Latest / Merryn Talks Money / The Stocks That Could Survive an AI Bust

Transcript

The Stocks That Could Survive an AI Bust · Merryn Talks Money

0:02
Speaker 1: Bloomberg Audio Studios, podcasts, radio, news.
0:07
Speaker 2: But if you do believe that AI is very powerful,
0:10
Speaker 2: that it's going to bring some positive to the human
0:13
Speaker 2: population across the board, I mean, obviously, there's going to
0:16
Speaker 2: be some negative as well. Then what you really want
0:19
Speaker 2: to do is to invest into companies that are receiving
0:23
Speaker 2: the spending or investment, which we call them the picks
0:25
Speaker 2: and shuffle companies. But then when you are thinking about
0:29
Speaker 2: the company that investing into AI to make an AI model,
0:33
Speaker 2: to come up with different use cases and all that,
0:36
Speaker 2: that's going to be a trial and error process in
0:38
Speaker 2: a way that a lot of them are not going
0:39
Speaker 2: to survive, but some of them are going to make
0:41
Speaker 2: it to be very successful. So it doesn't mean that
0:44
Speaker 2: you can't pick them, but then I think this is
0:47
Speaker 2: more of a venture capitalist market, isn't it? You invest
0:49
Speaker 2: into probably 100 companies and then some of them worked out,
0:53
Speaker 2: most of them doesn't, but then you still make a
0:55
Speaker 2: lot of money.
1:06
Speaker 3: Welcome to Merrin Talks Money, the podcast in which people
1:08
Speaker 3: who know the markets explain the markets. I'm Merrin Somerset Webb,
1:11
Speaker 3: and this week I'm speaking with Stephen Yu, Chief Investment
1:14
Speaker 3: Officer of the Blue Whale Growth Fund, which manages more
1:16
Speaker 3: than $ 4 billion.
1:18
Speaker 1: Stephen is a major investor in technology and has built
1:21
Speaker 1: something of a reputation for identifying the likely winners and
1:24
Speaker 1: indeed the losers. of the AI boom. His record makes
1:27
Speaker 1: him a voice worth listening to, and we have had
1:29
Speaker 1: him on the podcast a few times before. From his
1:32
Speaker 1: launch in September 2017 to the end of August, the
1:35
Speaker 1: fund returned over 320%. That's against just under 130% for
1:38
Speaker 1: its benchmark. So far this year, it has outperformed the
1:40
Speaker 1: index by more than 40%. But with so much of
1:49
Speaker 1: the global market now riding on enthusiasm for AI, can
1:52
Speaker 1: investors really diversify their portfolios enough to survive a disappointment?
1:57
Speaker 1: And if so, which companies might still prosper if the
2:00
Speaker 1: boom begins to fade? Stephen, welcome to Merion Talks Money.
2:04
Speaker 2: Thank you.
2:05
Speaker 1: Now, you are not really an AI fund. I know
2:08
Speaker 1: that you've had so much of your investments in AI-related
2:12
Speaker 1: stuff over the last couple of years that people look
2:14
Speaker 1: at Blue Whale and they go, well, that's an AI fund,
2:16
Speaker 1: or that's a tech fund, or that's just a growth fund,
2:19
Speaker 1: whatever it is. And of course, you are none of
2:21
Speaker 1: those things. You're a fundamental stock picker. However, before we
2:26
Speaker 1: move on to talking about stocks and how you pick them,
2:28
Speaker 1: do you mind if we just have a chat about
2:32
Speaker 1: the AI environment at the moment?
2:34
Speaker 2: Yes, of course.
2:35
Speaker 1: Can we start with the dangers and why we're talking
2:40
Speaker 1: so much about the dangers at the moment?
2:42
Speaker 2: I think that's a good starting point just because if
2:47
Speaker 2: you think about the pace of how the AI cycle
2:51
Speaker 2: has been gone through over the last few years, it's
2:53
Speaker 2: been very much on an exponential trajectory, which means everyone
2:58
Speaker 2: is now feeling part of AI, whether you have ended
3:02
Speaker 2: up investing in the AI or not, that's a different story.
3:05
Speaker 2: And of course, if a technological advancement is taking shape very,
3:09
Speaker 2: very quickly without all the other gut rails around it,
3:14
Speaker 2: then we could end up in some troubles in time.
3:17
Speaker 2: And hence, I think we now realize as a community
3:20
Speaker 2: that we need to start probably putting some policies in
3:23
Speaker 2: place and thinking about the implications of what this might
3:26
Speaker 2: bring is not all positive. There's certain negative implications. But unfortunately,
3:31
Speaker 2: I think maybe the policies that are still very much
3:34
Speaker 2: behind the curve and obviously thinking about the committees, how
3:37
Speaker 2: many meetings they need to do until they make some
3:39
Speaker 2: changes in terms of what AI is about or how
3:42
Speaker 2: we're going to deal with it is going to take
3:44
Speaker 2: a bit longer. There's been a narrative to suggest that
3:46
Speaker 2: we're going to see a lost generation. And what it
3:50
Speaker 2: means is like by the time that we would have
3:52
Speaker 2: something public that we can actually use and control AI,
3:57
Speaker 2: it might take longer than expected.
4:01
Speaker 1: But controlling AI is a global business, right? You can't
4:05
Speaker 1: just have one country controlling AI. You can't put regulation
4:08
Speaker 1: in place in the UK and the EU or just
4:11
Speaker 1: put regulation in place in the US. So when you
4:15
Speaker 1: hear these calls for regulation, it's very hard to understand
4:18
Speaker 1: how that could possibly work.
4:20
Speaker 2: It's probably not going to work. This is a competition.
4:22
Speaker 2: It's not going to work.
