The AI Supercycle, Asia, and Why This Time is Different

Anu Rajakumar: For much of the past decade, the technology sector has been the dominant engine of equity market returns. This year it ran hot, and then over the summer the sector sold off sharply. Markets have since recovered, but that drawdown is still on investors’ minds. Was it a warning about AI valuations, or was it just the volatility you’d expect from one of the biggest capital cycles in history?

My name is Anu Rajakumar, and today I’m joined by YT Boon, a Portfolio Manager and Head of Thematic for Asia at Neuberger. We’ll talk about the summer sell-off, where he sees opportunity across semiconductors and agentic AI, and what he thinks investors misunderstand about the China-US tech rivalry, and why he’s pushing back on the hype globally. YT, welcome back to the show.

YT Boon: Hi, Anu. Thanks for having me today.

Anu Rajakumar: YT, let’s start with this summer’s tech drawdown. Some folks are still rattled by it, but I gather that you are not one of them. Talk us through your perspectives.

YT Boon: The drawdown that we’ve seen during summer has a lot to do with the outperformance that we’ve seen throughout the year, and in fact, over the last two and a half, almost three years. Generative AI, the chatbots that we know of, are advancing a lot faster than you and I are able to catch up. That drove a massive supercycle, whether it’s companies’ earnings fundamentals or the market itself.

If you think about global macro today, almost every single aspect of the global economy is driven by AI, whether it’s infrastructure rollout, whether it’s the picks and shovels, such as semiconductors, hardware, servers, even the very boring things, components, such as copper, is now becoming very exciting because of the supercycle that we’re seeing driven by AI data center rollout. It’s created a lot of massive tech companies, top 10 of a lot of indexes. It’s dominated by all these AI or tech giants, rightfully so because they are driving a lot of efficiency.

I used to be a microprocessor chip designer in the semiconductor industry for about 9, 10 years. Throughout that career, I’ve seen the ups and downs of the industry. Now that I’ve been in the financial industry for the last almost 15 years now, this is the biggest I have seen. If you really talk to a lot of the technology companies, the CEOs, the CFOs, and so on, they’ll tell you the same. Even the very boring personal computer hardware makers, they’re telling you that this cycle is very different.

Now, with all these in the backdrop, market volatility is definitely going to be elevated, especially when we’ve seen such two- to three-year rally that’s pushed up a lot of valuations for hardware, semiconductor stocks. At the same time, we’ve seen a derating in software sector. A lot of software companies were seeing a derating throughout late last year and earlier this year, but market volatility continues. There’s been a lot of disruption in the market, but with a lot of noises that’s coming in, volatility, it’s a given.

Anu Rajakumar: Absolutely. I want to pick up on something that you just said. You said, and I quote, “This cycle is very different.” What I’ve noticed is that every time that technology rallies, lots of people wonder if this is the next dot-com bubble. Having covered the sector across multiple cycles, what is different today versus 1999?

YT Boon: Throughout I think the last 25, 30 years, we’ve seen many cycles, but this cycle is very different in a sense that, number one, the technology supply chain is far from being able to catch up. This is probably the first time we’ve seen so many bottlenecks in the technology value chain, whether it’s from semiconductors all the way to even materials. As I mentioned about copper, I should also highlight, for example, optical cables. We’re seeing massive bottlenecks, massive shortages in the value chain, which is driving a lot of large tech companies signing long-term deals with their suppliers. It’s not just a supply constraint that defines this very different cycle. It’s also the demand. This is real demand, a multi-year demand.

I think this really defines something that we’ve not seen before. Even, as I mentioned, the most boring companies are seeing strong earnings growth, not just valuation. In fact, if we really look into the fundamentals versus the valuations, we’re probably at a lower end of that valuation range right now, given the strong earnings visibility for the next couple of years.

Anu Rajakumar: Absolutely. If it’s not a bubble, where do you actually want to be putting money to work across, again, semiconductors, agentic AI, cybersecurity, et cetera? What are the key themes that you think are most exciting in this industry?

YT Boon: Anu, you just mentioned some specific key areas. We’re actually quite excited about, for example, optical networking because if we look at AI compute demand today, it’s just accelerating. Now, every single GPU has only certain amount of computing power, but if you can combine a cluster of these GPUs together, you get a much stronger, much bigger computing power. To do that, you need higher-speed networking.

If you think about the large companies like NVIDIA’s, the Google’s, the Amazon’s of the world, they’re all working with their suppliers to figure out what’s the best technology to remove this bottleneck from networking. Moving data from one GPU to another, from one data center to another, that’s now the biggest driver for the industry because it will help resolve a lot of bottlenecks.

