Latest / The Money Puzzle / How to invest in the data centre boom

Transcript

How to invest in the data centre boom · The Money Puzzle

• 0:10
  Speaker 1: Hello and welcome to the Australian's Money Puzzle podcast. I'm
• 0:13
  Speaker 1: James Kirby. Welcome aboard. Now, the $ 150 billion minimum rollout
• 0:20
  Speaker 1: of data centres across Australia to service the adoption of AI,
• 0:25
  Speaker 1: it's the biggest thing we have seen. It's the biggest
• 0:28
  Speaker 1: economic boom inside the wider economy since the comparable is
• 0:34
  Speaker 1: the LNG boom, which is at least 15 years ago.
• 0:38
  Speaker 1: How do you, the investor, get a slice of the
• 0:41
  Speaker 1: action here. How do you play this boom? And how
• 0:44
  Speaker 1: is it going to map out in the years ahead?
• 0:47
  Speaker 1: And what we have, first of all, we have that
• 0:49
  Speaker 1: uplift where we have a lot of investment and we
• 0:51
  Speaker 1: have winners and losers. Later on, if history is to
• 0:55
  Speaker 1: be relied upon, we'll have a takeover boom that comes
• 0:58
  Speaker 1: as everybody mops it up and basically you get a
• 1:01
  Speaker 1: number of key players at the end of any boom.
• 1:03
  Speaker 1: That's how it happens. We need, I think, a jungle guide.
• 1:06
  Speaker 1: And my guest today is Emmanuel Dat. He is from
• 1:11
  Speaker 1: the DAT Capital Group. Emmanuel has an absolute return fund
• 1:16
  Speaker 1: and a small companies fund. The absolute return fund is
• 1:19
  Speaker 1: probably the best known fund. About $ 200 million under management.
• 1:22
  Speaker 1: He's had an 11% TA return since inception, which is
• 1:26
  Speaker 1: pretty good. He promises that he will bring you in 5%
• 1:31
  Speaker 1: above the RBA cash rate in his funds. That's his,
• 1:34
  Speaker 1: I promise might be too strong a word, Emmanuel, I
• 1:36
  Speaker 1: should have said, but that would be the expectation you
• 1:38
  Speaker 1: put forward. Would that be fair to say?
• 1:40
  Speaker 2: Yes, yes, that's our benchmark, and we seek to achieve
• 1:44
  Speaker 2: double-digit net returns to our investors each financial year.
• 1:48
  Speaker 1: That's our objective. I came to you on this. I mean,
• 1:53
  Speaker 1: Emmanuel and I, folks, we often talk in the course
• 1:56
  Speaker 1: of my work, but we haven't had him on the
• 1:58
  Speaker 1: show before. But I knew that he had a particular
• 2:00
  Speaker 1: interest in the data center theme and is as such
• 2:03
  Speaker 1: all over it and would have been all over it
• 2:05
  Speaker 1: long before. The rest of us started to pay attention
• 2:08
  Speaker 1: to it. But let's just kind of map it out,
• 2:10
  Speaker 1: first of all, for investors, Emmanuel. Okay, we take it
• 2:15
  Speaker 1: as red folks that if you want to, you can
• 2:17
  Speaker 1: just go and buy, you know, data center ETF global
• 2:19
  Speaker 1: or whatever and see how you go. And there's nothing
• 2:22
  Speaker 1: wrong with that. But assuming you're a little bit more
• 2:24
  Speaker 1: ambitious than that and you want to see what's happening
• 2:26
  Speaker 1: in our world here, tell us a little bit about it, Emmanuel.
• 2:31
  Speaker 1: $ 150 billion rollout of data centers, at the very least.
• 2:36
  Speaker 1: Tell us, describe, if you would, what's going on, this boom,
• 2:41
  Speaker 1: and in what way investors so far have taken a
• 2:44
  Speaker 1: piece of action. Yeah, sure, James.
• 2:47
  Speaker 2: Thanks for having us. I think that the number one
• 2:50
  Speaker 2: aspect to be understood is what's driving this actual boom.
• 2:54
  Speaker 2: And I think for the general listener out there, it's
• 2:57
  Speaker 2: clearly the mainstream adoption of AI, which has certainly come
• 3:03
  Speaker 2: and progressed in leaps and bounds over the past couple
• 3:06
  Speaker 2: of years. And that sort of mainstream adoption is really
• 3:09
  Speaker 2: driving the demand for compute, which is basically what runs
• 3:13
  Speaker 2: these models.
• 3:15
  Speaker 1: So it's like the engine, it's the engine room of
• 3:16
  Speaker 1: the AI. I don't, to simplify it.
