DeepSeek's Disruption and What It Means for Your Investments
en
January 28, 2025
TLDR: Investor Stephen Yiu discusses reduced stakes in major US tech companies due to AI costs concerns by 2024 and his current investment outlook. His fund is backed by billionaire Peter Hargreaves.

In the latest episode of Merrin Talks Money, Stephen Yiu, founder and manager of the Blue Whale Growth Fund, shares insights into the recent sell-off in big tech stocks and the transformative potential of artificial intelligence (AI). Yiu discusses his investment strategy in the context of evolving market dynamics, emphasizing how to identify opportunities amidst uncertainties.
Key Takeaways
Understanding Market Dynamics
- Recent Market Trends: Yiu highlights a notable reduction in stakes in major US tech companies like Microsoft and Meta due to rising costs associated with AI development.
- Growth Fund Performance: The Blue Whale Growth Fund has yielded an annualized return of over 14% since its inception in 2017, significantly outperforming UK peers.
- AI's Economic Impact: Yiu expresses confidence that AI will revolutionize sectors, proving a deflationary force that may increase productivity and lower costs.
Investment Philosophy
- Quality Focus: Yiu emphasizes investing in high-quality companies with sustainable business models, capable of maintaining profitability in changing environments.
- Concentration Strategy: The fund targets 25-35 high conviction stocks, favoring comprehensive research and depth over quantity.
- Market Sentiment: Yiu stresses the importance of differentiating perspectives to outperform passive strategies and adapt to market shifts.
Reflections on AI and Tech
- Big Tech Concerns: The shift in strategy involves scaling back on some traditional tech giants, reflecting concerns about valuation versus performance.
- DeepSeek Breakthrough: The advent of solutions like DeepSeek, a Chinese AI product, threatens established players by lowering the cost and energy intensity of AI applications, prompting a reevaluation of investments in traditional infrastructure.
Opportunities Ahead
- Emerging Tech: Opportunities abound within AI and its applications, as companies rapidly adapt to new efficiencies without excessive capital expenditure. Yiu notes that such innovations can catalyze a new wave of tech companies and solutions.
- Sustainable Growth Sectors: Yiu discusses how sectors like gaming, particularly sports betting, are emerging in the US market due to shifting consumer spending habits, offering robust growth prospects.
Strategic Outlook
- Long-Term Perspective: Yiu’s research is grounded in long-term forecasts, targeting free cash flow projections over three to five years to assess company sustainability and potential market share capture.
- Market Adaptability: The fund's adaptability allows it to pivot away from overvalued companies while remaining open to innovative tech firms that present compelling growth narratives.
Practical Applications for Investors
- Evaluating Investments: Investors should consider both qualitative and quantitative measures to assess the sustainability of business models in tech and emerging sectors.
- Staying Informed: Keeping abreast of technological advancements, particularly in AI, will be critical in understanding market dynamics and investment viability.
- Risk Management: With increasing volatility in tech stocks, maintaining a diversified but highly concentrated portfolio may provide a buffer against market downturns.
Conclusion
Stephen Yiu’s insights on the interplay between AI developments and market investments serve as a roadmap for navigating the complexities of modern investing. The evolving landscape presents both risks and opportunities, requiring investors to stay informed and adaptable. As trends unfold, those who can identify quality businesses poised to thrive in these environments will likely succeed.
Reading Recommendations
- Yiu is currently reading Lucky Loser: How Donald Trump Squandered His Father's Fortune and Created the Illusion of Success, reflecting on the impact of personal narratives on public perception and economic policy.
By centering on quality stocks and leveraging understanding of AI’s implications, investors can position themselves meaningfully in this dynamic market.
Was this summary helpful?
Joel, the holidays are a blast, but the financial hangover. That can be a huge bummer. If you are out there and you're dreading the new statement email that reveals the massive balance that you may have racked up, well, you could use our help. That's right. I'm Joel.
I am Matt and we're from the How To Money podcast. Our show is all about helping you make sense of your personal finances so you can ditch your pesky credit card debt once and for all. Make real progress on other crucial financial goals that you've got and just feel more in control of your money in general. You know it for money advice without the judgment and jargon. Listen to how to money on the iHeartRadio app, Apple podcasts or wherever you get your podcasts.
The forces shaping markets and the economy are often hiding behind a blur of numbers. So that's why we created The Big Take from Bloomberg Podcasts to give you the context you need to make sense of it all. Every day, in just 15 minutes, we dive into one global business story that matters. You'll hear from Bloomberg journalists like Matt Levine. A lot of this meme stock stuff is, I think, embarrassing to the SEC. Follow The Big Take podcast on the iHeartRadio app, Apple Podcasts, or wherever you listen.
Are you looking for a new podcast about stuff related to money? Well, today's your lucky day. I'm Matt Levine. And I'm Katie Grifield. And we are the hosts of Money Stuff, the podcast. Every Friday, we dive into the top stories about Wall Street, finance, and other stuff. We have fun, we get weird, and we want you to join us. You can listen to Money Stuff, the podcast on Apple Podcasts, Spotify, or wherever you get your podcasts.
Bloomberg Audio Studios.
Welcome to Merrin Talks Money, the podcast in which the people who know the markets explain the markets. I'm Merrin Samsung, this week I'm speaking with Stephen Yu. Stephen is the lead manager of the Blue Well Growth Fund, a long-only and very concentrated global equity fund, which is the backing of British billionaire Peter Hageries. Stephen started his career as a fund manager at Hageries Lansdowne. He's since then held positions at Newstar, Artemis and Nevsky Capital.
Folders closure a few years ago, I did do some paid work with Blue Well, but I have not been commercially involved with the fund since 2023, so time has passed since then, and this week is actually the perfect time to get Steven on. First of all, the fund has an excellent track record, an annualised return since launch in 2017 of over 14% a year.