I think what’s also interesting is that we’re seeing also broadening out to even applications layer. Cybersecurity is another area where we’re seeing companies getting a lot of inquiries. No CTOs or company executives will want to have their companies end up to be on the headlines having a cybersecurity breach.

Today, as you know, with the advancement of AI, especially AI agents, we’ve already seen some really scary episodes of AI agents being able to hack more effectively than a human being. In fact, we have limited human beings who are really good hackers, but if you can program your AI agents to be as efficient as human beings and you can launch hundreds of thousands or millions of these, this creates a problem. This is why potentially we’re seeing a lot more vulnerability in our digital infrastructure today. This is an area where spendings to protect for cybersecurity has to increase.

Another area that I also like to highlight is that the broadening out is now just at the very beginning. More and more companies are looking to join the bandwagon because this is really one of the largest industrialization that we’re seeing. I think specifically in countries like China and the US, this is seen as a very strategic industry that one has to win because whoever gets AI implemented in the most efficient way will drive a lot of growth, a lot of economic returns, and more importantly, stay ahead.

Anu Rajakumar: Actually, on that point, that was something I wanted to ask you about because I’m based in the US and there’s certainly the perception, I think, in America that the US is leading the AI race to some degree. There’s certainly communication about how critical it is that the US does indeed win the AI race in that this is a zero-sum game. You’re based in Asia and you feel that China’s AI innovations are somewhat misunderstood. What I’d really like to understand is what do you think the Americans are misunderstanding about where China is in AI? Maybe more importantly, what can the US learn from where China is and what can the Chinese learn from where the US is when it comes to the AI race?

YT Boon: Just to kick off this discussion with my background because I used to be a microprocessor chip designer for British, US companies like ARM, Broadcom, but I also had a couple of years working with Chinese companies like Huawei. This gives me a good perspective on how the technology and AI developments are in both nations.

Now, in the current generative AI age, the US is leading because the US have semiconductor companies, compute resources. They’re much ahead of China.

Now, China has fallen behind. Now it doesn’t mean that China will be behind forever. They are developing their homegrown semiconductor industry. Their hands are tied. They have no access to UV machines. They have no access to the best technologies, but the Chinese are really smart. They’re looking for ways to outrun. The Chinese are not looking for the best technologies. They’re looking at mass market technologies. The Chinese typically have the strategy call. If you can capture the market through mass market, that’s how you’re going to be able to continue to dominate.

The US and China are taking a very different approach. Obviously, we’ve seen US tech companies pushing the envelope, developing the best frontier AI models. The Chinese are fast copiers. They can really copy the best, but they can also out-innovate. In the US, I think that the best analogy is that most companies are trying to be Apple, the best pricing at the higher end. China is taking a very different approach. Just like their electric vehicles industry, they’re pricing down, cheaper solution, but for the mass market. There’s a mass market that would go for China.

In a race where nobody slows down, this creates a very dynamic environment. I think this is very good for active investors to find opportunities. I would say the opportunities does not just lie in the US. In Asia, the Golden Triangle of Asia, Taiwan, Japan, and South Korea, and also parts of China, a lot of Chinese companies are exporting to the US. Asia is now becoming a very important part of the AI supply chain.

Anu Rajakumar: Maybe let’s dig into that a little bit, because there’s a lot of great context on the US versus China, and where each of them stand, and some of the pros and cons of their approaches. As you mentioned, that Golden Triangle: Taiwan, South Korea, Japan, why does that regional cluster matter so much in this race?

YT Boon: It’s a very unique geography. If you look at how the semiconductor industry has developed over the last 50 to 60 years, originally, it was very dominated by the US. You have big companies like Intel, Texas Instruments, and so on. In the ’90s, things were changing. Outsourcing became a very big thing. Then Taiwan, especially Japan and South Korea, started taking a lot more of these outsourced demand, specifically on Taiwan, where we’re seeing the establishment of TSMC, who today has become the single largest semiconductor foundry for the world.

Japan, obviously, in the late ’80s, early ’90s, were also dominating in the semiconductor industry, but because of lack of innovation, Taiwan has taken over. If you look at US companies, semiconductor companies like NVIDIA, AMD, Broadcom, they are pretty much reliant on just one Taiwanese company to manufacture their chips, and that is TSMC. Within Taiwan, it’s not just TSMC. There is a vast supply chain of technology companies, suppliers to TSMC, suppliers to NVIDIA’s of the world. We’re looking at hundreds of companies.

