Compiled & Edited: Deep Tide TechFlow

Host: David, Bonnie

Guest: John Wu (Former technology investor at Tiger Fund, now President of Ava Labs)

Original Title: Is TSMC's stock price seriously underestimated? Are Taiwanese engineers the "strongest moat"?

Podcast Source: Bonnie Blockchain

Broadcast Date: May 29, 2026

Edit Introduction

This episode's guest is John Wu, former technology investor at Tiger Fund and now President of Ava Labs. He stated that relative to other AI stocks, TSMC is still cheap; and that Asia holds a structural advantage in the hardware and electrical engineering aspects of AI, which are currently the most scarce resources.

He also unpacked two "non-price" logics of the current Bitcoin mini bull market. The OG whales have basically completed the chip transfer to Michael Saylor and ETFs, and the RIA (Registered Investment Advisor) channel, which controls tens of trillions in wealth, requires a four to six-month "educational cycle" before it truly recommends to clients.

Additionally, he disclosed for the first time the cooperation details of Broadridge, which processes $80 trillion in tokenized assets monthly, opting to land on-chain equity on Avalanche, and responded from an operator's perspective to the market panic over "AI disrupting SaaS".

Highlights

TSMC and Asian AI Hardware: An Underestimated Scarce Resource

• “TSMC has always been a low-volatility, stable, high-quality company, and its historical valuations have always been reasonable relative to its own history. But compared to other AI stocks, considering growth rates, it is still very cheap.”
• “Asia excels in the hardware and electrical engineering layers of AI, and this is precisely where many scarce resources are located, with TSMC likely at the top of this list.”
• There are multiple reasons to hold Intel, not just because the U.S. government has invested in it. With the arrival of the agentic era, CPUs have become scarce.

Continuous Innovation is the Only Moat

• “The reality of the technology world we live in today is that continuous innovation itself is a moat, and perhaps the only moat.”
• “The half-life of these quantitative models is becoming shorter and shorter, which is why some companies have hired countless PhDs, because new models must be constantly released.”

ETF, Quantitative and Retail: A Double-Edged Sword

• “The market today is essentially driven by quantitative and ETFs, with stocks being highly correlated like never before, rising and falling together, sometimes unrelated to the stocks themselves, purely due to ETF weightings.”
• “In the long run, the power of compounding and buying companies with relatively fair valuations will prevail; you wouldn't want to hold those crazy stocks at the wrong time, as they would be severely hit.”

AI Disrupting SaaS: An Operator's Internal Perspective

• “As long as someone on my engineering team comes up and says 'look what we can do', all SaaS stocks will drop, and safety stocks will drop too. If I were still a hedge fund manager, I would short those immediately.”
• “Removing a piece of technology doesn't mean you can simply replace it with another, because the next piece must be compatible with everything already in the tech stack, meaning you have to replace the entire tech stack.”
• “25 years ago, every time it was rumored that Google would enter travel, stocks like Expedia and Priceline would crash; 25 years later, they are still around and doing better, while Google has never been able to enter the travel sector.”

Bitcoin: Handovers Completed, Channel Funds on the Way

• “Around this year in February, I believed we saw the bottom, this is a transition from old to new, with old-school Bitcoin holders selling their chips to Michael Saylor and those ETFs.”
• “The RIA channel controls tens of trillions in wealth. Being allowed to recommend does not mean they will do it immediately. You need these advisors to be educated first, and this process takes four to six months.”
• “Crypto is entering a new phase; it is no longer the newest and coolest thing. Some people prefer to go from 0 to 1, while others prefer from 1 to 10.”

RWA and Value Capture: The Overlooked Infrastructure Layer

• “Everyone wants tokenized Real World Assets (RWA), but they forget that the underlying needs a system to operate smoothly.”
• “Value must be captured along the entire chain; top-level applications must be able to run a business, and Layer 1 also needs to clarify its token economics to survive and serve these applications.”

The Underestimated TSMC and Asian AI Hardware

David (Host): I am pleased to invite John Wu again. John is the President of Ava Labs (the development company behind the Ethereum-compatible public chain Avalanche). Before joining Ava Labs, he spent many years as a technology investor on Wall Street, managing his own fund, and had previously served as a portfolio manager at Kingdom Capital. We are delighted to have him talk about the future of technology investment and blockchain investment. Welcome, John.