That's against the return of more like 5.5% for UK-based peers, so it's always good to get someone on who's fund is actually performing very well. It's also true that about a month ago, Blue Well started talking about reducing its stakes in some of the big, big US tech companies, including Microsoft and Meta, on concerns about the costs of the capex of artificial intelligence. Service holdings in Microsoft was, for example, cut
to about 2% of the portfolio from 8% at the beginning of last year. We'll talk a bit about that. Now, Bluewell is thought of as a tech investor that may or not be fair. We're going to talk about it. Stephen is increasingly less positive on the magnificent seven tech stocks all around. We are speaking on Monday this week and his cautious attitude is looking to be spot on. In this conversation, we get more about Stephen's thoughts on the tech giants, why the fund has been reducing that AI holding and the deep seek upset.
and of course where he sees opportunity right now. Right, the way I've said that, it sounds like there's an awful lot to get through and we haven't got days, we've only got 45 minutes or so, so sorry about that Stephen. Thank you for joining us, you're gonna have to talk fast.
Thank you, Mary. Right. So let's start. I've talked a lot about the already about the kind of things you hold in the fund or perceived to hold in the fund. So why don't we dial back a bit and talk about what you see the fund is actually doing, how you invest, what kind of stock you choose, what your valuation parameters are, et cetera. So we started the strategy back in September 2017. So it is over seven years ago now.
The objective at the time was very simple is we want to launch a high conviction mandate investing into between 25 to 35 stocks with a quality bias. So we like high quality businesses, but at the same time to be able to significantly outperform the markets, which would include the passive tracker in today's context.
And I think that has what we have been doing. So I think on a very high level that we naturally would exclude low quality businesses that we would deem as not sustainable, highly intended. And then the opposite of the spectrum would be company with very strong balance sheet, good management team, sustainable business model, et cetera. But of course, I think if you look at what had happened in the market with some other peers who are running a very similar mandate,
is that there's been a bit of a deviation in performance terms over the last couple of years. And so then when I reflected back on what we are trying to achieve investor on top of what I just outlined seven years ago is at the same time of not only investing into high-quality businesses, you would want to pick the winners that the companies are actually able to capture more market share or revenue or profit in a new environment. And I think that is what we are set out to do.
Okay, can I take you back a little bit to the definition of quality? One of the things you said when you were talking about what makes a quality company is having a sustainable business model. And by that, do you mean one that you believe will last for a long time? Or do you mean sustainable in the sense of ESG sustainable?
More to deal with the business model in itself. I mean, we are not a yesh or high impact fund, but we do deem yesh factors as a business consideration or business risk consideration. So what I'm referring to is a sustainable business model in terms of recurring revenue stream, cash flow conversion that you are able to tap into your.
Customers with high switching calls, which means that then it's not easy for them to move to a new competitor to to use something similar in the service or products. So that is what what we like or prefer to invest into.
That kind of company, this kind of company with a great balance sheet, with great cash flows, with great management, with a sustainable business model, and one that you can see continuing to generate profit growth out to the...out indefinitely, that's automatically going to be expensive, right? So, what are the valuation parameters that you have here?
build financial model for every single company we have, and then we would make sensible or conservative forecasts in terms of how much money we'll expect our company to make in the next three to five years. So one valuation metrics that we will look at is free cash flow yield adjusted for debt. So it's how much free cash flow the company would have generated in five, three to five years from today.
And of course, the way that we assess the quality of the business is the sustainability of the free cash flow level of generation, three to five years from today, isn't it? It's not like a one number that you get in five years time and then that number disappear in the six or seven years. So that is not what a high quality business means. A good company would mean that they can continue to deliver that. But of course, we do not have a crystal ball so that we cannot just make a number to say, okay, I think this company is going to do well.
10 years from today. So you want to have a sensible time horizon which for us will be between three to five years. And at the same time you will be able to track the recent developments such as quarterly earnings, investor day or when the company announced their new guidance for this coming year in terms of their plan. And then you can marry that with your business analysis alongside the forecast that you have made to determine whether the company is likely to achieve those results. So that's what we do.
But the point that I want to make just to go back on the quality spectrum of companies that we look at, obviously, they're surprised to pay as a starting point. But I think what is really important as you can probably worked out across different holdings that could be deemed as good quality businesses, that the performance potential or how they have performed over time are very different. So identifying a good company or high-quality businesses
The starting point is like, we do know that these companies are not going to disappear. So it's one of the things that I'm trying to get at is let's say you've bought a quality company. It's great. Everyone else company, companies doing very well. Is there a point when you look at it and you go, it's still doing really well, but it's too expensive?
Of course. So what's the point at which you go? Well, actually, it's now so expensive. I can't have it in my portfolio anymore. Or as a growth investor, does that point never exist because you expect a good company to grow into any valuation? You see where I'm going here obviously. Yeah, we totally agree with you. That's always a price for everything, right? Like even like for high quality company, let's say NVIDIA or Microsoft or some others, that's always a price. So we do take that into account. So the way that we look at it is two things.
So one is we make four cars in terms of how much money or free cash flow our company is expected to deliver in the next three to five years. And of course, up to three years, you can look at the market consensus numbers. So what we want is our forecast to remain ahead of the market expectation, which means that if we ended up to be correct based on our analysis, that a company would continue to beat expectations.
Right. And of course, that will be from time to time that the market has caught up with our expectations and we are no longer ahead of market expectation. And in that manner that we would deem the shares to be overvalued within a short time horizon, which is up to three years. That's one which we do look at. Secondly,
is about the five-year number, which we look at a single valuation metric free cash flow yield adjusted for debts in five years time. And the way that we look at it, obviously, there's no market consensus to refer to, but we were trying to adjust that number alongside the quality of the business. So if you're a good company, of course, there's a spectrum of good company versus better company, high quality company, et cetera. And the price would be very different from one to another.