• 3:18
  Speaker 2: Yes, that's right. That's right. And so that obviously requires
• 3:22
  Speaker 2: data centers, which are effectively, you know, delivering this computing
• 3:28
  Speaker 2: power at scale. And as you mentioned, James, that this is,
• 3:33
  Speaker 2: an extraordinary amount of capital that's being put forward. And
• 3:39
  Speaker 2: it's probably one of the largest, you know, coordinated CapEx
• 3:42
  Speaker 2: cycles that we've seen in recent times. Yes.
• 3:47
  Speaker 1: I was looking at the ABS report and they have
• 3:49
  Speaker 1: these columns of where they say things are being built.
• 3:52
  Speaker 1: They had commercial buildings, you know, and they had the 3.5%
• 3:55
  Speaker 1: or whatever improvement. They had this column which says, other, other.
• 4:02
  Speaker 1: And the other category was 280%. And that was where
• 4:05
  Speaker 1: the data centers were. So we've already had quite a
• 4:11
  Speaker 1: construction lift in this area. And there's much, much more
• 4:14
  Speaker 1: to come. And it would compete, won't it, with other areas.
• 4:16
  Speaker 1: It literally competes with home building, for instance. So it's,
• 4:20
  Speaker 1: to some extent, politically a hot area. But let's just
• 4:24
  Speaker 1: look at it as an investor. Let's just take it
• 4:27
  Speaker 1: that people are familiar with what it is. Let's take
• 4:30
  Speaker 1: it that they know that, There are issues here that
• 4:34
  Speaker 1: are political. And one of them, the first ones that
• 4:37
  Speaker 1: emerged was that the government in its enthusiasm to foster
• 4:40
  Speaker 1: economic growth, needless to say, have supported in principle data
• 4:46
  Speaker 1: center building because it's going to be a construction boom,
• 4:49
  Speaker 1: if nothing else. However, on close inspection, we were just
• 4:54
  Speaker 1: talking about how it's a very particular boom, right? Tell
• 5:00
  Speaker 1: us a little bit more of what it's really like
• 5:01
  Speaker 1: on the ground. Yeah, absolutely.
• 5:04
  Speaker 2: Look, I think that ultimately with all these plethora of
• 5:08
  Speaker 2: data center projects being constructed, it's really driving demand in
• 5:14
  Speaker 2: areas such as the contractor space that you pointed out
• 5:17
  Speaker 2: that apps would work on home building or government infrastructure
• 5:21
  Speaker 2: or even private infrastructure, for instance. So all of a
• 5:24
  Speaker 2: sudden there's all this demand for services and Particularly in
• 5:28
  Speaker 2: the electrical contracting businesses, those data centers obviously are very
• 5:34
  Speaker 2: heavily reliant on electrical infrastructure. So that's where we're probably
• 5:39
  Speaker 2: observing the most pronounced demand.
• 5:42
  Speaker 1: Stress, if you would. And energy too, Emmanuel.
• 5:45
  Speaker 2: Yeah, that's right.
• 5:46
  Speaker 1: Tell us a little bit about that. I mean, didn't
• 5:48
  Speaker 1: the government, have the government actually said to these data centers,
• 5:50
  Speaker 1: you must, what's the deal, that they must provide their
• 5:53
  Speaker 1: own energy or?
• 5:54
  Speaker 2: Yeah, yeah. Well, that's the really interesting aspect to it, James,
• 5:58
  Speaker 2: because with data centres, we've observed that the location of
• 6:03
  Speaker 2: the data centres is obviously critical due to matching demand
• 6:07
  Speaker 2: and supply, for instance.
• 6:08
  Speaker 1: So we've seen sort of the.
• 6:10
  Speaker 2: Majority being located in Sydney and Melbourne, being the sort
• 6:15
  Speaker 2: of metropolitan areas in Australia, but also What's quite notable
• 6:21
  Speaker 2: is that it's actually access to the power grid that
• 6:25
  Speaker 2: seems to be the gating item here. It's not about
• 6:29
  Speaker 2: capital or land itself. And that's why we're seeing a
• 6:33
  Speaker 2: lot of these previously industrial real estate operators suddenly become
• 6:39
  Speaker 2: far more prominent in this aspect, given industrial sites have
• 6:43
  Speaker 2: that sort of access to higher voltage power.
• 6:48
  Speaker 1: I saw there's an application then from a Singaporean group
• 6:50
  Speaker 1: to use the old Ford factory in Melbourne, which is enormous.
• 6:54
  Speaker 1: I mean, it's several times the size of the MCG.