John Wu: David, it’s great to see you again.

Bonnie (Co-host): We were just discussing TSMC. I have a heavy investment in TSMC and can say I am a whale-level holder, having held it for a long time and I’m very passionate about it. You just said it is not considered expensive.

John Wu: I know quite a few investors in TSMC. As David mentioned, I have been a professional tech investor for a long time and know not only professional hedge fund investors and venture capitalists but also operators within these companies. TSMC has always been a stable, high-quality company with low volatility, and its valuations relative to its historical position have always been reasonable. Because it is in such a crucial position and is the best performer in its peer group, it might appear a bit expensive now. But compared to other AI stocks, considering the growth rates, it is still very cheap.

Bonnie: Is it because of geography?

John Wu: I think there is some discount due to geography. But I also feel that overall, people’s attention to this company is not as high as when it was perceived as a U.S. company. U.S. companies naturally get more exposure.

Bonnie: So if I have the same amount of money and face two equivalent companies, would you choose the U.S. market?

John Wu: Let me put a disclaimer: I was an investor, and now I am merely an enthusiastic observer. I have friends who are still investing and keeping up with this field, so this does not constitute investment advice. If I had a portfolio, I would definitely allocate some, or even a higher weight to TSMC. And not just TSMC, there are many AI supply chain-related companies in Taiwan, many memory companies in South Korea, and Tokyo Electron in Japan. Asia excels in the hardware layer and electrical engineering layer of AI, which is exactly where many scarce resources are located. So there are many companies in Asia worth paying attention to, with TSMC likely at the top of this list.

Bonnie: But would you bet on U.S. companies being acquired by the government or at least receiving government investment? This is exactly what happened with Intel; it jumped double digits in a few trading days. I'm not saying it will necessarily happen, but comparing East and West, isn’t that a factor to consider?

John Wu: As an American, I hope Intel can build enough precision in its foundry business and create jobs in the U.S., but they still have a long way to go. Intel is also strong in making CPUs. One of the reasons for its increase is not just good management, not just their attempt to replicate foundry-level success in the U.S., but also because with the arrival of the agentic era, CPUs have become scarce. So there are multiple reasons to hold Intel, not just because the U.S. government has invested in it.

Bonnie: A few weeks ago, I was at Mar-a-Lago, and President Trump said, "We want 50% of chips to be made in the U.S."

John Wu: The President wants jobs and manufacturing to return to the U.S., which is not wrong; I think it's a good thing. He has been very consistent on this theme and is willing to support U.S. companies. As I just said to David, this is not a bad thing.

Bonnie: But I want to know your position on whether this could be achieved?

John Wu: In what sense could it be achieved?

Bonnie: For example, in Asia, when an earthquake occurs, engineers do not first look for their children or spouses, but head straight back to the company.

John Wu: Are you talking about Taiwan? This kind of thing has always been happening. In Japan, there is a specific term for dying at the desk due to overwork, and it is considered a very noble way to die, unlike being hit by a bus.

Bonnie Chang: I can’t really say about that.

John Wu: There is really such a term, believe me. Everyone appreciates the hardworking professional spirit in Asian society and culture.

Bonnie: But it does not mean Americans don't work hard.

John Wu: Yes, Americans have a slightly different culture, for instance, a strong entrepreneurial spirit and a greater willingness to take risks. If you can blend the two together... that’s what I did when I started my entrepreneurial journey at Ava Labs, in the early stages we had a team of just 10 people, all engineers from Cornell, many of whom were Asian, possessing a hardworking professional spirit; but it didn't matter, we were all diligent while also having an entrepreneurial spirit. If you can combine these two elements, that would be an ideal scenario.

How to Design a Portfolio

David: Before we continue, I want to ask both of you a question about TSMC: how long can this company's technological advantage be maintained? Because the core logic here is that, at least for now, no one can do what they can do.