And I think in today's context, as going back to Microsoft to make it a bit simple, is that we believe that the quality of the business for Microsoft is not improving from here. But when you look at the business seven years ago, when they went through the digital transformation to make it a cloud subscription model, that the quality of the business was improving for a good period of time.
But what is changing now is when you look at the valuation of Microsoft, let's say using a five-year free-careful yield number in the next five years, compared to previously when we look at Microsoft, that number is very much similar. So Microsoft is trading at a all-time high valuation compared to when we first invested. But then the quality of the business in our view is deteriorating.
So we think the share is overly expensive, but we're not here to suggest that Microsoft is going to disappear as a company, right? So we are still going to use Microsoft continuously. But when you look at the return on investor capital profile for Microsoft from here in the next five years, that number is likely to come down rather than going up from in our view base on how much money they have spent into artificial intelligence investment. So if I
just link it back to the context of NVIDIA. I'm sure we'll talk about some other stuff like deep-seek later on. But when you look at a very simplistic P-ratio one-year forward consensus on Bloomberg number, NVIDIA and Microsoft is trading at the same level, which is about 30 times earnings. But of course, from NVIDIA to be able to sell you many more products, let's hope that is going to continue that journey. And
It's higher than what Microsoft has been doing in terms of the co-pilot. And of course, the end customer for Nvidia would be coming from the free cash flow generation of Microsoft previously. So that number that free cash flow has been passed on to Nvidia's top line. OK, so in part, to go back to what you were saying at the beginning of the answer to that question, a lot of it is about wanting or needing the rest of the market to have a positive surprise relative to you.
That's right. Yeah. So this very simplistic way that investors should be thinking about is in order to do a lot better than your peer group, the markets, the ETF, is you need to have a differentiated view on the outlook of a company. And that's more likely to be correct than wrong, because it's very easy for anyone to have a contrarian view. You can take a view on a company on a market to be a contrarian for extended period of time.
and you're probably going to get carryouts in terms of that view because the market is right most of the time, right? So then in order to do something different, which is in performance term that you need to deliver more than your competitors or the market, you need to have a differential angle and you need to maintain the differential angle with the probability that the pressure angle is likely to be right or maintain it to be correct. And as long as there's still a decent gap,
Then we think there's likely that the market is going to follow us and vice versa if we end up to be behind the market and we are likely to be correct and we will think the shares could be expensive because the market is going to follow us that maybe there would be earnings revision downgrades on the other side.
Yeah, and this is easier said than done, of course. And as you say, the market is more often right than wrong. And so if you're going to run a portfolio, like this really has to be a concentrated portfolio. So 25, 26, 27, that's the amount of companies you can realistically find that will fit inside these parameters.
That's right. And the other points that I want to just illustrate to the audience that this is very important in today's context is obviously the market is very efficient. Everyone have access to the same level of information. In order to have this differential angle, you need to do something original. You need to make it very deep dive in a way that you have outworked your competitors or your the market audience. So the way that we try to translate that at Blue Well is that our five or fast on the investment team.
Okay, this doesn't sound like a lot of people, right? But when you marry that with the number of stocks that we have in the fund, which on average 25, then you're saying that we're saying that there's on average an individual covering five companies. So if a professional full-time workers is covering five companies for the entire year versus another individual who may be covering 10 or 15 or more,
and maybe not full-time too because you have a day job, then the level of research that we could put behind the scene to understand what the company is trying to do or whether we have a friendship angle. I think the chance of that happening is higher. We're not an expert on anything, right? If you ask me about the stock that's outside the fund, I probably cannot have a proper conversation. But for the stock that we end up having invested, then yes, I think we could claim to be an expert just because we have spent a lot more time
when you're very focused in terms of that resource. So one of the things that has happened as a result of the way that you invest is that you have ended up quite a heavy tech investor. And as I said at the beginning, a lot of people when they hear blue whales, they think tech, they think of you as a tech investor as one of your biggest tech areas has been in AI. You know, you and I have talked a lot about your position in video before. And let's go back then to Microsoft to matter to why you've been selling out for some of these stocks, but staying in the video.
So just on a very high level that we, we, we still say that we are global fund, but with a tech bias. And if I look back to the last seven years that our tech exposure has been fairly consistent, I would say just about 40% throughout that journey. But what has been interesting within that exposure.
was that the composition of that mix was very different seven years ago compared to now. So seven years ago before the pandemic, we were heavily exposed to digital transformation opportunities. And since then we have moved on and since 2022 or 2023, we have switched those exposure into artificial intelligence, which is the hot topic of the last few years. And of course, we have done very, very well on the back of that.
But when you look at it for a very high level, yes, over the course of last seven years, we did have a bias to a technology company. I think the point that I want to make is when I try to look back on what happened over the last seven years, I've done a bit of analysis myself, looking at all the companies that's being listed in the stock market, whether they are American companies, British companies, European, Chinese or Indian companies,
That over the last seven years, the market share winners in terms of the world revenue shares has been American companies. I think people would have worked out even without these studies. But at the same time, if you look at which part of the North American company have got the most share in terms of the dollar or time that we're spending with them now would be the big tech companies.
And then what is it at the expense of the expense of some very traditional business model that are more maybe high street facing not in the digital world and hence I think as far as we are concerned that we like companies or sectors that can capture a larger share of the world GDP or when we say maybe a larger share of our dollars
spend or a wallet or maybe from an enterprise perspective, whether they invest their money into. And so we are very dramatic whether those companies would end up to be technology company, which has been the case for the last seven years.