• 6:58
  Speaker 1: But it did strike me that sitting out there, of course,
• 7:00
  Speaker 1: and it had the infrastructure around it, it had the
• 7:02
  Speaker 1: power coming into it. Okay. And then also you have
• 7:06
  Speaker 1: things like fascinating, like Firmus, which we'll probably talk about
• 7:08
  Speaker 1: in a while, getting approval to build a data center
• 7:12
  Speaker 1: in Bell Bay, which most people, of course, would regard
• 7:15
  Speaker 1: as heavy industry. Must... On the political side, there is,
• 7:20
  Speaker 1: in very broad terms, there's political support for an economic mini-boom,
• 7:27
  Speaker 1: and data centers fit the bill there. There is the
• 7:30
  Speaker 1: sense that there will be an enormous investment, and certainly
• 7:33
  Speaker 1: in the initial phases, enormous construction, activity, and employment, though
• 7:38
  Speaker 1: not necessarily later on, when these are built a bit
• 7:41
  Speaker 1: more like toll roads really than LNG gas. At least
• 7:44
  Speaker 1: LNG gas continues to export for years and years and
• 7:47
  Speaker 1: that's actually good for the economy. So I'm just trying
• 7:50
  Speaker 1: to put some political context in here and community unease
• 7:54
  Speaker 1: and protest about the location of data centres. Can you
• 7:57
  Speaker 1: tell us a little bit about that? Data centres, do
• 8:00
  Speaker 1: they need to be in cities or how does it work?
• 8:04
  Speaker 2: Yeah, look, I think that ultimately data centres... It's not
• 8:08
  Speaker 2: mandatory for them to be in cities themselves. But however,
• 8:13
  Speaker 2: to minimize the computational lag in terms of the time
• 8:18
  Speaker 2: required to dispatch and receive information, I think that really
• 8:24
  Speaker 2: will validate the rationale why they're located typically in metropolitan areas, preferably.
• 8:30
  Speaker 2: But however, as the build-out progresses, we're increasingly seeing more
• 8:36
  Speaker 2: data centers being located in less sensitive areas.
• 8:39
  Speaker 1: Right. And that access to energy, if it's as important
• 8:42
  Speaker 1: as it sounds, then we were just talking about the
• 8:45
  Speaker 1: Bell Bay one. What they are attracted to, of course,
• 8:48
  Speaker 1: is that they can sell, if you like, themselves to
• 8:51
  Speaker 1: global investors as clean because they have hydropower down there.
• 8:55
  Speaker 1: So you have this sort of maturing of the industry
• 8:57
  Speaker 1: and what it actually is. Just before we take a break,
• 9:00
  Speaker 1: we'll go to the part that I know you want
• 9:02
  Speaker 1: to hear about, folks, which is how do we get
• 9:03
  Speaker 1: in here and what do we invest in? Very broadly,
• 9:06
  Speaker 1: risks and opportunities here. It's very global, right? Many of
• 9:11
  Speaker 1: these data centers are being built for global clients, aren't they?
• 9:15
  Speaker 1: And money is being raised for global clients. How much
• 9:18
  Speaker 1: of this is innately Australian? Look, I think.
• 9:22
  Speaker 2: That my best estimate would be that only the minority,
• 9:26
  Speaker 2: I guess, as you say. I think that obviously a
• 9:30
  Speaker 2: lot of demand would be from the large global hyperscalers.
• 9:34
  Speaker 1: Just explain to listeners who they are and what that means.
• 9:38
  Speaker 2: Yeah, sure. So these would be the large AI companies, Anthropic, PanAI,
• 9:44
  Speaker 2: even Google and Meta as well. So these are sort
• 9:49
  Speaker 2: of very much consumer-facing businesses that are looking to build
• 9:53
  Speaker 2: out their Australian infrastructure to service Australian clients primarily. That's
• 9:59
  Speaker 2: my opinion.
• 10:01
  Speaker 1: So we don't necessarily have champion companies in this area.
• 10:06
  Speaker 1: And even the data centers, when they are built, Tim Toohy,
• 10:10
  Speaker 1: who has been on the show a few weeks ago
• 10:11
  Speaker 1: of Yarra Capital, the economist, has made a really interesting report,
• 10:16
  Speaker 1: which clearly has been picked up and has now become
• 10:18
  Speaker 1: part of the political debates. which is that when they
• 10:21
  Speaker 1: do build the data centers, roughly 66% of those data
• 10:26
  Speaker 1: centers are imported material, like literally the servers and the chips,
• 10:31
  Speaker 1: et cetera. They're coming from overseas. So there is that aspect.