Now, Taiwan is a stronghold in terms of the foundry and also the electronics manufacturing, the server manufacturing. South Korea plays a very different role. South Korea has a very dominant memory chip industry. It dominates over 70% of the global markets. It has two big players like SK Hynix, Samsung, and the supply chain on the ground in South Korea supporting that.

Now, even though Japan has faded since the 1990s, what’s recently changed is that Japan is a powerhouse in terms of providing semiconductor equipment to Japan and South Korea, and the US, and also materials. There are a lot of materials companies in Japan that rooted from the technologies they had since the 1980s. They still are very strong. This makes up the Golden Triangle of Asia.

Not to forget China. China today dominates in a few areas such as optical networking and printed circuit boards. They’re a powerhouse. Most of us know that China still is the factory of the world, not necessarily in semiconductors, but in areas like optical networking modules. They’re one of the biggest assemblers. In printed circuit boards, China still dominates with more than 70% market share. They’re really the mass market player here, which you can’t ignore. This pretty much makes up 80% of the manufacturing you can think of for the entire electronics and semiconductor world. That’s why Asia has become more and more important, especially into this AI supercycle.

Anu Rajakumar: You’ve made a bull case quite thoroughly so far. You spoke about the supply and the demand side both being so strong. Now, let’s look at the flip side. You did mention earlier on vulnerability, particularly when it comes to cybersecurity, but tell me more about what you’re cautious about, where you’re saying no, where you’re pushing back on the hype around AI, whether that’s at the scale of capital being committed, bad actors, or the longer-term human impact.

YT Boon: When it comes to the social human impact on AI, I think we’re still learning how artificial intelligence, especially this current version of generative AI, can impact humanity, which is why I think it’s interesting that we’re seeing more frontier AI companies now saying, “Let’s pace. Let’s figure out whether we’re doing this responsibly.”

Anu Rajakumar: Which is hard to do when you’re in a race. (laughs)

YT Boon: Exactly. I think what’s more interesting to see is that AI developments have gone to every aspect of the economy. It’s no longer a matter of whether you’re adopting artificial intelligence. It’s already part of our lives. If you don’t already notice, any social media apps or any advertising apps, they’re all powered by very, very sophisticated AI models. In some ways, we’re already living in a world where AI is manipulating a lot of the society behaviors.

Like it or not, I think based on what’s been happening over the last five years, you’ve seen how society, how AI has influenced even political outcomes. That’s something which we have to live with. Now, developing artificial intelligence, the next-generation models as we achieve AGI, hopefully over the next few years or next decade, we’re doing it more responsibly.

Anu Rajakumar: As we start to wrap up here, I’d like to hear your thoughts about where we go from here. You wrote recently about AI moving off the cloud and onto our devices, literally into our pockets and onto our faces. For example, smart glasses, which sitting across the table from you, I see you’re wearing the Oakley Meta glasses. We’ve seen robots coming up around the world. What do you think is actually going to work and what do you think is still science fiction?

YT Boon: I think none of what you talk about are science fiction because they’re already here. It’s just that how quickly we can find a commercial aspect to it. I’ve been wearing these smart glasses for more than a year. In fact, I own four pairs of smart glasses. I use them on a daily basis. It’s cool to have a smart glass with you, where you can snap pictures a lot quicker than anyone else. It’s cool that you can ask questions to AI that can answer your questions very quickly, answer the phone, listen to the music, do everything hands-free.

Robots. I think AI robots are going to be a few years away. Their commercial aspect of it, we’re still not able to find a very good use case. I think initially, physical AI or robots will find its way into smart manufacturing factories. That’s where it will excel. Then hopefully, later, find itself into our homes instead of having washing machines or refrigerators. You should have your AI do most of that with the robot. Obviously, there will be some anxiety that, oh, having a robot at home, that’s going to be very interesting. I think a lot of these are just at the beginning. We’re looking for more use cases.

Today, agentic AI is really what’s dominating. If you think about generative AI today, it’s able to generate texts and videos. That’s human beings’ outputs. Most employees are doing that on a daily basis. Today, those who are able to utilize agentic AI to produce work, to become more efficient, that’s going to be very important; that’s going to matter the most.

Companies or individuals who are able to leverage the power of artificial intelligence in a smart way, they’re going to be the winners. Companies who position themselves in the right place in agentic AI, they’re going to be winners in the space as well. I think we’re still at the pathfinding of how different types of AI applications go into a more commercial environment, but we’re seeing very innovative companies now coming up with a lot of new solutions. I’m very excited about these.