But it might not be the case in the next seven years. So then if you ask me to go back in time, would I have done a lot better by launching a technology fund? I think looking at the stock pick that we ended up with, like, I think that performance will be even more exceptional than what we have achieved for the Global Fund, which is a global mandate.
But the one thing that I'm not changing is I don't want to be fixated on technology because as far as a global investor or a fund is concerned that we want to be able to navigate that market environment. But today, that is where you want to be, which is technology.
OK, so let's talk then about technology. And let's talk about AI in particular. We're talking, as I said in the beginning, we're talking on Monday. And over the weekend, there's all this information that come out about deep-seak with Chinese AI product and how it can now do everything with a cheaper kind of chip.
represent a genuine threat to everyone else who spent billions and billions of Catholics on developing AI products to suddenly find that this exists, right? So we see, as we're speaking, we've seen the video of 10, 11% and we're seeing the
tech across the board beginning to come off. So do you see this? Have you looked at it in particular yet? Do you see this as a threat to the big tech companies in the US? Do you think it's a bit overblown or does it justify your view that you need to get out of the companies that have been spending huge amounts on AI?
So we are a true believer in how generous AI is going to change the world, whether it's on a consumer basis or in the enterprise professional basis. So deep seek have actually reinforced our view that this is happening. So what deep seek has managed to do in terms of breakthrough, that it actually soft two things that has been the bottleneck of the AI industry.
The first thing is the price or the cost of training models of doing AI. And you might record the same ultimate, the co-founder of OpenAI recently make a comment a few weeks ago saying that the chat GPD version that charge you $200 a month, that he's seeing a lot of demand, but they are actually loss making, even by charging you $200 a month to use it. So to ask, there's no doubt that there's many more things that AI can do.
But then the question is, what is the price that you want to pay for that? So if that subscription is $2,000 a month, you might not be paying for that. But there's no doubt that there's enough use cases that people want to embrace AI in terms of their day-to-day. So what DeepSeq had managed to do is they managed to use some very dated NVIDIA technology to produce a very similar output to what chat GBT has done.
after they have spent billions and billions of dollars investing into that. And on the back of that is the big tech companies so far in the US in particular, they have a monopoly.
to drive AI applications. Because the entry ticket or the barrier to entry is whether you have the billions to spend on NVIDIA GPU. If you don't have access to the GPU compute, you cannot develop any AI application, right? So what DeepSick had managed to do is I think they have started a revolution for the startup AI company globally.
that anyone with no access to billions of investment can start to work on some AI applications. The other bottleneck, which DeepSeed have managed to solve, I think, is energy demand. As far as the current energy infrastructure is concerned, that is not capable to handle what is yet to come from the AI world, right? And hence, when you look at the big tech companies, they have been spending money now exploring the nuclear energy.
But of course, nuclear is not just not the answer because it does take a long time and you need to probably get the permission to build it. So it's not going to solve anything imminent. So what deep seek had managed to do is to optimize the software and the AI algorithm that they no longer need as much GPU compute, which means energy demand or in energy intensity for the use cases that we have so far to date.
could be replicated with lower energy supply. So it's actually very helpful in terms of development.
So what we have here now is the possibility of an AI model, which is much cheaper and much less energy intensive than previous assumptions. That's right. It's a big positive supply shock to the global economy, surely. So that will take us a big step further down towards the productivity boom that we've all been waiting for for decades now. So this is actually, it's a huge and disruptive supply change, yeah?
That's right. Yeah. So basically it has leveled the playing field that is more startup AI companies is able to join the race to be more creative, innovative. At the same time, it's lower the price or the cost to run certain applications.
So if you look big picture, then that should mean less inflation, more productivity. It's yet another wave of deflation coming out of China. I think on a very high level that is of our view, even before before this week is like AI is very deflationary over time, right? Like in terms of increasing productivity, in terms of maybe having fewer headcounts in certain more mundane kind of operations like back office or maybe some customer service function. So I think over time,
A lot of things that we're doing today could be done by AI. But of course, the question is like, how long does it take, right? Even that is so expensive to invest into AI as a starting point and to be able to come up with the applications that worked. And then you then end up, you need to have the customers happy to pay for this at the same time, right? So I think that's a lot of box to take.
But by lowering the course, I think it does open up a lot of opportunities. The analogy I want to make on this is if you look at the smartphone ecosystem, let's say the App Store on iPhone or Android, when you first launch, you probably have maybe 50 to 100 apps. But over time, as the platform become more level and in a way that is making more accessible to any software developers, app developers, now you have a few million apps in your App Store, which we don't even know.
of them. And the number of applications of use cases that you can do on your smartphone is a lot more. And hence, the penetration is not as Apple's winning, right? Apple is only a small player in the market. You have a lot of entry-level smartphone that can do most of the stuff. So this is what is going to be very helpful from the AI adoption penetration kind of conversation going forward. So what DeepSeq has done is accelerating the adoption and penetration of AI.
What does that mean for your portfolio? What does that mean for Nvidia in particular? Because we've talked about you selling down other companies, possibly ones that have over-invested in AI and you think that weakens them long-term. But Nvidia, you've maintained your holdings and you haven't wanted to get out of that at all. What does this news mean for them?
So we've seen the AI domain, we are still a pro AI infrastructure company. So companies who are on a receiving end, so including NVIDIA and Broadcom, for example, both of them in our top 10. And the way that to think about this, obviously our view has not changed yet. I mean, it doesn't mean that we might not change that view in time, but our view has not changed that by freeing up some of the GPU capability. Let's say if you assume that
All the big tech companies, the utilization of all the GPU they have got in-house is 100%. So basically, they utilize everything they have got and hence they need to continue to order more GPU going forward, right? Whether it's from NVIDIA or working with Broadcom to do in-house GPU. Okay, what deep-seak have managed to suggest is maybe there's certain level of optimization you can do to free up some existing capacity.