• 10:36
  Speaker 1: Watch how the big numbers you're hearing may, I won't
• 10:40
  Speaker 1: say they're exaggerated, but the extent to which it percolates
• 10:43
  Speaker 1: through to the local economy is limited. That is only
• 10:47
  Speaker 1: of limited interest for investors. They want to know What
• 10:50
  Speaker 1: is the way in? And we will deal with that
• 10:52
  Speaker 1: in a moment. Hello, welcome back to the Australian's Money
• 11:11
  Speaker 1: Puzzle podcast. James Kirby here with Emmanuel Datt from Datt Capital.
• 11:15
  Speaker 1: Emmanuel is a fund manager and the fund is accessible
• 11:19
  Speaker 1: through a lot of the very well-known platforms that are
• 11:22
  Speaker 1: out there. He's had a very interesting career so far,
• 11:25
  Speaker 1: but his big goal was to start an absolute return fund,
• 11:28
  Speaker 1: which is long and lonely, doesn't short. So it's really
• 11:31
  Speaker 1: all about stuff picking. How did you, when you would
• 11:36
  Speaker 1: have been familiar, obviously, with the notion of data centers
• 11:38
  Speaker 1: way back, Next DC, et cetera, that it was coming
• 11:41
  Speaker 1: down the line, and then AI just sort of put
• 11:44
  Speaker 1: this stuff, put rocket fuel into this whole scene, didn't it?
• 11:46
  Speaker 1: How did you as an investor approach it? Did you
• 11:48
  Speaker 1: go looking for AI stocks, or did you look for
• 11:52
  Speaker 1: proxies that were like picks and shovels of this boom? Yeah, sure, James.
• 11:57
  Speaker 2: I think that as an investor, you have to consider
• 12:03
  Speaker 2: the layers that are investable. As you mentioned, DC is
• 12:08
  Speaker 2: over direct exposure in the sense that it builds and
• 12:11
  Speaker 2: operates data centers. However, we took a more derivative approach
• 12:16
  Speaker 2: that was investing in companies that were providing services to
• 12:22
  Speaker 2: deliver upon this capex and delivery, essentially. So these companies
• 12:29
  Speaker 2: can be those such as Mass Group, MGH, Southern Cross.
• 12:35
  Speaker 1: MAS, the Mars Construction Group, which has pivoted towards data
• 12:41
  Speaker 1: centers big time, haven't they? Yeah, that's right.
• 12:44
  Speaker 2: And somewhat unusually, they've gone upstream into more direct exposure
• 12:49
  Speaker 2: via Firmus, but however, obviously still continue to operate an
• 12:53
  Speaker 2: electrical contracting business. And a couple of others that I
• 12:58
  Speaker 2: thought probably worth mentioning are Southern Cross Electrical Group. They've
• 13:02
  Speaker 2: been sort of a long-established electrical contract on the ASX,
• 13:07
  Speaker 2: on that as SXE.
• 13:09
  Speaker 1: And they strategically went for the data centers. They just
• 13:12
  Speaker 1: became a data center electrification specialist.
• 13:16
  Speaker 2: No, they are quite diversified. So typically, traditionally, they've been
• 13:21
  Speaker 2: more in mining and industrial sort of electric services. However,
• 13:27
  Speaker 2: due to the enormous emphasis on electrical infrastructure that these
• 13:34
  Speaker 2: data centers require, that certainly pulled demand into the the
• 13:39
  Speaker 2: data center sector away and made it relatively a larger
• 13:44
  Speaker 2: slice of the pie than perhaps mining or traditional industrial contract. Yes.
• 13:48
  Speaker 1: And so they won some big tenders, did they, to
• 13:51
  Speaker 1: build these? Yes, yes. Yeah, I see. So that's what
• 13:54
  Speaker 1: you followed. You followed the money, as they say. What
• 13:58
  Speaker 1: about property trusts? Goodman, for instance, where 80% of the
• 14:01
  Speaker 1: future revenue is now in data centers. How did you
• 14:04
  Speaker 1: do it? How did you assess them?
• 14:06
  Speaker 2: Yeah, so Goodman are an interesting company. case study, I guess,
• 14:10
  Speaker 2: of a large established industrial player, realizing that it held
• 14:16
  Speaker 2: aspects within its portfolio that were highly attractive for data centers,
• 14:20
  Speaker 2: which is number one, high voltage power access into its
• 14:26
  Speaker 2: industrial properties, and of course, water as well. So we've
• 14:30
  Speaker 2: seen Goodman really put a lot of emphasis on developing
• 14:35
  Speaker 2: a data center division.