4:25
Speaker 1: Okay, right. Okay, fair enough. So we're not going to
4:27
Speaker 1: have a regulatory effort that slows AI down, but might
4:31
Speaker 1: it be the case that we've reached a point where
4:33
Speaker 1: it naturally slows, because after a very fast shift, we're
4:38
Speaker 1: hitting a couple of bottlenecks, a couple of places where
4:42
Speaker 1: things have to slow down. So we hit that in
4:44
Speaker 1: terms of power, in terms of resources, in terms of
4:48
Speaker 1: the capacity of what LLMs can really do, and we
4:50
Speaker 1: haven't made the next leap. So Might it be that
4:54
Speaker 1: we're about to see a kind of natural slowdown in
4:57
Speaker 1: the move of AI anyway while what's happened so far
5:01
Speaker 1: consolidates inside the economy?
5:02
Speaker 2: I think that's definitely probably a given in time. Just
5:06
Speaker 2: think about in any technological advancement, think about the penetration
5:12
Speaker 2: of internet, the penetration of smartphones. how we have evolved
5:17
Speaker 2: through the digital transformation, through COVID, shopping online, working from
5:21
Speaker 2: home and all that. It took a long time to
5:23
Speaker 2: get there. And obviously, there's always this thing about the
5:27
Speaker 2: rate of change, like it's how quickly that would have happened.
5:30
Speaker 2: How long does it take? And I think when you
5:32
Speaker 2: look into AI today, that there's a lot of capital
5:35
Speaker 2: that has gone into the space, which is the Quebec cycle.
5:38
Speaker 2: There's a lot of money that has gone into developing
5:41
Speaker 2: the model to make the model more advanced. But what
5:45
Speaker 2: we have not seen yet, which takes a bit longer
5:48
Speaker 2: than just investing into a model to make it better,
5:51
Speaker 2: is the penetration is taking time to get it penetrated
5:55
Speaker 2: within the consumer space, within the enterprise space, within the
5:58
Speaker 2: government space. So then what it means is in time
6:02
Speaker 2: when the technological advancement slows down, then the penetration can
6:07
Speaker 2: start to take shape. Obviously, you want that cost of
6:10
Speaker 2: compute to be quite low so it's affordable for most people.
6:14
Speaker 2: And hence, it's still a question about for the technological
6:20
Speaker 2: advancement of models to slow down doesn't mean AI is
6:24
Speaker 2: unreal because you still have a lot of use cases.
6:28
Speaker 2: yet adopting just think about all the art all the
6:30
Speaker 2: productivity that we are yet to realize across.
6:35
Speaker 1: Functions and then we also i mean at this point
6:38
Speaker 1: in a technological revolution like this it's unclear what form
6:42
Speaker 1: the penetration will take right so in the the early
6:45
Speaker 1: parts of, say, the data center buildup, there was a
6:47
Speaker 1: general assumption that it would be LLMs across the board.
6:50
Speaker 1: But now we see people talking about using different types
6:53
Speaker 1: of models, using more of the Chinese type of model,
6:56
Speaker 1: using SLMs, et cetera, small language models, so they can
6:59
Speaker 1: keep things more contained. So already, the way that it's
7:03
Speaker 1: being adopted and penetrating through ordinary companies, it's slightly different
7:07
Speaker 1: to what might have been expected even a few years ago.
7:09
Speaker 1: So it's very hard to predict how this unfolded.
7:13
Speaker 2: AI is very much of an unknown or uncharted territory
7:16
Speaker 2: in a way that we don't know how this is
7:19
Speaker 2: going to end up. But if you do believe that
7:21
Speaker 2: AI is very powerful, that it's going to bring some
7:24
Speaker 2: positive to the human population across the board, I mean,
7:27
Speaker 2: obviously there's going to be some negative as well. Then
7:30
Speaker 2: what you really want to do is to invest into
7:33
Speaker 2: companies that are receiving the spending or investment, which we
7:37
Speaker 2: call them the picks and shuffle companies. But then when
7:40
Speaker 2: you are thinking about the company that investing into AI
7:43
Speaker 2: to make an AI model, to come up with different
7:46
Speaker 2: use cases and all that, that's going to be a
7:49
Speaker 2: trial and error process in a way that a lot
7:51
Speaker 2: of them are not going to survive, but some of
7:52
Speaker 2: them are going to make it to be very successful.
7:55
Speaker 2: So it doesn't mean that you can't pick them, but
7:58
Speaker 2: then I think this is more of a venture capitalist market,
8:01
Speaker 2: isn't it? You invest into probably 100 companies and then
8:04
Speaker 2: some of them worked out, most of them doesn't, but
8:06
Speaker 2: then you still make a lot of money. But I
8:07
Speaker 2: think from our perspective that we invest into a publicly
8:11
Speaker 2: listed company for a high conviction, only about 30 companies,
8:14
Speaker 2: then you probably need a bit more certainty in terms
8:17
Speaker 2: of what the direction of travel is going to be.
8:21
Speaker 1: And one of the big conversations at the moment when
8:23
Speaker 1: looking at the big AI companies are listed ones, the hyperscalers,
8:26
Speaker 1: et cetera, is you look at the valuations and you're like, well,
8:28
Speaker 1: those aren't bubble valuations, but is there maybe a bubble
8:32
Speaker 1: in earnings expectations? So, you know, the PE looks relatively low,
8:37
Speaker 1: but well, not low, but you know, not bubble levels.
8:40
Speaker 1: But is that because the E number is grotesquely over-optimistic?
8:45
Speaker 1: As you say, we can't really predict the future with
8:46
Speaker 1: any of this stuff. So is this more of an
8:48
Speaker 1: earnings bubble?
8:49
Speaker 2: Yeah, I think you also need to differentiate amounts of
8:53
Speaker 2: a few probably major companies, because if you're referring to
8:56
Speaker 2: the likes of SpaceX, OpenAI or Anthropic, I would definitely
9:01
Speaker 2: say that they are in a valuation bubble, right? Two
9:03
Speaker 2: of them yet to be listed in the stock market.
9:05
Speaker 2: They're not profitable at this moment in time. Their revenue
9:08
Speaker 2: is growing very quickly. The projection for them to be
9:10
Speaker 2: profitable would be in a few years' time. And obviously,
9:14
Speaker 2: you're paying up for that. But then when you look
9:15
Speaker 2: at the other side of the fence, let's say, talk
9:17
Speaker 2: about NVIDIA, because obviously, it's the biggest company in the world.