So rather than 100% utilization, maybe the number is like 50% utilization. So you suddenly have like a 50% free up capacity, you can work on other applications or development that you might not need to then order more GPU from NVIDIA or Broadcom. So that is the narrative happening in today's market.
Our view is a bit different. Our view is that sure, okay, that is fine. If everyone is trying to take a step back to say, oh, actually, we can take it easy now because our competitor have got capacity free up. We do have that too. Let's slow down the development of AI and not to invest more into GPU. If everyone is doing that, that's fine, which is not the case. People are going to utilize more of the free up capacity to develop more things.
I mean, the start, the AI use cases that we have been talking today or most people have access to. We're talking about GPT technology, which is like a searching algorithm or some model giving you a very strict answer, quickly correcting your email, suggesting a few things going on. But what we haven't talked about, which is very new now is happening, is called AI Agents.
A.I. agent means this agent would actually help you to do things. So instead of speaking to the B.A. appreciate the airway chatbot saying, oh, my flight got canceled. Can you help me to do the refund? Where do I go? And then they give you all the answer. Is that? No, no, they don't give you the answer. They don't give you the answer. OK. It doesn't happen. You know nothing about the B.A. chatbot.
But what is happening, if BA is smart enough, like when this AI agent comes along, is the AI agent itself not only able to answer your question, but at the same time it's going to fix it for you in the backhand, right? When I talk about AI agent, it would need to be embedded within the
back-hand system or British airway, not just on the surface chatting with you on some random queries, it's going to link up to the back-hand and hence like, when I say, okay, my flight got canceled, can I get me a fun bet? Done for you, like you get the payment straightaway, just talk about to the AI agent. So that journey has not even started, right?
So what I'm trying to say here is like there's many, many more use cases that we have not yet come to use and any free up capacity would basically accelerate those development without like basically saturating the utilization to date. And I think everyone is racing to do that. And the other thing which is very important if you look at what deep-seak have said or maybe some of the comments on Twitter have been
trying to tackle is like those GPU they have got is not enough to go forward right they managed to utilize a very limited data number of GPU to get to where they got to but that is not the end story that's not the end story that China no longer need any GPU because I would imagine if tomorrow that Trump decided or actually we want to become friends with China again we will let you have as many GPU as you want from Nvidia
I would best China is going to spend the biggest amount of money to buy up the entire GPU stock in order to take AI to the next level.
is a big breakthrough in terms of the success in optimizing the algorithm that some, it seems like maybe the Western company have not managed to do so, but it doesn't change the narrative that AI is still changing the world. There's a lot more things that you need to do, you can do with AI. And of course, in order to perform the trainings on AI, you need to have the GPU compute.
Interestingly, I'm looking at the analysts putting out almost immediate comments on this person today, on Monday as they say. I've seen all sorts of things from, you know, this will lead to a huge capex unwind and a mild recession in the US and an unwinding of equity inflows into the US and eventually a weaker dollar. That's all a bit OTT from what you're saying. You don't think analysts should be allowed to write instant responses, do you?
No, I think the thing with the deep-seak breakthrough is, I think that's going to be more data point that we'll get from some of the big tech companies. You probably want to hear from them, right? Because I suggest that, okay, they're going to free up some capacity. But what is that number, right? If you say to me, they can free up like 50% of the capacity. That's a lot. It's a big number. If they can free up like 20% of the capacity, that's not a lot.
So then I think that's still a debate in terms of what is happening. But the other thing I just want to throw it out there, because we're going to hear from Matta too, was a suck-and-burr on Friday, a few days ago, before the deep-sick announcement, Friday he announced or pre-announced the capital expenditure for 2025. So that number has gone up from $50 something trillion expected for 2025 into the range of more than $50 billion to $60 billion.
So that number's gone up by over $10 billion, but and I would imagine he actually was connected, right? Like he's not like he's going to hear the same news on Deepseeker's, you and I just reading on on the Zlumba channel. So by him doing that, just being a bit speculative was that he already know what what this means for them. They already have worked out that, okay, I we can maybe copy some of the
the optimization algorithm that deep seek have introduced because it's open source, but then they still need more GPU to do more things. So I think that's very interesting, but we will probably hear more from them this week because they're going to report their results. Okay, more on that later. So you're hanging on to NVIDIA and presumably you're hanging on to your other investments in the AI infrastructure area Tuesday. Vertiv is another one in your top 10, which they do cooling infrastructure for data centers. That's right. Yeah.
I mean, things have quite fluid, right? So at the moment, we're not selling. I think that's a starting moment, but we'll get more clarity as this week goes on. I think we need a few more data points to suggest whether our thesis might have changed from on that perspective.
Yeah. And eventually, you mentioned earlier something I was going to ask you about energy in a particular about nuclear. You know, if you're very, very pro AI, very pro data centers, AI infrastructure, et cetera, et cetera, does that make you want to lean into holding anything in the nuclear energy sector?
Not currently, I think just from maybe a public equity investor perspective, there are just not many of these companies around, right? Like if you ended up finding a company that have some exposure to nuclear, that probably have other stuff that you don't want to own, which is a typical maybe like a utilities company, they have like a nuclear operation or some nuclear plants.
So you don't get a pure play in terms of nuclear, right? And of course, then whether we would end up investing into uranium, which we couldn't, then there's a more of a very different game to play. OK, so let's move on then and look at some other parts of the portfolio, some other sectors and themes that I know you're interested in. And one of the ones that always pops out to me when I look at your portfolio is what you liked called gaming and I liked called gambling, flutter.
For a lot of, for example, gaming I think is a bit of a euphemism here. It's outright gambling, right? But you think that sports gambling in the US is an area that you are exposed to and you have great hopes for it. Yeah, so this is a very good illustration of what we're trying to achieve without 25 companies, right?