• 14:37
  Speaker 1: I mean, I would always say they've bet the company
• 14:39
  Speaker 1: on it. I mean, 80% of their future revenues are
• 14:41
  Speaker 1: linked with data centers, eh?
• 14:44
  Speaker 2: Yeah, well, I think that there's such a value differential between,
• 14:48
  Speaker 2: you know, running sort of an industrial portfolio or developing
• 14:52
  Speaker 2: an industrial portfolio and developing data centers, given the weight
• 14:55
  Speaker 2: of capital.
• 14:56
  Speaker 1: So just tell us as investors, what do you think
• 14:59
  Speaker 1: investors were right to run behind Goodman and see it
• 15:02
  Speaker 1: as a data center proxy? Or are you skeptical about
• 15:05
  Speaker 1: that notion?
• 15:07
  Speaker 2: I think that it is the investors are correct. Goodman
• 15:12
  Speaker 2: is such a dominant player in Australia's industrial space, full stop.
• 15:19
  Speaker 2: And I think that ultimately, you know, due to the
• 15:23
  Speaker 2: importance of location and access to these critical services, I
• 15:27
  Speaker 2: think that investors are well warranted to, you know, and
• 15:31
  Speaker 2: a proven, you know, seasoned management team as well, not
• 15:35
  Speaker 2: to forget that aspect. So I think that, yeah, it
• 15:38
  Speaker 2: was an attractive play to take that exposure.
• 15:42
  Speaker 1: MAAS, we mentioned Southern Cross Electrical, we mentioned Next DC,
• 15:47
  Speaker 1: we mentioned Goodman as a property proxy into this area.
• 15:51
  Speaker 1: Have I missed anything? Who else, if I was an
• 15:53
  Speaker 1: investor looking at the data center space, who else should
• 15:56
  Speaker 1: I be aware of?
• 15:57
  Speaker 2: Yeah, I think that IFT is another example. Infratel is
• 16:02
  Speaker 2: another company.
• 16:03
  Speaker 1: Infratel, of course, yes. The big New Zealand-owned infrastructure player.
• 16:08
  Speaker 2: Yes, that's right. They're sort of a little bit more diversified,
• 16:12
  Speaker 2: and they have exposure across a number of segments. Upstream Energy,
• 16:17
  Speaker 2: for instance. And the exposure is a little more diffuse,
• 16:21
  Speaker 2: but they should benefit over time, I'd imagine, if this continues.
• 16:26
  Speaker 1: And they saw Diacenters as an infrastructure player. They just
• 16:29
  Speaker 1: saw it As new era infrastructure, basically. Yeah, that's right.
• 16:36
  Speaker 2: That's right.
• 16:36
  Speaker 1: Okay. So now we went in firmness. This is going
• 16:40
  Speaker 1: to be the biggest IPO we've seen for many years.
• 16:43
  Speaker 1: And when it comes to the market, it's an Australian
• 16:46
  Speaker 1: owned data center player. It has all sorts of colorful
• 16:50
  Speaker 1: aspects to it. And it's a very much a plan,
• 16:54
  Speaker 1: very much a plant, but it's a big plan. And
• 16:57
  Speaker 1: we've seen a lot of people go in, big names
• 16:59
  Speaker 1: go in early in this, into this group. So there's
• 17:02
  Speaker 1: a lot of pressure on it to list on the
• 17:04
  Speaker 1: share market. When it does arrive at the share market,
• 17:07
  Speaker 1: it will be pitched up to mom and dad investors
• 17:09
  Speaker 1: who invariably are late to the party on any boom.
• 17:13
  Speaker 1: And where do you think we are in this boom?
• 17:16
  Speaker 1: Are we still in the early stages? Are we well
• 17:19
  Speaker 1: down the line?
• 17:20
  Speaker 2: Look, I think my best estimate would be that we
• 17:23
  Speaker 2: are probably over the peak of this boom.
• 17:27
  Speaker 1: Of the capital expenditure. Yeah, yeah, absolutely.
• 17:31
  Speaker 2: In the sense that I think that Firmus as a
• 17:33
  Speaker 2: company have managed to capture the Z just very well
• 17:38
  Speaker 2: and raise a lot of capital privately. However, as you
• 17:42
  Speaker 2: point out, I think that there must be significant pressure
• 17:45
  Speaker 2: to bring this to the public market itself, given private
• 17:50
  Speaker 2: investors or financiers can be sort of, you know, The
• 17:55
  Speaker 2: duration of capital varies according to the investor, ultimately. And
• 18:00
  Speaker 2: I think that this has been very much pitched as
• 18:02
  Speaker 2: a public market story for some time. And so I
• 18:06
  Speaker 2: think that it's fair to say that there's a fair
• 18:09
  Speaker 2: amount of emphasis on getting this IPO over the line.