9:20
Speaker 2: NVIDIA currently is trading at probably 16, 17 times earnings.
9:24
Speaker 2: I mean, it's the biggest company, $ 5 trillion. Then you
9:27
Speaker 2: can say that, okay, NVIDIA is not exposed to valuation bubble,
9:31
Speaker 2: but it's exposed to earnings bubble. Okay, but the challenge
9:34
Speaker 2: with earnings bubble is a lot unclear. So it's less
9:38
Speaker 2: clear about versus valuation bubble because when you think about
9:42
Speaker 2: valuation bubble, it's all about future expectations. It's about, okay,
9:46
Speaker 2: what is my payback period? But when you think about
9:48
Speaker 2: earnings bubble, earnings bubble means that the companies are actually
9:52
Speaker 2: earning a lot of money as we speak. Most companies
9:55
Speaker 2: want to buy alongside Broadcom, because there's only two companies
10:01
Speaker 2: that could probably give them the product they want to
10:03
Speaker 2: build AI infrastructure. And ultimately, you have to ask the question,
10:08
Speaker 2: is more to do with whether AI is going to
10:11
Speaker 2: be more game-changing or not? Whether these companies who have
10:14
Speaker 2: invested money are going to get a return on investment?
10:17
Speaker 2: If they do get a return on investment, then the
10:19
Speaker 2: earnings bubble is actually not a bubble. It's actually right,
10:23
Speaker 2: because like this is the product they're selling that people
10:25
Speaker 2: actually want. And then I think also when you think
10:28
Speaker 2: about the valuation bubble versus earnings bubble is the drivers
10:32
Speaker 2: for that to collapse. It's very different. When you think
10:34
Speaker 2: about earnings bubble, it's a lot more difficult for that
10:36
Speaker 2: to change. You mentioned about AI is more of a
10:39
Speaker 2: global domain, which means that it's not one single player.
10:44
Speaker 2: It's not one single use cases that decide to say,
10:47
Speaker 2: I want to buy more GPUs. It's across multiple companies.
10:50
Speaker 2: They have different priorities, different agendas, different use cases, different
10:54
Speaker 2: time horizon. And so for Nvidia's earnings to collapse tomorrow,
10:59
Speaker 2: just like, let's say, for example, like a valuation bubble,
11:03
Speaker 2: you need to have a coordinated effort among all the
11:07
Speaker 2: AI spenders to say, let's stop spending now, let's build
11:10
Speaker 2: something ourselves, or we don't need Nvidia anymore for that
11:13
Speaker 2: to happen. At the same time, when you look at
11:15
Speaker 2: some of the commitments the companies, some of these companies,
11:18
Speaker 2: customers has made, it's now been very much extended, which
11:21
Speaker 2: means that you can't actually pull out your orders so quickly.
11:24
Speaker 2: So hence the visibility for some of the picks and
11:27
Speaker 2: shuffle company that we like, you can easily see that
11:30
Speaker 2: they would continue to earn super normal profits or earnings
11:34
Speaker 2: for the next 18 months at least. Of course, you
11:36
Speaker 2: keep tracking in terms of what's going to happen from
11:39
Speaker 2: 2028 or 2029 onwards from now, but it's not going
11:42
Speaker 2: to change. Nothing major is going to change in the
11:44
Speaker 2: next 18 months.
11:47
Speaker 1: Okay, so there's no valuation bubble and.
11:49
Speaker 2: There's no earnings bubble. I think for the company that
11:51
Speaker 2: we invest in too, there's more of our earnings. They're
11:54
Speaker 2: over earnings, for sure. But over earnings doesn't mean the
11:57
Speaker 2: earnings will go back to zero when the clock stops, right?
12:02
Speaker 2: It's the same thing when you remember the COVID period.
12:06
Speaker 2: Think about all the digital winners like your Amazon, your Alphabets, Meta, Microsoft,
12:13
Speaker 2: and all that at a time. Everyone was talking about, oh, yeah, okay,
12:16
Speaker 2: once we come out from COVID, we don't need any
12:18
Speaker 2: of this anymore. I mean, that's not true. Now we
12:21
Speaker 2: live or operate in a very hybrid world. And when
12:23
Speaker 2: you look at the big tech companies, the Mach 7,
12:26
Speaker 2: people love to talk about Mach 7, when you total
12:29
Speaker 2: the revenue of these seven companies, is about $ 3 trillion.
12:35
Speaker 2: It's similar size to the UK economy, seven companies combined revenue.
12:39
Speaker 2: I mean, that is a crazy number, right?
12:41
Speaker 1: That might say more about the UK economy than it
12:43
Speaker 1: does about the company.
12:44
Speaker 2: So what I'm trying to say here is like, obviously
12:46
Speaker 2: 10 years ago, they weren't earning $ 3 trillion, but today
12:50
Speaker 2: they are. But the reason, very simple reason being, is
12:53
Speaker 2: because we are now very much entrenched in the ecosystem, right?
12:57
Speaker 2: Think about Apple, Alphabet, Microsoft, and all that. And we
13:01
Speaker 2: can't live without them. And then we end up continuing
13:04
Speaker 2: that journey. And then the question is like, is that
13:07
Speaker 2: the same thing for AI or not the same thing
13:10
Speaker 2: for AI?
13:20
Speaker 1: Well, I think let's pick up there on this idea
13:23
Speaker 1: that tech is an ecosystem and AI in particular is
13:27
Speaker 1: an ecosystem. It's not just a software thing. AI is physical, right?
13:32
Speaker 1: Is there a way to say I would like to
13:35
Speaker 1: get away from the great AI boom? So if something
13:41
Speaker 1: goes wrong, I'm not in it. That kind of isn't anymore, right?
