megatrend is very strong. We talk to the customers and also the customers' customers to understand their need and their economics. We talk to the customers' customers who are mainly the hyperscalers. And these companies are financially very strong with a lot of financial resources. So we believe that they are able to continue to invest, which they also announced this year's CAPEX. It's much higher than last year's. So we think that megatrend is there. Now, semiconductor is a cyclical business. So at some point of time, you may see a drawback. But the whole thing is the megatrend.

7:42will continue. Will the drawback ever constitute a bubble? It depends on what you say the bubble is. But when you draw back, you continue to grow the following years. That's why we are saying this, this is the multi-year structural demand, a mega trend. So given that and given your response to what you think the trends are likely to be, is there a risk of overcapacity and producing too much product? For us, we talk to the customers, the customers' customers. We also make our own judgment. And then we build our capacity in the disciplined approach. That's how we do this. Is there anything that the world is getting wrong about the AI infrastructure build?

8:34I think the demand is there. The structural megatrend is there. There may be some time that there will be a drawback, but our conviction is strong. Where do you see the most demand coming from within AI in the future? Going forward, humanoid AI serving older people. Humanoid robots, right? Yes. It will be very, very important because a lot of people are getting aged and you need help. And this is a very good application. We do see that there will be a lot of need, demand for this kind of services. And any other platforms or devices that you see demand picking up? Yeah, auto, for example. This is another area.

9:32Autonomous vehicles, is it? Yes. You're providing the building blocks of artificial intelligence. Do you think that the technology will replace jobs in the future? We think the technology will help us, help people to free up time to do more value-added works. Do you think it would increase productivity and TSMC? Yes. How is TSMC using AI in its production processes? We are using it both in fab operations and also in R &Ds. In fab operations, you can imagine, I cannot go into details, but we have the most data among all the semiconductor manufacturers in the world. So using AI help us utilize all this data, improve our efficiency.

10:24And do you see it taking jobs? I don't think so. It will actually relieve people and have more time in doing even more meaningful jobs. The U.S. has pushed TSMC to invest$100 billion in American fabs, and the U.S. Commerce Secretary wants 40 % of Taiwan's chip production moved to America. Is that a realistic ask? We go out of Taiwan to build capacity based on customers' demand. The customers want us to go there. It's not the request of government. For example, we go to the U.S. to build 5 nanometers and more advanced technologies because most of the customers of advanced technologies are in the U.S.

11:15We go to Japan. The phase one in Japan's JASM is to build for a CMOS image sensor. We have big customers over there. We go to Germany to build a specialty technology fab because our customers in auto industry and industrial customers are there. So we go to overseas to build the capacity based on customers' demand. So those are not your most advanced semiconductors. So in the future, are the most advanced chips going to continue to be made in Taiwan or Arizona or both? The most leading edge will continue to be RAMP in Taiwan for practical reason. When the new technology comes out of R &D, you need R &D people and operations people to work really closely together.

12:14In Taiwan, we're able to move hundreds of engineers from, say, Taichung site to Tainan site and or back. So it's for practical reasons that the ramp of the most leading edge technologies will stay in Taiwan. After that, when that's stabilized, we can move that to the US, for example, and ramp it over there. How long do you think it will take to build that ecosystem? I mean, can the Taiwan ecosystem be replicated elsewhere? It will take time. Think about this. The Taiwan ecosystem, it takes Taiwan a few decades to build to today's status. In foreign countries in the US, we think it will at least take five or 10 years or even longer.

13:04You mentioned your operations in places like Europe. Companies are looking for more regional production, closer to the product and the manufacturing. You mentioned Japan, the US and Europe, but everyone has said that it is cheaper to manufacture here in Taiwan. Are you concerned that customers will choose companies that are able to offer more competitive prices because they continue to be based here? Our customers always know that our value to them is probably the highest. And we want to reflect the value as well. So price is not what we use to compete. We compete on the technology, the manufacturing, and the trust that the customers gave us.

13:58But does moving manufacturing to other locations weaken Taiwan's semiconductor advantage over time? As long as the customers want us there, we will go there. And we will be the most efficient manufacturer in the region that we operate. That we are very confident in. What about Taiwan's geopolitical advantage if manufacturing is moving out? For TSMC, as we said, we also expand in Taiwan. It's definitely bigger than we expand in overseas. And the most leading edge technology will be ramp in Taiwan for practical reasons. Huawei unveiled a new technology that it says will produce the most advanced semiconductors within sort of five years.

14:55How close do you think Huawei is to making the kinds of chips that TSMC makes? Well, we don't really know at this moment, but that principle, that is not a new one. Other semiconductor companies did that before, and they obviously find a good way of using it because of no access to the most advanced tools. I mean, there has been some criticism that U.S. export controls have actually helped the Chinese semiconductor industry to progress faster than it ordinarily would. I mean, is it? And are you concerned? I don't think we are in a position to comment on that. We just comply to the rules and regulations.

15:47There are reports that Chinese companies are buying chips through third parties or third countries. I mean, does a company like TSMC have a responsibility for where the chips go after they leave the FAP? To comply with the export control, we do have a very robust system within the company to check. However, for us, there is so much we can do. If it goes out of our door to the customers, the customers, the customers, then we really are not able to identify the end result. But we do have an internal robust system to make sure when we ship out, those are the complying with the regulations. Is it challenging for a company of your scale to do that?