• 18:12
  Speaker 1: What don't you like in this boom? You know what
• 18:14
  Speaker 1: you're buying, right? You're in early. You saw the electrical
• 18:18
  Speaker 1: contractors in there, very safe way in, the electrical contractors.
• 18:21
  Speaker 1: Unless the hyperscaler or the data center goes under, they're
• 18:23
  Speaker 1: going to get paid, right? So it's very, it's old-fashioned business, right?
• 18:26
  Speaker 1: They want a tender. As long as they don't blow it,
• 18:28
  Speaker 1: they're going to get paid. What don't you like? Where's
• 18:31
  Speaker 1: the danger?
• 18:32
  Speaker 2: Yeah, look, I think the danger is always in investors
• 18:37
  Speaker 2: committing capital late in the cycle. And as I mentioned,
• 18:42
  Speaker 2: I think where I can't quite characterize it as being
• 18:45
  Speaker 2: at the tail end of the cycle because who knows
• 18:47
  Speaker 2: how long it will go on. But I think that
• 18:49
  Speaker 2: we are definitely over the peak. That's my perception because
• 18:54
  Speaker 2: we've seen a lot of big deals. We've seen a
• 18:56
  Speaker 2: lot of large delivery plans announced. So my question is always,
• 19:02
  Speaker 2: what's next? Where does the next bit of incremental demand
• 19:05
  Speaker 2: come from? And if it's not forthcoming, then was that
• 19:10
  Speaker 2: the peak a month or two ago?
• 19:13
  Speaker 1: Right. Very interesting. We'll take a short break and we're
• 19:15
  Speaker 1: going to have a look at some potential winners and losers.
• 19:18
  Speaker 1: Back in a moment. Hello, welcome back to the Australian's
• 19:35
  Speaker 1: Money Puzzle podcast. James Kirby with Emmanuel Dat of DatCapital.
• 19:39
  Speaker 1: And we're talking data centres. how to get in there,
• 19:42
  Speaker 1: should you wish to do so, if it's not too late.
• 19:45
  Speaker 1: All right. One thing I mentioned at the start of
• 19:47
  Speaker 1: the show, how the comparison is being made usefully, I think,
• 19:54
  Speaker 1: strictly on numbers, that it's very similar to the LNG
• 19:58
  Speaker 1: gas boom of the, what was roughly 2009 onwards, I
• 20:03
  Speaker 1: would think. And what happened, there was this concept that Australia,
• 20:08
  Speaker 1: could suddenly be a major LNG exporter because you could
• 20:11
  Speaker 1: suddenly convert the gas, basically, and put it on a ship,
• 20:14
  Speaker 1: ship it around the world. All came true. There was
• 20:17
  Speaker 1: many companies. Some were wonderful, some were hopeless, some were
• 20:22
  Speaker 1: in the middle. And there was all sorts of investment action.
• 20:25
  Speaker 1: And then later, it sort of calmed down. It became
• 20:29
  Speaker 1: part of everyday life. That big expansion phase was over.
• 20:33
  Speaker 1: That CapEx phase was over. There was a takeover boom
• 20:36
  Speaker 1: where a lot of the big companies around the world
• 20:37
  Speaker 1: mopped up the smaller ones that had emerged. How do
• 20:40
  Speaker 1: you think this data center boom will play out from here, Emmanuel,
• 20:44
  Speaker 1: which I think has been really interesting for our listeners
• 20:46
  Speaker 1: to see what you have to say. Sure.
• 20:49
  Speaker 2: Well, I think the LNG analogy is very apt because
• 20:54
  Speaker 2: my recollection at the time is that you had multiple
• 20:57
  Speaker 2: LNG projects competing for the same limited pool of contractors.
• 21:04
  Speaker 2: And that's quite similar. However, also a similar dynamic in
• 21:09
  Speaker 2: the sense that LNG prices are not static. They're very
• 21:14
  Speaker 2: fluid global markets. And it's the same with compute. For instance,
• 21:19
  Speaker 2: token prices do fluctuate. So at a high level, these
• 21:25
  Speaker 2: could be viewed as very similar sort of exposures. Ultimately,
• 21:31
  Speaker 2: you know, the lowest cost producers always win in any
• 21:34
  Speaker 2: sort of industry with this sort of dynamic. However, you know,
• 21:38
  Speaker 2: I guess the difference is in this situation is that
• 21:42
  Speaker 2: a lot of the consumers themselves are Australian-based, rather than
• 21:48
  Speaker 2: being based in Asia.