13:44
Speaker 2: I think there's still certain companies or sectors or asset
13:48
Speaker 2: classes that you can explore. I think what comes to
13:52
Speaker 2: mind is like a holding in the fund is European defence.
13:56
Speaker 2: We really like European defence just because this is also
13:59
Speaker 2: a new normal that European Union have concluded that actually
14:03
Speaker 2: we do need to spend some money to defend ourselves
14:06
Speaker 2: on the back of Trump's policies. And this is going
14:08
Speaker 2: to be a much more prolonged kind of investment cycle
14:12
Speaker 2: into European defense. So that is something unrelated to AI.
14:16
Speaker 2: I think if AI does change in terms of the
14:18
Speaker 2: course direction of travel, it shouldn't have any impact on
14:21
Speaker 2: the European defense budget, just because we do need it.
14:24
Speaker 2: And then secondly, when you think about maybe the more
14:26
Speaker 2: true economic environment, like two other areas that we like
14:32
Speaker 2: would be, let's say, oil and gas companies. Like we
14:34
Speaker 2: have a company called Canadian Natural Resources. Of course, I mean,
14:37
Speaker 2: things are changing every single probably minutes or every day.
14:41
Speaker 2: But if you think about what we have experienced since
14:45
Speaker 2: 2022 is what we have gone into, and we have
14:47
Speaker 2: gone into a new regime of geopolitical uncertainties. We started
14:51
Speaker 2: with the Ukraine invasion, followed by some of the Middle
14:54
Speaker 2: Eastern crisis and recently with the Iran situation. And of course,
14:58
Speaker 2: in time, I'm sure we're going to see a lot
15:01
Speaker 2: more uncertainties maybe in the Asia Pacific region as well.
15:04
Speaker 2: So hence, could be one of the things that are
15:07
Speaker 2: less probably tied to AI. Last but not least, we
15:10
Speaker 2: also have initiated a company last year in buying into
15:15
Speaker 2: a gold miner called Newmont. And of course, I mean,
15:19
Speaker 2: gold prices has not done very well recently. I think
15:21
Speaker 2: there's a bit of technical factor. But if you think
15:24
Speaker 2: that maybe AI is not going to be as game changing,
15:28
Speaker 2: I think gold shouldn't really be that impacted as much.
15:32
Speaker 2: So I think there's some areas. But of course, in
15:35
Speaker 2: the meantime, there's a bit of opportunity cost, isn't it?
15:38
Speaker 2: Because if you think about What you're trying to achieve
15:40
Speaker 2: is that are you just trying to make a decent return,
15:43
Speaker 2: let's say maybe up to 10% a year, which I
15:45
Speaker 2: think is a good return for most investors. But from
15:47
Speaker 2: our perspective, being an active manager, which is a very
15:52
Speaker 2: competitive space, that you need to outperform. So when you
15:54
Speaker 2: need to outperform, then you need to earn more money
15:58
Speaker 2: or performance versus your peers, versus the market. And ultimately,
16:02
Speaker 2: if you decided to be very defensive and being a
16:05
Speaker 2: contrarian too early, even that you might be right, then
16:09
Speaker 2: you are basically trading off the opportunity cost that could
16:13
Speaker 2: cost your investors performance. So hence, I think it's a
16:18
Speaker 2: very difficult question to tackle, but it's a question that
16:22
Speaker 2: we have to keep asking ourselves.
16:24
Speaker 1: Yeah, that was the nightmare of active management summed up
16:26
Speaker 1: very succinctly. Listen, let's stick with gold mining and resources, right?
16:32
Speaker 1: Because it takes us right back to what we were
16:34
Speaker 1: saying at the beginning is that people look at you
16:36
Speaker 1: and they go, yes, Stephen runs an AI fund, but
16:38
Speaker 1: you really don't. You run a stock picking fund. So
16:40
Speaker 1: what is it about Newmont mining, which I think may
16:43
Speaker 1: be a listener favorite that you find attractive?
16:46
Speaker 2: Yeah. So before I kick off with that, I just
16:48
Speaker 2: want to quickly introduce this concept called follow the money,
16:52
Speaker 2: which is a part of our investment philosophy. It's very simple.
16:55
Speaker 2: So globally, if you think about how the world operates,
16:58
Speaker 2: how companies operate, there are only three sources of spending,
17:01
Speaker 2: consumer spending, enterprise spending, and government spending. Where you want
17:05
Speaker 2: to be is to invest into companies or sectors that
17:08
Speaker 2: are going to capture a lot of incremental spending. And
17:13
Speaker 2: the world is a zero-sum game. So if your companies
17:15
Speaker 2: are capturing incremental spending away from consumers, enterprise, or government,
17:20
Speaker 2: someone is going to lose out on the other side
17:22
Speaker 2: of the fence. So then you think about what's going
17:25
Speaker 2: on within the gold mining space. You think about the
17:29
Speaker 2: global deficit within the different governments, let's say Japan, US, UK,
17:35
Speaker 2: or Europe. Basically, fiat currency is, I mean, I wouldn't
17:41
Speaker 2: call it a scam, but then in a way that
17:43
Speaker 2: basically we're trying to borrow into the future that no
17:46
Speaker 2: one is able to pay it all back. And then
17:48
Speaker 2: if you think about what is the alternative asset class
17:51
Speaker 2: that may protect you from some of this quantitative easing
17:55
Speaker 2: in a way for the governments to borrow. There's only
17:58
Speaker 2: two asset classes. One is crypto. The other one is gold.
18:02
Speaker 2: Gold is going to be a good hedge against the
18:04
Speaker 2: depreciation of fiat currency in general.
18:07
Speaker 1: Because we know that gold has a 3,000-year history of
18:10
Speaker 1: maintaining its real value in the face of the collapse
18:13
Speaker 1: of every single fiat currency there's ever been. So you
18:15
Speaker 1: want to hold gold, but why do you want to
18:17
Speaker 1: hold Newmont?
18:18
Speaker 2: So, obviously, we cannot invest into gold. I'm sure some
18:22
Speaker 2: of our audience can.