John Wu: Unfortunately, when I began investing, the technology cycle was longer; now the technology and economic cycles are much shorter. Nvidia and AMD also face similar issues. The reality of the technology world we live in today is that continuous innovation itself is a moat, and perhaps the only moat. Even in quantitative trading firms, the half-life of these quantitative models is becoming shorter and shorter, which is why some companies have hired countless PhDs, because new models must be constantly released. This is continuous innovation. Ultimately, it has the potential to bring good results for society, but those involved will find it very hard and must constantly think of new ideas.

David: This insight is enlightening. So how would you design a portfolio now? We’re talking about AI, energy, and defensive stocks?

John Wu: A professional portfolio and a personal portfolio are two different things; the latter is for ordinary people with daily jobs.

David: Let’s start with personal.

John Wu: For that individual, I think this is an incredibly exciting time, reminiscent of the Internet era. If you can find a portfolio, even if you get some wrong, if you can identify today's Amazon, Google, and Nvidia, as an individual you don’t have to deal with daily fluctuations, just pick out the best companies across all segments in the AI world, and then sit tight and hold for the long term. Imagine 25 years ago; you might pick some wrong ones, but if you identified Google, Netflix, and Amazon, you just hold on and ignore the ups and downs.

David: Now let’s talk about the professional portfolio.

John Wu: The caliber of professional portfolio managers today is astonishing. We are talking about discretionary investment managers, not quantitative. They are very adept at identifying every bottleneck, which is why you suddenly see a shift from single bets on Nvidia to network equipment, then to photonics, and memory, with Micron and SK Hynix in Korea—the bottlenecks are different, and they continually find the next bottleneck and position themselves ahead of time. Micron is a perfect example; it was previously seen as a cyclical and slow memory business, and in terms of price-to-earnings ratio, it still looks cheap, but people do not know how long this will last. But its stock price seems to trade as if it has no earnings, while I believe its trading price is less than 10 times next year's earnings.

David: I heard that Micron was hyped up on Reddit.

John Wu: Well, you are entering a completely different realm. There are pros and cons to that, and I have mixed feelings. When I was a professional investor, everyone focused more on valuations, more on active management, believing that they could genuinely create alpha. The volume traded in systems like Nasdaq, NYSE, etc., is basically contributed by active managers, whether mutual funds or hedge funds. Now the market is essentially driven by quantitative and ETFs, with ETFs accounting for a large portion of the volume. So stocks are more correlated than ever, sometimes completely unrelated to the stocks themselves, simply due to ETF weightings. For example, during times of safety panic, companies like Palantir that perform extremely well, because they are classified within some software ETF, will also drop as they are dragged down. So back to the Reddit thing, my contradiction is that I like retail investors participating, feeling that they can engage in this world, which is great; but at the same time, I worry that their excessive speculation may reduce their focus on valuations. In the long run, the power of compounding and buying companies with relatively fair valuations will prevail, which is the reason Buffett is great. You would not want to hold those crazy stocks at the wrong time, as they would be severely impacted. I still remember during the internet era, some people bought into pets.com and never recovered; it’s unfortunate that if they had just held onto those 10 to 15 really good stocks, or bought an index, they would be doing quite well now.

David: We were just discussing the old forces and new forces in today’s investment era. For instance, the IGV software index (U.S. software industry ETF) has dropped about 27% this year; some would say the market is pricing in "this is the old force." You just mentioned Micron, theoretically asking, since the current innovation direction is reducing model use of memory and RAM, why are there still investments in RAM companies? For example, when Google’s Turbo Coin (which refers to a model that reduces memory usage) was released, Micron's stock price dropped 12% to 15% on the same day. Also, remember that even following a 12% drop, it had already risen about 300% before that.