• 21:49
  Speaker 1: Are you saying not all the data centers we hear
• 21:52
  Speaker 1: will get built?
• 21:53
  Speaker 2: Yeah, I believe that is a fair statement to make.
• 21:57
  Speaker 2: I think that a lot really depends on you know,
• 22:00
  Speaker 2: the demand for compute rising over time. However, that doesn't
• 22:06
  Speaker 2: necessarily account for the progression in the underlying technology itself.
• 22:10
  Speaker 2: So in the AI models itself, and as we know
• 22:13
  Speaker 2: with technology, costs typically tend to trend down over time.
• 22:17
  Speaker 2: So I think the efficiency of these models will rise
• 22:21
  Speaker 2: over time and make it, you know, cheaper, you know,
• 22:24
  Speaker 2: cost to deliver over time.
• 22:26
  Speaker 1: If you're listening to the show today and you hadn't
• 22:28
  Speaker 1: been in so far, where would you What would your
• 22:32
  Speaker 1: top choices be?
• 22:34
  Speaker 2: Yeah, look, I think that if push came to shove
• 22:37
  Speaker 2: at this stage in the cycle, I would probably opt
• 22:40
  Speaker 2: to invest in derivative exposures, so the service providers to
• 22:46
  Speaker 2: the sector. And I think that irrespective of the continuing
• 22:51
  Speaker 2: data center investment, these obviously have other sort of exposures
• 22:58
  Speaker 2: to the broader economy rather than anyone going all in
• 23:03
  Speaker 2: on just sector. But also, you know, conscious of the
• 23:07
  Speaker 2: fact that once a data center is delivered, it does
• 23:10
  Speaker 2: require continuing maintenance to keep it functional. So I think
• 23:16
  Speaker 2: that that in itself can be beneficial for these derivative
• 23:20
  Speaker 2: exposures like electrical contractors.
• 23:23
  Speaker 1: They get the maintenance contracts. Is that what you mean?
• 23:26
  Speaker 2: Yes, that's right.
• 23:27
  Speaker 1: Right, right. Yes. Though the maintenance contracts probably need to
• 23:31
  Speaker 1: where every hundred that were in there, there's four in there,
• 23:34
  Speaker 1: if you know what I mean.
• 23:35
  Speaker 2: Yeah.
• 23:37
  Speaker 1: But there'll be a fair drop-off between active electrification and maintenance.
• 23:42
  Speaker 1: One last thing I want to ask you about, which
• 23:44
  Speaker 1: we've already tipped off on the show, but I think
• 23:46
  Speaker 1: people will find it very interesting. We're talking about these
• 23:50
  Speaker 1: potential floats. We're talking about how some of these companies
• 23:53
  Speaker 1: have done very well so far and may continue to
• 23:55
  Speaker 1: do so, servicing, if you like, the data-centric boom. But
• 24:00
  Speaker 1: in terms of dollars, the really big action has been
• 24:04
  Speaker 1: off the market. And one of them was in the
• 24:06
  Speaker 1: bond market where Google came in quite recently and literally
• 24:09
  Speaker 1: raised effortlessly, raised $ 5 billion in the biggest corporate bond
• 24:14
  Speaker 1: raising we've ever seen only a few weeks ago. There's
• 24:17
  Speaker 1: reports this morning that Amazon is going to follow with
• 24:20
  Speaker 1: something broadly similar. These guys, I mean, tipping off $ 5
• 24:24
  Speaker 1: billion at a go on the bond markets and these
• 24:27
  Speaker 1: raisings and private investors have been in there too. That
• 24:30
  Speaker 1: surprised you. These are the hyperscalers. These are, Emmanuel mentioned
• 24:34
  Speaker 1: at the start, the hyperscalers. This is literally Google, Amazon,
• 24:37
  Speaker 1: and Co. coming into Australia and saying, we want to
• 24:40
  Speaker 1: finance the AI boom, possibly through data centers, possibly through
• 24:45
  Speaker 1: whatever costs we have. They don't even specify to any
• 24:49
  Speaker 1: granular detail what it's for. And the money's flowing. People
• 24:54
  Speaker 1: are literally giving them $ 5 billion at a go on
• 24:56
  Speaker 1: our market. Think about it for the share market. if
• 25:00
  Speaker 1: you had five $ 1 billion floats that were linked with AI,
• 25:04
  Speaker 1: that would be the biggest thing we've ever seen. That
• 25:05
  Speaker 1: would just dominate all financial headlines. So what do you
• 25:08
  Speaker 1: think is going on? And just tell us a little
• 25:10
  Speaker 1: about your observations about what that means. Yeah, sure.