Anu Rajakumar: Absolutely. You were saying, before we started recording, how you were in Mexico City recently and the smart glasses really helped with some of your communication abilities.

YT Boon: Yes. Now, a lot of Mexican friends thought that I could speak Spanish. I love watching La Casa de Papel, which is a very famous Spanish drama. My Spanish is still far from being able to converse, but with my Meta glasses, I can do live translation very quickly and understand, at least get the gist of things very quickly.

I was walking through Polanco in Mexico City and looking at certain places where it looked interesting. I just asked Meta, “Where am I? What’s the story? What’s the history about this place?” I can get by very easily with artificial intelligence, but there’s a lot more to this. I think what’s interesting is the internet or tech companies that are launching these smart glasses, they’re not just thinking about the smart glasses. They are thinking about how to monetize based on these traffic.

Anu Rajakumar: YT, I can’t let you go without a quick bonus question. You may be the first guest that we’ve had on the Neuberger podcast to record an episode wearing a leather jacket. I understand there’s a story behind you and leather jackets. I’d ask if you don’t mind please sharing that with our listeners.

YT Boon: It all started with a silly question I asked my AI agent late last year. Jensen’s leather jacket looks cool. How much do they cost?

Anu Rajakumar: We’re talking about Jensen Huang of NVIDIA.

YT Boon: Jensen Huang of NVIDIA. This got me down a path where my AI agent helped me to source a leather jacket without me having trying it because it knows my weight, my height, my dimensions well. With the AI agent knowing that it’d be great to actually get a signature from NVIDIA CEO Jensen Huang on the leather jacket. Throughout this journey, my AI leather jacket agent even helped me to find the right pen for the signature. When I was at a conference where I was able to meet with Jensen Huang, my AI agent even asked me, “Jump on stage before the bodyguard moves.” I did that.

Obviously, to get Jensen Huang’s attention, it’s a bit of a different thing. My AI told me that either you capture his attention with food, obviously, I can’t bring food to the conference, or I speak a local Taiwanese dialect. Now, I don’t speak a local Taiwanese dialect, but AI has prepared me for that. I asked my friend to even teach me the right pronunciation. I shouted to Jensen in the Taiwanese dialect, which sounds like, pang guá tshiam miâ, which means, “help me sign” the jacket.

Jensen Huang was actually leaving the venue because he was already late. He turned back and he says, “Of course.” He walked back to me, and he says, “Turn around.” I gave him the pen and he signed on it. You’ve seen the picture. It took him almost 10 seconds to sign it. Initially, I was like, has it done? He says, “Don’t move.” Once I saw the signature, he really put an effort. He’s an artist. It’s great. I really thank Jensen for actually spending some time signing my leather jacket. Now I’m going to actually follow Jensen’s footstep, hopefully auctioning this for charity-

Anu Rajakumar: Oh, that’s great. Wonderful.

YT Boon: -to fund university AI research to give back.

Anu Rajakumar: Oh, that’s wonderful. I think I speak on behalf of our listeners. I’m glad that you didn’t get taken away by bodyguards before he was able to sign your jacket. I’ve seen on your LinkedIn profile lots of posts with your jacket. That is a very special story. Thank you for sharing.

Okay. YT, this has been an absolute blast. Thank you for being here today and for covering so much ground.

We started talking about the summer’s tech drawdown, but you also commented about the strong earnings growth and visibility for multiple years ahead. We also spoke about supply and demand, the supply chain that can’t quite keep up and demand coming across every sector which are getting involved. Really, the AI rollout has extended beyond the key sectors and into other sectors. You talked about some areas that you’re really excited about, like optical networking, while also acknowledging some of the vulnerabilities that we need to be aware of, including cybersecurity.

Really appreciate the comments on the US versus China in the AI race. I think both countries have something they can learn from each other. You also spoke about the future of AI. We don’t know exactly where it’s going to go next, but we’ve already seen some of the new innovations, smart glasses like the ones you’re wearing today, as well as physical AI and potentially robots coming into the house. Still very potentially far away, but also could happen fairly soon. A new chapter that nobody has quite fully priced in yet, but will be exciting no matter what happens. Finally, thank you for sharing your story about the jacket that Jensen Huang has signed. I felt like he was a little bit in the studio with us today, if only in spirit. Thank you again for joining us, and look forward to the next one.

YT Boon: Thank you.

Anu Rajakumar: To our listeners, if you’ve enjoyed what you’ve heard today on Disruptive Forces, you can subscribe to the show from wherever you listen to your podcasts, or you can visit our website at nb.com/disruptiveforces, where you can find previous episodes as well as more information about our firm and offerings.