18:23
Speaker 1: I mean, you could. You could buy a bullion-backed gold ETF, right,
18:28
Speaker 1: and hold it in the fund, but you wouldn't do that.
18:30
Speaker 2: Yeah, we can do that too, but we decided to
18:33
Speaker 2: pick a company instead. So, Newmont is actually quite a
18:36
Speaker 2: nice company. It is the biggest gold miner in the world,
18:39
Speaker 2: accounting for about 5% of the world production. Majority of
18:43
Speaker 2: their assets are based in North America. So I would
18:46
Speaker 2: say it's pretty safe jurisdiction. And then the beauty investing
18:50
Speaker 2: into oil mining obviously comes with risk as well because
18:52
Speaker 2: it's a company, is that when you do some sort
18:55
Speaker 2: of valuation on the gold reserve, taking into account the
18:58
Speaker 2: mining costs, overtime, etc., inflation, into account, etc., that they
19:03
Speaker 2: are trading at a discount to the current gold prices.
19:06
Speaker 2: So what it means is if you invest into a
19:08
Speaker 2: gold ETF, at the moment, let's say the gold price
19:11
Speaker 2: is about $ 4, 000, if gold prices doesn't change in the
19:15
Speaker 2: next two years or three years, then you're not going
19:17
Speaker 2: to make any money. But for a gold miner such
19:20
Speaker 2: as Newmont, because they have a lot of reserves, again,
19:23
Speaker 2: they are very efficient to mine the gold out, assuming
19:26
Speaker 2: that the gold prices doesn't change for the next few
19:28
Speaker 2: years at $ 4, 000, they will still make a lot of money.
19:31
Speaker 2: So then, hence, you get a bit of, I would say,
19:34
Speaker 2: margin of safety in a way that we don't actually
19:37
Speaker 2: need gold prices to go up for Newmont to make money.
19:41
Speaker 1: And if gold prices go down, not so good.
19:44
Speaker 2: Then we're wrong.
19:45
Speaker 1: Then you're wrong.
19:46
Speaker 2: Of course.
19:46
Speaker 1: Again, crypto, you say that you could invest in either
19:49
Speaker 1: gold or crypto as a hedge against global devaluation. There's
19:53
Speaker 1: nothing that you could invest in to leverage your interest
19:57
Speaker 1: in crypto inside this fund, right?
19:59
Speaker 2: No, no, we are not interested in crypto. I was
20:01
Speaker 2: just giving an example that people can consider crypto as
20:06
Speaker 2: an alternative to fiat currency.
20:07
Speaker 1: I was interested in that because I Yeah, I didn't
20:10
Speaker 1: have you down as a crypto guy.
20:12
Speaker 2: No, no, no. I think that's something that we would
20:15
Speaker 2: try to stay away from just because it's something a
20:18
Speaker 2: lot more difficult to understand in terms of the ecosystem. Okay.
20:22
Speaker 1: And have you been looking at anything else in the
20:24
Speaker 1: commodity space?
20:24
Speaker 2: So recently, we started to, I would say quite early
20:28
Speaker 2: days that we started to build a small position into
20:31
Speaker 2: a refinery called Marathon. And basically, if you think you
20:37
Speaker 2: follow the landscape within the oil and gas space, so
20:40
Speaker 2: obviously you have the oil producers, but you also have refiners.
20:44
Speaker 2: If you track some of the spread between the oil
20:47
Speaker 2: that's coming called crack spread and then the output that's
20:51
Speaker 2: coming out, which is gasoline and other fuels, that has
20:55
Speaker 2: been at a very elevated level. And of course, then
20:58
Speaker 2: when you look into the dynamics behind it, it's because
21:01
Speaker 2: that a lot of the refinery have been shut down
21:05
Speaker 2: on the back of the geoparticle uncertainty. Some of them
21:07
Speaker 2: has been damaged. And at the same time, when you
21:10
Speaker 2: look at, okay, what happened if we do have a
21:13
Speaker 2: deal tomorrow, for example, with Iran and everything go back
21:16
Speaker 2: to normal. I mean, does all this refinery capacity go
21:19
Speaker 2: back live like tomorrow, 100%? It doesn't because it's actually
21:25
Speaker 2: a very expensive exercise to rebuild some of this infrastructure.
21:28
Speaker 2: So hence, unfortunately, I would have to say that consumer
21:32
Speaker 2: are the one who's going to suffer for longer than
21:34
Speaker 2: you think. So even though maybe oil price might come down,
21:37
Speaker 2: the end product, which is the product we end up
21:40
Speaker 2: paying for, could remain at a very elevated level just
21:43
Speaker 2: because of supply and demand imbalance within the refinery are
21:47
Speaker 2: very now constrained. So it's a bit like the follow
21:50
Speaker 2: the money approach again, isn't it? You think about why
21:53
Speaker 2: the AI infrastructure, the picks and shuffle company are earning
21:57
Speaker 2: so much money is because of supply and constraint are imbalanced.
22:02
Speaker 2: And hence, you just need to pay them a lot
22:05
Speaker 2: more money. So hence, the refiners are making a very
22:08
Speaker 2: big margin in terms of the input cost versus the
22:12
Speaker 2: price they end up selling the product to us.
22:14
Speaker 1: And the other area that is interesting, of course, is medical,
22:17
Speaker 1: medical and healthcare, which I guess you could say is
22:20
Speaker 1: AI related in lots of ways. And there's huge expectations
22:23
Speaker 1: that AI can bring us to lots of wonderful new
22:26
Speaker 1: places in healthcare, but that's a little way away.
22:28
Speaker 2: Yeah, so we have two companies, but I will tackle
22:30
Speaker 2: the one that's related to AI first. So there are
22:33
Speaker 2: two segments to that exposure. One area is biologics equipment companies.
22:38
Speaker 2: So we have Danaher and Repogen. And if you follow
22:42
Speaker 2: what has been announced recently by Anthropic, that they are
22:44
Speaker 2: now having this Anthropic, I think... AI2 that's going to
22:50
Speaker 2: perform preclinical trial, which is, I would say, that's very promising.