John Wu: I believe it is about 50% higher than it was then. First of all, just because a technology is announced does not mean it will destroy the entire industry. I am now an operator. Interestingly, most of my team are engineers, very forward-thinking, and they enjoy experimenting with new tools and various vibe coding. We basically coded our own CRM software, cut out two service providers, and created our custom system; we only need a database, and handle everything else ourselves; we also use AI for vulnerability testing. So for a while, whenever someone on my engineering team came to me and said, 'look how cool we can do this,' all SaaS stocks would drop, and safety stocks would drop too. If I were still a hedge fund manager, as soon as my team brought me such news, I would short. But they also tell me more detailed situations, like you can't completely disintermediate many aspects; a lot of claims are actually overhyped. Yes, if you are just a UI/UX level, high-level software automation tool, then it is clear that sustainability will be very difficult; but you cannot remove those core codes. Although the user levels can be rewritten, everything below that—core consensus, core blockchain, deep tech code—cannot be replicated. As for the memory issue, you are actually assuming two things: one is that memory companies themselves are not innovating, unaware of what is happening; the second is that you underestimate the 'business' itself, which is far more than just technology; it also involves their existing networks and connected supply chains. When you remove a piece of technology, it does not mean you can easily replace it with another, because the next piece must be compatible with everything that already exists in the tech stack, which means you have to replace the entire tech stack. This is exactly the lesson the crypto industry is learning; if you want to build something from scratch, there are indeed better approaches, but in the real world, there are legacy systems, rules, and massive businesses; you cannot rely on recklessly shifting to new things and failing; you must progress stepwise.

David: By the way, during the Internet era, you probably remember: every time it was rumored that Google was entering travel, Expedia, Priceline, and Booking.com stocks would crash. 25 years later, they are still around and doing better, while Google has never made it into travel.

Bitcoin and RWA Value Capture

David: I want to talk for a few minutes about Bitcoin. If Bitcoin closes at 76,000, we will enter a new bull market. This is what Tom Lee (co-founder of Fundstrat, a well-known market strategy analyst) said a few hours ago at this conference; I am reading a review of his speech. He stated that during the bear market, Bitcoin has never had three consecutive months of closing gains; if Bitcoin closes above 76,000 this month, the bear market will definitely be over. He also said that agentic AI financing is part of the reason this wave of prosperity will continue. This is his viewpoint.

John Wu: How much is Bitcoin now?

David: 81,000.

John Wu: So he’s basically saying the bear market is over, right?

David: If it closes at the current level by the end of the month, we are above 76,000. It’s May now, and we have 23 days left; as long as we maintain the current level for 23 days, the bear market will end.

John Wu: Okay, so what’s your question?

David: Do you agree with this viewpoint?

John Wu: First, I really like Tom Lee; I've known him for 25 years. When I was on the buy-side at Tiger Fund, he was a technical analyst at JP Morgan. He is very smart and talented with statistics, and can pull up a bunch of numbers easily, so I never question his numbers. Speaking of my view, around this year in February, I thought we saw the bottom, and we are experiencing a transition from old to new, with old-school Bitcoin holders selling chips to Michael Saylor (founder of MicroStrategy and a representative figure of Bitcoin accumulation strategy) and those ETFs. I think this situation stabilized earlier this year, which is also part of the reason Bitcoin's performance has improved. I cannot specify exact timing or the price at year-end, but I really believe there are two things helping Bitcoin, or at least propping up the current mini bull market in recent months. Firstly, the handover of holders or chips has occurred—the uncertainty has passed. Secondly, an important point is the RIA channel, which controls tens of trillions in wealth. Being allowed to do something does not mean they will do it immediately. For instance, Morgan Stanley said they would be able to conduct related business through their wealth channel, but you need those individual advisors to be educated first, and only after they are educated will they recommend to clients; this process takes four to five or six months. Now it’s May, and since the beginning of the year, these channels have been allowed to recommend to individual investors, and the advisors have been educated, and end users are informed enough, so the flow of funds from this group will also be helpful. Based on these two reasons, combined with my belief that the macro environment is calm and quite good, and there is hope for the passage of the CLARITY Act (U.S. crypto market structure legislation), these are all reasons I believe we will see a nice Bitcoin mini-market.”

David: Can you expand on the “we are in a calm situation” point?

John Wu: What I just said is that there has been a massive handover of holders. Now entering the Bitcoin space are a new group of holders, many of the OG whales are either no longer around or have reduced their positions to a level where they don’t need to offload further.

David: Why do you think this is?