16:37A very robust system. Yeah. Has China's dominance in rare earths affected TSMC? No, not at this moment. Not that we are aware of. What about when the response to the tariffs came in and they restricted the export of rare earths? For our supplies, we maintain multiple supply sources, different suppliers from different regions, and we have our safety stocks so that we have the resilience in our supply chain. Xi Jinping did warn at the Beijing summit he had with President Trump recently that Taiwan could put the U.S.-China relationship in an extremely dangerous situation. How much of a pressure point is TSMC in that relationship?

17:33We're not politicians, so I don't think we are able to answer that. As I said, we expand based on customers' demand. Beijing has previously raised concerns about the US government's relationship with TSMC and other Taiwanese companies. I mean, how do you respond to that? We just said we, based on customers' demand, we invest to meet their demand. And given that both Beijing and Washington need TSMC to continue functioning, can you give us any insight on how easy or difficult it is to manage both those relationships and which one might be slightly easier than the other? We focus on the fundamentals of our companies, which is the trinity of strength that I just mentioned.

18:22we also maintain healthy dialogue with the government in the region that we operate in. Apple, NVIDIA, AMD, Broadcom, these make up a majority of your advanced chip customers. That's a lot of reliance on a handful of American tech companies. What happens if in the future the U.S. government decides to become more protectionist and maybe protect its own companies and manufacturing? The customers, they have to have very successful suppliers so that they can be successful in their end markets. You have to have advanced technology, the most leading-edge technologies. You have to be able to manufacture the chips, the wafers.

19:12You need to be able to ramp to a large volume in a very short period of time with very good yield. Not many companies in the world can do that. So if you're the customers, at the end of the day, you still need to find such a partner to work with. Companies like Samsung, Intel and SMIC, your competitors, are receiving a lot of government subsidies and also buying very expensive equipment. I mean, how does TSMC retain its edge in such an increasingly competitive market? Number one, this is our 39th year. We never lack of competitors. We know how to compete. Number two, the foundry industry, the semiconductor foundry industry, money alone is not enough.

20:06You need to have this technology, the manufacturing capability, and then you can have your customers' trust. So you put those together. you see examples in the past that government heavily subsidized companies may not be as successful as they thought. But you're receiving subsidies from the American government for your Arizona operations, right? That's right. That's right. But not U.S. and other regions as well. In the sense? We don't need the subsidies as long as it's equal footing. What do you make of the fact that Elon Musk says that he wants to start making chips? Maybe there's an innovation way we'd like to hear, but there's no shortcut in semiconductor manufacturing.

21:03Are you concerned that his ambitions will catch up with TSMC anytime soon? We monitor everything. But as I said, there's no shortcut in this industry. Are you likely to increase your prices for customers, as other competitors have done? CCWays also said, we treat our customers as partners. We don't increase the price suddenly, four-folds, five-folds. We reflect our value. It's a long-term strategic relationship with our customers. But you think it will be necessary? To reflect our value. Geographic diversification, for example, is a value. Technology leadership is a value. Manufacturing excellence is a value, and we want to reflect those values.

21:56So increased costs for your chips are on the horizon? Inflation, yes, did cause the cost to increase. I want to ask you about what's going on in the Middle East now. A lot of supply chains, a lot of companies' supply chains have been affected by the war there and the blocking of the Strait of Hormuz. What's TSMC's position? We diversify our sourcing, multiple sources and in multiple regions. And we keep our safety stocks. Therefore, we don't see any short-term impact to our operations. What about the use of helium and hydrogen, which of course is essential to chip making? Including in what I just said.

22:43Are you stockpiling anything? We have safety stock. We keep the safety stock. How long do you think your stockpiles will last? Sufficient for us at this moment. And can you just speak to the challenges of acquiring talent in your overseas locations? For example, I believe in the U.S. it has been difficult to get the right talent into your fabs that are being built there. First of all, yes, there was a learning curve. But I believe we pretty much overcome that. If you look at the achievement that our Arizona FAB has had, the yield in the Arizona FAB is as good as the mother FAB in Taiwan. The culture issue is both ways.

23:39We also need to, well, first of all, we need to recruit the people who agree to our value. And we also need to adjust. Our managers also need to adjust to a foreign environment, a foreign culture, so that we can work together and make the FAB very successful. And I think we have already achieved that in Arizona. So would you say you've overcome the biggest challenges of operating in the U.S.? Yes. I believe we've overcome the biggest challenges, the people in Arizona.

24:47Transcription by CastingWords Well, that's a not-so-great conversation with the CFO. LinkedIn has a word for that. Bull spend. Now you can invest in what looks good to your CFO. LinkedIn ads generates the highest ROAS of all major ad networks. You'll reach the right buyers because you can target by company, industry, job title, and more. So cut the bull spend. Advertise on LinkedIn. The network that works for you. Spend$250 on your first campaign on LinkedIn ads and get a 250 credit for the next one. Just go to linkedin.com slash broadcast. That's linkedin.com slash broadcast. Terms and conditions apply.

25:47Wendell Huang:We would movie director Steven Spielberg. Football superstar Cristiano Ronaldo to name just a few. And as ever, we're asking you to decide whether they're good, bad or just another billionaire. That's good, bad, billionaire from the BBC World Service. Listen now wherever you get your BBC podcasts.