Because the way that you do it is not only you find a good company or a high quality business model and at the same time you want this company to capture new market share from someone else. So as far as US sports gaming or US gambling is concerned, sports gambling is concerned, that is a completely new opportunity since the US started to legalize sports gaming state by state.
few years ago. So at the moment, when you look at the market segment in the US, it's only about 50% of the states or population would have access. And then at the same time, it's a two-player market, which is, we like that in a way that's a draw, probably. So in the US, it's between Futter's sub-brand Confendure versus Troft Kings, which is a US-listed company.
And these two companies, they control over 80% of the market share. So they are a very big company. Let's do a few more small players out there, but then they do get a economic scale. But what we really like in a way to think about this. So between these two companies, their top line is about $20 billion. So these are the money that they would be
Uh, be keeping after they pay out all the winnings, right? So then the question to ask is like, where does this 20 billions come from? Right? Because a few years ago that number was zero, right? It's either the consumer in the US have spent less money with a restaurant, gone bought a few of things from Amazon or maybe gone a few times to cinema or some other discretionary spending. So this is what we really like in a way that is idiosyncratic.
That is not depending on Trump. It's not depending on the war in Ukraine or the Middle East. It's about American consumers shifting away some of the spending they're already doing to the recreational sports gaming or sports gambling or like whatever you want to call it, but it's a new market for them.
And of course, the other thing that we know, which is very positive, is the American sports market is multiple times bigger than the rest of the world sports market to start as a starting point. So we know that they love their sports. Secondly, the GDP per capita as the American consumers, they are a lot wealthier than the rest of the world. So when you put these two things together and then you have a new opportunity to capture some of the spending, then you know that
This is going to be very idiosyncratic and could be very positive. When you put all these things together, then you know that it's more likely that they would do well. The only thing that I would say is a headwind or a hurdle is the text.
coming from the state, because like when the state decided, oh, actually, it is a very profitable, or maybe you get this yes she angle that I don't want you to do too much. They probably would increase the tax, but what we would say, this is a win-win situation, right? Because the only reason a state would decide to legalize sports gaming is because they want more income.
So they probably don't want to kill the industry, otherwise then they might as well not to have legalised it in the first place. Yeah, I see that, that makes sense. You mentioned in your remarks there that this is an industry that isn't threatened by President Trump. Does his election change anything for you when you look at your portfolio, even your soap watermarks, so stock specific? But is there anything in Trump's plans
That changes the way you look at your portfolio or the specific companies inside it.
Yeah, so not so much on maybe stock on the bottom level, just because when we were running into election, we were very much indifferent in terms of whether Harris or Trump is going to win, so we were not positioned either way. And since then, we probably haven't actually got a lot of benefit on the back of Trump being elected, because I think there's a few policy that is probably going to benefit more from the stocks that we don't currently have exposure to.
I think what is interesting or what we are likely to get in this coming four years is, first, I think the headline from kind of driven volatility in the market is going to be a lot higher in terms of headline risk. So then he's going to make a lot of comments about different sectors or companies.
and the shares is probably going to move according to that comment. I think that it plays in the strength of a high conviction active investors who are actually doing some work behind the scenes to assess whether some of these are just a rumor speculation or some of that is going to get translated in proper terms in terms of impact. So I think that's one. Secondly, I think that's probably more of a macro debate, which we're not an expert of, but of course we follow the dynamics.
is whether Trump's policy is going to be a lot more inflationary, which means interest rate is going to stay high for longer or maybe it would even go out from here rather than coming down. So I think that's more of a macro-positioning in terms of valuation, in terms of whether money should be coming out from equities to fixed income and all that, but not so much on a bottom-up company perspective.
Not much you can do about money moving out of equities into fixed income is the given what your your fund does. OK, what else is in the in the fund that we haven't discussed that do you think would be interesting? So I want to suggest another very controversial stop, which is Philip Morris that has done very well for us, gambling and smoking. I don't know. And I have to stop writing this podcast as an ESG podcast at this rate.
It's the way that I want to reason for Filmora to just to have a very quick discussion was. So it's a new stock for us. I mean, we've only been holding it for over a year or so. So we haven't got it since the beginning of the fund back in 17. But what we really like about a company is it's a good company to start with, right? The margin, the return of investor capital profile, I mean, how sustainable the business model on a very high level, we all like that.
But what have changed for us, which I think is the important part, is when you look at Filamores as a business today, 50% of the business is in non-traditional cigarettes products. So the 50% is the one that is probably the more relatively healthier way of consuming nicotine. Okay, we can have a debate on whether... Well, we do have a debate about that, couldn't we? Because there's a lot of conversation about when you say non-traditional, you're talking about vaping, right? I'm talking about heat not burnt, and also
uh, nicotine pouches, by a British match. So not raping. So they don't do raping. So they don't do raping. Okay. They don't do raping. So that's not what, what we're after, but it's more nicotine pouches. They have this protocol I course, which they use heat, not burn technology to smoke a cigarette, but not basically making it combustible. So they use the heating technology. It's a solid complicated, but they, they pioneered that more than 10 years ago and they're market leader in a few markets globally.
So basically, the view that we talk is obviously, okay, nicotine, we can debate whether nicotine is good for anyone, but to us, it's more of a choice. It's like alcohol, like maybe it's not good, but people should have a way to consume the product they want to consume. But of course, from a company perspective or from the regulator perspective, it's want them to be consuming healthier manner, right, or less damaging manner.
Anyway, it's better that you don't consume any of this stuff. We can stop justifying it now. It's okay. So the point I won't try to make is like so 50% of the business now is in a non-traditional secret world, which is growing very fast because people are adopting a healthier way to consume nicotine. And secondly, the reason we like it is because since they've been tiled from Altria back in 2008, Philip Morris business is anywhere but the US.