John Wu: I think there are several reasons. I was initially attracted to tech investment because I love innovation and new things; it's exciting, and I think everyone is like that at 18. And crypto is now entering a new phase; it is no longer the newest and coolest thing. It’s like asking why some serial entrepreneurs like to take a company from 0 to 1, then hand it off to someone else to take it from 1 to 10? Why not do it all the way from 0 to 10 themselves? Because some people enjoy going from 0 to 1, while others prefer from 1 to 10. So for some of them, it has already reached a natural exit point, especially now that liquidity has finally come in.

David: Broadridge (a global technology leader, NYSE code BR) processes $80 trillion in tokenized assets monthly and is now expanding its business into the digital asset field, partnering with Avalanche. Please talk about the significance of this cooperation.

John Wu: This is fantastic and very important. Broadridge is a leader in the traditional securities world, responsible for processing corporate actions, proxy voting, and many things most people don’t realize but require someone to manage on their behalf. They have announced a partnership, and we are proud to say they will record on-chain, tokenized equity on the Avalanche blockchain. Soon, Galaxy Digital's equity will be recorded on the Avalanche chain through Broadridge. This is a significant part of traditional equity, so I see this as a major move. How does it fit into the world I see, into the vision of Avalanche and Ava Labs? Since we began communicating six or seven years ago, we wanted to serve both crypto-native clients and become the default infrastructure for putting all institutional assets on-chain. What we do is to build all the various components and pieces together. We understand it as having an asset layer and an infrastructure layer. The asset layer is all those things—billions in stablecoins, money market funds and treasury bonds, private credit, private equity, various funds, even public equity tokenized on Avalanche; while Broadridge is embedded in the infrastructure layer, which is the underlying system necessary to make everything run smoothly. Everyone wants tokenized real-world assets (RWA), but they forget that the bottom layer needs a system to operate seamlessly. There are obviously crypto-native infrastructures like wallets, fiat and crypto entry and exit channels, custody, etc.; but there are also traditional financial aspects—who handles corporate actions, who manages those operations that make everything work smoothly regarding ownership, proxy voting, compliance, regulation, etc. Broadridge is one of the partners and pieces helping to build the future infrastructure layer for the RWA world. So this is a great thing, and we can't wait to cooperate with them.

David: For investors, the key question is always value capture. StraitsX is a new Singaporean company that is working with you, creating its version of a stablecoin and has significant influence in Asia. But my question is, for all these companies building applications and apps on top of your infrastructure, who captures the value?

John Wu: First, back to the first part. StraitsX is a fantastic project; they have their own stablecoin and clearly target Asia, and they are doing this on Avalanche, which indicates how they view Ava Labs and Avalanche. They also integrate with Alipay and Grab Pay, giving them a large distribution channel. We are proud to cooperate with them and to be chosen by them. That’s what I want to say first. Now regarding your question of who is there, in what way capturing value? This is a very good question. Let’s exclude Bitcoin from this discussion: Bitcoin's product-market fit is “digital gold”; some believe in it, some do not, but now that is roughly its positioning, occupying about 50% of the total market cap of approximately 2.7 trillion. So what about the rest? The question is where value capture occurs. I believe value must be captured along the entire chain, ultimately at the top-level applications we call DApps (Decentralized Applications); they must be able to run a business, providing a reason to stay on some infrastructure; while Layer 1 must also clarify its token economics to survive and serve all these applications. The architecture of Avalanche is a “network of networks”; you have a main chain, but you can launch your own customized, dedicated, self-sovereign blockchain. Many traditional financial institutions keep it in closed private chains, which do not belong to the permissionless part, and they are already capturing value just because they are in their private chains. We capture a little bit of value because they have to pay us for it; but ultimately they want a connection to a world that links to other hundreds of Layer 1s within the Avalanche ecosystem, and at some point to connect for liquidity to the Avalanche main chain. It is EVM (Ethereum Virtual Machine) compatible, so they can connect to the Ethereum world. So I believe this starts from customers being able to create value for themselves, and once they start connecting, natural value will flow to the rest of the Avalanche ecosystem.

David: A very good answer, thank you very much. I wish for the continued smooth evolution and success of Avalanche, let’s talk again next time.