22:54
Speaker 2: And this is something that we have been talking about
22:56
Speaker 2: for a long time. It's about personalized medicine. Just think
22:59
Speaker 2: about biologics, think about COVID vaccines and all that. Everyone
23:02
Speaker 2: is different. You want to have personalized medicine. Today, we
23:05
Speaker 2: do have personalized medicine, but it's very expensive. I think
23:08
Speaker 2: the treatment for some late-stage cancer is going to cost
23:11
Speaker 2: you $ 50, 000. to $ 100, 000. That's also a note most people
23:15
Speaker 2: can't afford it. And then in time, if we do
23:17
Speaker 2: have AI to accelerate drug discovery, then of course, we're
23:20
Speaker 2: going to have more biologics. And when you look at
23:23
Speaker 2: the drug under research that's not yet approved, the development cycle,
23:29
Speaker 2: two majority of the drugs under research are actually using
23:32
Speaker 2: biologics rather than chemical drugs. kind of base approach. So
23:36
Speaker 2: it's actually very promising. The only thing I would say
23:39
Speaker 2: that it's going to take much longer because obviously you
23:41
Speaker 2: have to take regulation in place. It's probably going to
23:45
Speaker 2: be 10 years away rather than two years away. It's
23:47
Speaker 2: unlike other AI use cases that we already can get
23:50
Speaker 2: our hands on already. I think medicines is something that
23:53
Speaker 2: we are going to take a more cautious approach. The
23:55
Speaker 2: other area that we have been quite excited about is
23:58
Speaker 2: the GLP-1 area. So this is not related to AI.
24:02
Speaker 2: It's Eli Lilly. And I'm sure everyone is familiar with
24:06
Speaker 2: that space by now that we i mean basically is
24:09
Speaker 2: very game-changing in the way that we is making the
24:12
Speaker 2: world i would when i say making a world better
24:15
Speaker 2: place uh that that we are consuming a lot less
24:18
Speaker 2: uh i would say unhealthy um uh i would say
24:21
Speaker 2: products and and and and obviously we could live even
24:25
Speaker 2: longer in in a way that um with some of
24:27
Speaker 2: these trucks not just like losing your weights but then
24:29
Speaker 2: i think there's other other other things that developing as
24:31
Speaker 2: well so you like lady i would say also if
24:34
Speaker 2: you think about hedging against AI. I think that is
24:37
Speaker 2: totally not related to AI, that it could continue to
24:39
Speaker 2: think about the world population. 20% of the world population
24:43
Speaker 2: potentially could be part of that customer cohort, which is
24:46
Speaker 2: very big. It's like a billion people.
24:48
Speaker 1: What other portfolio haven't we talked about that is interesting
24:50
Speaker 1: and we should talk about?
24:54
Speaker 2: So I would say that within the position of the
24:57
Speaker 2: portfolio that we very much still centric on the AI industry,
25:02
Speaker 2: infrastructure companies, the pitch and shuffle, just because we do
25:06
Speaker 2: think if they are already making the money today, rather
25:09
Speaker 2: than hoping them to make the money of tomorrow, then
25:12
Speaker 2: I think you get paid. I think you have a
25:15
Speaker 2: bit of margin of safety from an investor perspective. And
25:19
Speaker 2: then we have talked about the different other ideas, trying
25:22
Speaker 2: to tackle the geoeconomic environments. uncertainties like your own gas company,
25:28
Speaker 2: your maybe gold miner, European defense, and then obviously we
25:32
Speaker 2: have some healthcare companies. I think that probably sums up all.
25:36
Speaker 2: I think one thing that I would want to just highlight,
25:38
Speaker 2: just going back to your point that, yeah, we are
25:40
Speaker 2: not an AI fund, not a technology fund, not an
25:43
Speaker 2: American fund. We run a global mandate. But I think
25:45
Speaker 2: how we ended up coming to this conclusion about what
25:50
Speaker 2: We are very privileged because we run a global mandate
25:52
Speaker 2: that we are very agnostic about where the opportunities are,
25:56
Speaker 2: which sectors do we want to own or not own,
25:58
Speaker 2: that we are not constrained that if we are running
26:00
Speaker 2: a tech fund, then of course we have to pitch
26:02
Speaker 2: about AI. Otherwise, then what are you going to own
26:04
Speaker 2: within the tech space? And hence, I think the conclusion
26:08
Speaker 2: that we come to are probably the best ideas that
26:11
Speaker 2: we have concluded globally that this is the best ideas
26:14
Speaker 2: today for looking forward.
26:18
Speaker 1: Yeah, so when people look at the fund now and
26:20
Speaker 1: have over the last couple of years looked and gone, well,
26:23
Speaker 1: that's an AI fund. You've ended up having a lot
26:26
Speaker 1: of AI-related companies by default because they've been the companies
26:29
Speaker 1: that look most interesting to you given the parameters in
26:32
Speaker 1: which you stock pick. When you look out now, say
26:46
Speaker 1: where the fund will be in three years? What do
26:48
Speaker 1: you think people will be saying? Will they be saying,
26:50
Speaker 1: oh yeah, that's a gold mining fund? Oh yeah, no,
26:53
Speaker 1: that's Stephen here. He runs that natural resources fund. He's
26:57
Speaker 1: an oil guy. What do you think people will be
27:00
Speaker 1: saying about your fund in three years?
27:03
Speaker 2: I don't know about that yet. I think it's even
27:05
Speaker 2: though that we, I would say that it's unlikely that
27:07
Speaker 2: we would We would be an oil guy. I think
27:12
Speaker 2: it's just not a natural part of what we do.