• 25:14
  Speaker 2: So I think that ultimately it goes back to demand
• 25:18
  Speaker 2: for fixed income. And I think that's become something that
• 25:23
  Speaker 2: has become increasingly prominent.
• 25:25
  Speaker 1: So they're offering 6%, weren't they? Sorry, Manuel, but they
• 25:28
  Speaker 1: were offering 6% on these bonds, yeah.
• 25:30
  Speaker 2: Yeah, but, you know, that sort of chase for yield,
• 25:33
  Speaker 2: I guess, in a population that is aging like Australia's,
• 25:38
  Speaker 2: but also, you know, the real deficit of alternatives, I guess, perhaps.
• 25:44
  Speaker 1: Yes, and high quality, the quality alternatives. Because whatever I say,
• 25:48
  Speaker 1: these guys are rated as good as, well, they're certainly rated,
• 25:50
  Speaker 1: they're rated, so under the joke, they're rated as well
• 25:53
  Speaker 1: as states. Yes, yes. As well as Victoria. Yes. It
• 25:59
  Speaker 1: is a state rate, it's a government class rating that
• 26:01
  Speaker 1: they're able to get.
• 26:03
  Speaker 2: Yeah, that's right. And I think that, look, ultimately, it's
• 26:06
  Speaker 2: also perhaps an exercise of these brands monetizing their recognition.
• 26:11
  Speaker 2: For instance, if you're a regular Google user, and you
• 26:14
  Speaker 2: have been for the last couple of decades, you probably
• 26:18
  Speaker 2: have a very, you have the impression that Google is
• 26:21
  Speaker 2: a very safe company and trusted. So I view this,
• 26:25
  Speaker 2: you know, they're probably legitimate treasury reasons for, you know,
• 26:30
  Speaker 2: partial issuance over here, just for diversification benefits, of course.
• 26:34
  Speaker 2: But I suspect that a lot of it comes down
• 26:37
  Speaker 2: to the fact that these are large global brands with
• 26:40
  Speaker 2: a reasonably high level of trust, I would say. And
• 26:43
  Speaker 2: there's clearly appetite for secure fixed income, you know, instruments
• 26:47
  Speaker 2: in Australian market itself.
• 26:50
  Speaker 1: It's the side of the data boom that not everyone
• 26:53
  Speaker 1: thinks of. They think of stocks. It's happening on the
• 26:57
  Speaker 1: bond market too, isn't it? On very big numbers. Yeah, absolutely.
• 27:01
  Speaker 2: Absolutely.
• 27:02
  Speaker 1: That's right.
• 27:03
  Speaker 2: And, you know, it's, I mean, to my knowledge, I
• 27:06
  Speaker 2: don't believe any of these, you know, like the Alphabets,
• 27:10
  Speaker 2: the Amazons. I think they have a pretty good track
• 27:13
  Speaker 2: record of repaying their bond issuances. So, and I don't expect,
• 27:18
  Speaker 2: I don't expect there to be any surprises basically on
• 27:21
  Speaker 2: that part.
• 27:22
  Speaker 1: So, Very good. I think we've covered the basis there, folks.
• 27:25
  Speaker 1: Very interesting in that Emmanuel's notion that, in fact, it
• 27:28
  Speaker 1: may have peaked in terms of that explosive period of
• 27:32
  Speaker 1: growth and CapEx announcements. And actually, maybe to some extent,
• 27:37
  Speaker 1: it's going to be a battle now for who actually
• 27:39
  Speaker 1: gets to build. Yeah, this is the rational part, like
• 27:43
  Speaker 1: who gets to build. They won't all get built, as
• 27:45
  Speaker 1: you said at one stage. So we'll keep a very
• 27:47
  Speaker 1: close eye on this. And it's something that I think
• 27:50
  Speaker 1: our listeners have been asking about. I thought it was
• 27:52
  Speaker 1: a great way to get you in this morning. Thank
• 27:54
  Speaker 1: you very much, Emmanuel Dart.
• 27:56
  Speaker 2: Thank you very much, James. Thanks for having us.
• 27:58
  Speaker 1: Great to talk to you. And that was Emmanuel Dart
• 28:00
  Speaker 1: of Dart Capital Group. Keep the emails rolling. I didn't
• 28:04
  Speaker 1: do questions there. I wanted to do just a theme today,
• 28:07
  Speaker 1: but we will be back to questions very quickly. themoneypuzzle
• 28:10
  Speaker 1: at theaustralian.com.au Today's show was produced by Leah Samaglou. Talk
• 28:15
  Speaker 1: to you soon.
• 28:20
  Speaker 2: Thank you.