But recently, because of the iCOS product, the heat not burn technology, which is the healthier version of a smoking a combustible cigarette, is they got the permission from the FDA in the US to market the product back to the US. So this is a completely new market for them. So the way that you think about, we think about Philip Morris, is that
This company has been around for years, I mean doing what it's doing, but there's a bit of a transition going on to start with, which is growing a lot quicker than the traditional business. At the same time, they have, on the back of this, they have got the license to go back to the US to take market share from your likes of Aotria or some other company that have been only in the US.
for a long time. So that is very idiosyncratic. And the reason I want to outline Flutter and Villamoris is just to showcase that we're not just a tech fund, that we do have some other opportunity. They're very different to your AI conversation. That's how you define tech, I guess. I mean, that, you know, heat not burn stuff that's a bit techy, isn't it?
I want to take you back to something that we talked briefly about earlier, the role of an active fund. And obviously, the active industry has been shrinking and shrinking and shrinking as a percentage of the industry as a whole over the last decade or so. And most people will now listen to what we're saying and saying, well, yeah, he's done all right. But in the main passive is better. Do you see there's any danger in that or any? I mean, we've been talking here and I and the other active fund managers who I talked to for years now about how it's going to turn around.
You know, the active fund will have a great resurgence, particularly anyone who invests in the way that you do in a very concentrated, high active share way, that your day will come again. Do you see signs of that over the next decade, or is it really just going to keep shrinking? We're shrinking as an industry, as we speak. That journey is going to continue for another good few years, I think.
I think the debate, this is very interesting in a way that all the time, if the majority of money in the market, which is not the case today, I think today is about 50-50 split, but let's give it another five years and then passive would become like 70 to 80% of the market in terms of flows or AUM, that it would actually make the market less efficient because it's a lot easier to compete with a passive driven market.
If you ask me to compete with another active investors, I mean, we're having a podcast here, I'm sure like that another team of people would be doing a lot of work on investment opportunities. So then it's also at the same time, it's a lot harder to work out what they're trying to do or how they're doing it. But when you're trying to compete with passive, this is my point I'm trying to make is it's very easy to beat the passive.
It's a lot more difficult to be an active investor, because you don't know what they're doing, you don't know the resource they have, you don't know the information they have. To be the passive, you only need to work out the top 15 or 20 companies that are very big.
And then you take an active view on why you're not owning them because you have something better, right? Like if you think you have something better than your max seven, feel free to not own them. Just make sure that you deliver the same level return of those max seven. But if you haven't got as good of a company to do the same thing, just own them. Or if you can own them more in size like we have, then you outperform the market, right?
So then I don't understand or this is the, I think the reason why the industry is dying a little bit. I think there's a bit of legacy before the rise of the ETF that there's many more funds, right? You have different fund manager with a different investment philosophy, different way of operating and then they're trying to sell you the approach. Oh, we are a value manager, we are growth manager, we are quality manager, we are UK manager. They're trying to sell you the approach and we are not going to deviate from our investment philosophy.
And of course, I think that was a period of time that it did work. But over time, actually, the markets worked out that the ETF can replicate a lot of this factor in itself, whether it's a quality basket of stocks, value stocks or growth stocks.
So then I think there's a lot of low quality in a way, a value for money proposition from an active manager funds perspective that it could disappear, right? If everyone is similar to what we're trying to achieve, which I do feel we do have a cohort of outperforming active managers that are trying to deliver more value for investors,
And I think the money will come back to the active because when everyone else has got passive, everyone is getting the same return. So then you want to be differentiator. I want to be wealthier than my neighbor or my colleague. And then so I probably want to have a bit of active. But at the moment, it's very difficult for people to differentiate a good active funds versus bad active funds. I think that at the moment, you just don't want to have any active.
Why don't we just go everything in the past but I think over time people worked out actually. I want something different, I want something better and hopefully we will still be here to join the conversation. And hopefully to it'll become clear to active managers if they want to survive a long time they need to take a leave out of your book and at least be a concentrated active manager, a real active manager as we might put it as opposed to an active manager who knocks around the benchmark hoping for the best.
There are two more quick things I wanted to ask you about. I know that you're not a crypto investor, but I did just want to ask you briefly about Bitcoin. Do you ever look at it? And we've seen it come down today. We're talking again. I repeat on Monday and as Nvidia and as the Nasdaq have come down quite a bit. So has Bitcoin followed? Do you see it just as a non-asset class that is simply correlated to tech stocks in the US? Or do you see it as something more special?
I think the way for people to think about Bitcoin, I'm sure people would have. A lot of people have made a lot of money from the crypto is that it's about the use cases. So from our perspective, the reason that we are not, firstly, we can't invest into Bitcoin as far as the strategy is concerned, but even personally, I'm not an investor in Bitcoin.
The reason only being is that I don't see too much of a use cases. So then it becomes more of a speculative asset class. I can see the argument. Why is that some sort of like digital goal in a way that you want to protect your money away from the US dollar or what the central banks are doing. You probably don't want to have too much stirlings. Maybe you can have a bit of a big con on that note.
But then it's all about the use cases, right? If you tell me in the next five years that Bitcoin is going to be used in supermarkets in our day to day, or maybe some of our outpay or bonus are going to pay in Bitcoin, then there may be, there's an argument to be made that you can justify why the crypto is trading at a certain level. But before that, which at the moment, we're still not seeing any of this.
then it's become more of a speculative asset class, which then is a very different playbook, right? Like if you know they're speculative, then the way that you try to make money would be more short-term rather than you take a long-term view to think, okay, why don't I buy in whole and thinking that Bitcoin would go up in high level. Okay, so you're not going to be launching a blue whale coin?