27:16
Speaker 2: But what we do like is we do like the
27:17
Speaker 2: game changes. We do like companies that are actually shaping
27:21
Speaker 2: the journey for the human population. So then if you
27:26
Speaker 2: think about maybe not three years, maybe more than three years,
27:29
Speaker 2: there's a lot of exciting thing, which today I would
27:31
Speaker 2: say is too early to call it today. But it's
27:33
Speaker 2: something that you want to watch very closely for. You
27:36
Speaker 2: think about space ecosystem. I think it's too early today,
27:40
Speaker 2: but space could be a very exciting part. You think
27:43
Speaker 2: about robotics, human rights, which is where it's going to
27:47
Speaker 2: be even bigger than your AI today, right? You think
27:50
Speaker 2: about what human rights could do in time. It's all
27:53
Speaker 2: powered by AI.
27:54
Speaker 1: Physical AI, right?
27:56
Speaker 2: Physical AI, exactly. Human rights. You think about autonomous vehicle.
27:59
Speaker 2: I mean, obviously it has been very slow to get adopted,
28:02
Speaker 2: but then that is also going to be powered by
28:04
Speaker 2: AI in a way, but you can invest into companies
28:07
Speaker 2: that might be providing the nitty gritty behind this. And
28:11
Speaker 2: then you think about maybe certain drugs, certain healthcare opportunities, etc.
28:18
Speaker 2: I think there's a lot of things. Basically, what I'm
28:21
Speaker 2: trying to say is no one knows where the world
28:24
Speaker 2: is going to take us. But if this is the
28:26
Speaker 2: world that we're going to end up with, we want
28:28
Speaker 2: to be there. So if in time... space is where
28:32
Speaker 2: everyone is going in a way that we are actually
28:34
Speaker 2: going to, I don't know, go to Mars or go
28:37
Speaker 2: to the moon and have another ecosystem there. I think
28:39
Speaker 2: you should expect Blue Whale Growth Fund to be there
28:42
Speaker 2: in a way that, okay, we are now a space fund.
28:45
Speaker 2: We are now like a, I don't know, personal medicines fund.
28:48
Speaker 2: I think this is like very exciting and this is
28:50
Speaker 2: what excites us.
28:51
Speaker 1: I think.
28:53
Speaker 2: Natural resources or maybe gold or all that is good
28:57
Speaker 2: investment opportunity, but that's not really new.
29:00
Speaker 1: What about if there's mining on the moon and you
29:03
Speaker 1: could suddenly be a.
29:04
Speaker 2: Space mining fund? We would be interested.
29:08
Speaker 1: What are the bits about space that really interest you?
29:11
Speaker 2: I would say that space is very complicated in a way.
29:15
Speaker 2: I personally think, obviously, early days, I mean, SpaceX just launched,
29:18
Speaker 2: so maybe Elon Musk definitely would be more, probably will
29:22
Speaker 2: have more context in terms of where we're heading from here.
29:25
Speaker 2: But from my take on this, it's like he has suggested, oh, we,
29:28
Speaker 2: in time, we're going to have AI data center up
29:30
Speaker 2: in space. So then, hence, you don't need a liquid
29:34
Speaker 2: cooling system anymore. You don't need fiber optics anymore, which
29:37
Speaker 2: we have some of this company, but then you need
29:40
Speaker 2: to think about, okay, when you're really going to have
29:42
Speaker 2: AI data center up there, you need to have a
29:44
Speaker 2: defense mechanism. You're going to probably shut up hundreds of
29:51
Speaker 2: billions of AI data centers. What if you're going to
29:55
Speaker 2: take them out? I mean, what happened? Who is going
29:57
Speaker 2: to do that? So that's going to be a mega ecosystem.
30:00
Speaker 2: I mean, we talk about a lot of money. At
30:02
Speaker 2: this moment in time, I don't think we are ready
30:04
Speaker 2: for that yet. Of course, when you look at some
30:07
Speaker 2: of the Europe defense company, they already have a bit
30:10
Speaker 2: of space, a defense technology, but still quite, I would say,
30:14
Speaker 2: when I say primitive means that it's not as big
30:17
Speaker 2: as the land defense kind of technology. So in time,
30:22
Speaker 2: it's actually a very big ecosystem. That's why when you
30:25
Speaker 2: follow SpaceX, I mean, they're talking about, I don't know,
30:28
Speaker 2: $ 100 trillion opportunity or something, $ 100 trillion, which is the
30:32
Speaker 2: world GDP.
30:33
Speaker 1: Yeah.
30:33
Speaker 2: Don't even understand how that number comes about, but that
30:36
Speaker 2: is what we're talking about.
30:38
Speaker 1: What about the idea that we don't have to worry
30:39
Speaker 1: about energy anymore because it won't be long before we
30:42
Speaker 1: can beam solar energy direct from space to.
30:45
Speaker 2: It could happen. I think what is really exciting about
30:50
Speaker 2: AI today is that I think it makes a lot
30:52
Speaker 2: of things possible. Just when I say makes a lot
30:55
Speaker 2: of things possible, that it can make things possible within
30:58
Speaker 2: a shorter space of time.
30:59
Speaker 1: Yeah, we can do the thinking a lot faster.
31:01
Speaker 2: A lot faster. And then obviously a lot of testings,
31:03
Speaker 2: a lot more trials and in the background without spending
31:07
Speaker 2: money to do it physically. And then the other thing
31:09
Speaker 2: we haven't talked about is quantum computing.
31:11
Speaker 1: Yes, we did a podcast on that the other day.
31:13
Speaker 1: So exciting.
31:15
Speaker 2: In time, if quantum computing comes, we want to be
31:17
Speaker 2: there to evade. But then the one thing I would
31:19
Speaker 2: have to say that as a bottom-up stock pickers or
31:22
Speaker 2: being very fundamental-based, to be profitable. So we rarely invest
31:31
Speaker 2: in a company that are in a very unknown territory that, oh,
31:34
Speaker 2: you might be able to make it to become a
31:36
Speaker 2: winner or maybe you end up becoming a loser. We
31:39
Speaker 2: try to avoid those opportunities. So we need to see
31:43
Speaker 2: that this company would end up being the Nvidia or
31:46
Speaker 2: the Alphabet or the Amazon of tomorrow. But at this
31:50
Speaker 2: moment in time, I think it's too early to call.
31:52
Speaker 1: And how soon do you think that quantum will start
31:55
Speaker 1: having effect? When's Q Day for you? We ask when
31:57
Speaker 1: Q Day is quite a lot. We get a lot
31:59
Speaker 1: of different answers.
32:00
Speaker 2: This is actually quite a relevant question to ask. I mean, obviously,
32:03
Speaker 2: we don't have an answer yet. a proper answer because
32:08
Speaker 2: the biggest risk that we see within the AI exposure
32:13
Speaker 2: that we have got is not about whether AI is real.
32:17
Speaker 2: It's not about the rate of innovation. It's not about
32:20
Speaker 2: competitions from China and all that. I mean, we all think,
32:23
Speaker 2: I mean, sorry, we concluded that it's all manageable. The
32:25
Speaker 2: biggest risk, some of the biggest risk is actually coming
32:28
Speaker 2: from within. So then if you think about the biggest
32:30
Speaker 2: risk to NVIDIA, it's not Broadcom. It's not China. It's
32:34
Speaker 2: actually quantum computing. Because if we do have quantum computing,
32:38
Speaker 2: you don't need GPUs anymore. Okay, then according to, I think,
32:42
Speaker 2: Alphabet and maybe IBM, I think we might have something
32:47
Speaker 2: maybe in a few years' time, by 2030. Okay, how
32:50
Speaker 2: game-changing that is, no one knows at this moment in time, right?
32:54
Speaker 2: But it's something that you want to follow quite closely.
32:57
Speaker 2: But unfortunately, when you look at some of the more
33:00
Speaker 2: independent quantum computing companies that are listed in the stock
33:03
Speaker 2: market today, I think they are probably a bit more
33:04
Speaker 2: speculative in a way that I think most of them
33:07
Speaker 2: are not going to make it. I'm sure some of
33:09
Speaker 2: them are going to be very successful, but then it's
33:12
Speaker 2: probably more of hit and miss. So I would say
33:14
Speaker 2: probably a few years from today, but it's something you
33:16
Speaker 2: need to watch quite closely because it's.
33:18
Speaker 1: A huge risk to a lot of the other things
33:20
Speaker 1: that we've talked about.
33:20
Speaker 2: Correct. But then what is really positive on the back
33:23
Speaker 2: of that is going to even accelerate AI innovation if
33:28
Speaker 2: we do have that.
33:29
Speaker 1: Yeah, because AI plus quantum is an extraordinary speed of thinking.
33:33
Speaker 2: Correct.
33:34
Speaker 1: Okay, so that's all positive. What are you worried about, Stephen?
33:39
Speaker 1: What keeps you up at night? I know there's something.
33:41
Speaker 1: What could make everything go horribly wrong?
33:44
Speaker 2: I do have something. So if I think about the
33:49
Speaker 2: companies within the portfolio, which is about 30 companies, that
33:52
Speaker 2: we are not so much worried about them. When I
33:54
Speaker 2: say not worried, it doesn't mean that we don't worry
33:56
Speaker 2: about them. We do, but we can actually analyze them.
33:58
Speaker 2: We can actually follow them. We can track the quarterly earnings.
34:01
Speaker 2: We can track the news flow. We can see the
34:03
Speaker 2: products development and all that. And then we can make
34:04
Speaker 2: a decision, right? So it's something that we can actually
34:07
Speaker 2: understand and we can monitor, we can reassess. But there
34:10
Speaker 2: are certain things that we can't actually understand and reassess.
34:13
Speaker 2: I would say that it's the geopolitical environment. And the
34:16
Speaker 2: one thing that I have been saying for the last
34:18
Speaker 2: few years is that probably the biggest risk ever from
34:23
Speaker 2: my perspective is what's going to happen in Asia-Pacific. in time,
34:26
Speaker 2: that is something that you can't model, like what is
34:30
Speaker 2: going to happen to Taiwan. Taiwan is a very important
34:33
Speaker 2: part of the ecosystem. Think about if Taiwan is no
34:35
Speaker 2: longer autonomous, the world come to a halt, right? This
34:39
Speaker 2: is not just about AI. It is related to every
34:42
Speaker 2: chips that we ever have in any equipment or consumer gadgets,
34:48
Speaker 2: et cetera. So hence, like, is a very, I would say,
34:52
Speaker 2: material moment But of course, then you think about, okay,
34:56
Speaker 2: is that going to happen? How is it going to happen?
34:58
Speaker 2: When is it going to happen? There's nothing that you
35:01
Speaker 2: could put your hands on. You can't predict it. You
35:04
Speaker 2: can't predict it, right? But if it happens tomorrow, for
35:07
Speaker 2: whatever reason, which unlikely it would be the case, it's
35:10
Speaker 2: going to be chaos. And there's nothing that you can
35:14
Speaker 2: really protect yourself apart from maybe a handful of companies,
35:17
Speaker 2: maybe gold, maybe defense companies, I think most things would
35:23
Speaker 2: suffer quite badly. So hence, I think this is the
35:27
Speaker 2: thing that you worry about, but it's not like there's
35:30
Speaker 2: something that you can actually do about this because otherwise
35:32
Speaker 2: then you say, okay, why don't we just have everything
35:35
Speaker 2: in cash for the next 10 years until it happened?
35:37
Speaker 2: It might never happen as well.
35:39
Speaker 1: Yeah, exactly. But at least you have Newmont.
35:42
Speaker 2: Yeah, at least we have.
35:43
Speaker 1: You've got your insurance.
35:44
Speaker 2: So we could become maybe a gold mining fund.
35:48
Speaker 1: By default, by mistake.
35:50
Speaker 2: Yeah.
35:51
Speaker 1: And in five years, we'll be saying, yes, Stephen, he's
35:53
Speaker 1: that gold guy. That would be unexpected. Stephen, thank you
35:57
Speaker 1: so much for coming on today. Hugely appreciate it.
35:59
Speaker 2: Been really interesting. Thank you, Maren.