If there's demand that we probably want to, like Trump did, right? Yeah, I was wondering. That's worth, like, $10 billion, right? Well, the one that goes, right? Well, we can do a lot of dual coin, but who is going to buy that? I mean, I was thinking that we could have a new way to judge our podcasts. You know, the moment we just look at how many people listen to them, but a much more effective and immediate way might be to simply launch a meme coin for every single podcast and see which one reached the highest price.
Then we would know who would be our most popular guest without even having to bother to look at the statistics. That's right. I'll suggest that to our producer at the end. I'm sure she's all over it. Stephen, final question. What are you reading at the moment? What's the book on your bedside table?
Oh, actually, I've just started that. But I'm reading a book on Trump. I can't remember the name of the book, but it just came out. And so I'm just trying to learn about what we're dealing with in the next four years and get some insight in terms of Trump's upbringing and his exposure. OK, excellent. Thank you. I'm going to ask you to email us the name of that book so that we can put it into our show notes because we would like that very much. We like to know what everyone's reading. Thank you.
And so I think that that's probably it. I've spent far too long asking these questions, but thank you so much for explaining Deepseek and its impact to us. I think that's a very important thing to know about.
Thank you for listening to this week's Merin Talks Money. If you like our show, rate, review, and subscribe wherever you listen to podcasts and keep sending questions or comments to MerinMoney at Bloomberg.net. You can also follow me and John on Twitter or X. I'm at Merin SW and John is John Underscore-Stapak. Stephen, are you on Twitter? Sorry, X. Stephen, you, my name? Yep.
The show is produced by Sam Masari and Moses Andam Sound designed by Blake Maples and the executive producer is Brendan Francis Newnam. Special thanks to Stephen Yu. Joel, the holidays are a blast, but the financial hangover. That can be a huge bummer. If you are out there and you're dreading the new statement email that reveals the massive balance that you may have racked up, well, you could use our help. That's right. I'm Joel.
I am Matt and we're from the How To Money podcast. Our show is all about helping you make sense of your personal finances so you can ditch your pesky credit card debt once and for all. Make real progress on other crucial financial goals that you've got and just feel more in control of your money in general. You know it for money advice without the judgment and jargon. Listen to how to money on the iHeartRadio app, Apple podcasts or wherever you get your podcasts.
The forces shaping markets and the economy are often hiding behind a blur of numbers. So that's why we created The Big Take from Bloomberg Podcasts to give you the context you need to make sense of it all. Every day, in just 15 minutes, we dive into one global business story that matters. You'll hear from Bloomberg journalists like Matt Levine. A lot of this meme stock stuff is, I think, embarrassing to the SEC. Follow The Big Take podcast on the iHeartRadio app, Apple Podcasts, or wherever you listen.
Want to understand exactly how interest rate rises will impact your mortgage? Or how New York City gets fresh produce? Or exactly what on earth was going on over at FTX before the whole thing collapsed?
Twice a week, we sit down with the perfect guests to answer these sort of questions and understand what's going on with the biggest stories in finance, economics, business, and market. I'm Tracy Alaway. And I'm Jill Wiesenthal. And we are the hosts of Bloomberg's All-Bots Podcast. Look us up wherever you get your podcasts. The All-Bots Podcast from Bloomberg.
Was this transcript helpful?
Recent Episodes
Trump's Tariff Threats and Central Banks, Gold's Surge, and US Equities

Merryn Talks Money
In this week's roundup, senior writer and Money Distilled author John Stepek speaks with Opinion columnist Marcus Ashworth. They discuss what Trump’s tariff threats mean for your money, gold's surge, and why there’s more to equities than just US stocks. Marcus's latest column: https://www.bloomberg.com/opinion/articles/2025-02-03/trump-tariffs-ecb-dithers-while-euro-economy-flatlines See omnystudio.com/listener for privacy information.
February 05, 2025
Don't Leave Your Family a Nightmare When You Die

Merryn Talks Money
On this week’s personal finance edition of Merryn Talks Money, the conversation revolves around effective financial planning for one’s eventual death.Merryn Somerset Webb sits down with Matthew Hutton, author of Your Last Gift: Getting Your Affairs in Order, to discuss practical advice on how to navigate the inheritance tax, write a will, choose an executor, manage digital assets and ensure one’s heirs avoid delays when it comes to accessing money after your death. Read more: https://www.bloomberg.com/opinion/articles/2024-04-29/death-admin-don-t-leave-your-family-a-nightmare-when-you-dieMatthew Hutton's book: https://www.yourlastgiftbook.com/See omnystudio.com/listener for privacy information.
February 04, 2025
Reeves' Heathrow-Fueled Growth Plan, DeepSeek and the Effect on US Big Tech, and Mini-Bubbles are Bubblin'

Merryn Talks Money
Chancellor Rachel Reeves' growth speech and potential challenges are discussed with John Stepek, along with the implications of Chinese AI startup DeepSeek for US Big Tech. John advises investors to expect uncertainty in stock prices and never overpay for perfection.
January 31, 2025
To Get The UK Growing Again, We Need Cheap, Secure Energy with Simon French

Merryn Talks Money
Simon French discusses encouraging North Sea oil and gas production as a potential growth strategy for the UK, and recommends reading 'Digital Minimalism'.
January 24, 2025

Ask this episodeAI Anything

Hi! You're chatting with Merryn Talks Money AI.
I can answer your questions from this episode and play episode clips relevant to your question.
You can ask a direct question or get started with below questions -
What was the main topic of the podcast episode?
Summarise the key points discussed in the episode?
Were there any notable quotes or insights from the speakers?
Which popular books were mentioned in this episode?
Were there any points particularly controversial or thought-provoking discussed in the episode?
Were any current events or trending topics addressed in the episode?
Sign In to save message history