Presenting now, we are joined by Marc Jocum, Senior ETF Strategist from Global X.
Marc is a returning presenter with investment markets and his talks and his presentations and the information that he provides are always worthy of your time.
This time he's speaking to us about the next leg of the AI trade and where investors should look in 2027.
Great to have you Marc, over to you.
Thanks, Darren, and thank you everyone for tuning in.
It's great to be back on this platform.
I know I've done a few of these before, but this time it's going to be a little bit of a different lens where we're actually going to be talking about where investors should be positioning in 2027.
How should investors be thinking about the next leg or what we're calling of this AI trade?
And there's a few key things to paint a bit of the lay of the land of what's going on right now.
So AI has been a very topical focus in the markets.
It's driven majority of what you can see in the share market returns.
There's some great data showing the difference between companies that are exposed to AI and those that are not.
And most of the returns, since we had that ChatGPT moment back in 2022, has been driven about anything relating to AI.
However, there have been some wobbles and I think it's important to address some of those wobbles, particularly now leading into 2027, where at first it was about, we just need to spend as much as possible.
There's a great quote from the Microsoft CEO, Satya Nadella, who said that the biggest risk is not overspending, but actually underspending.
But now companies are reporting Q3 earnings in the US.
And certain companies are being punished for spending too much while others are getting punished for not spending enough.
So now the question turns to, well, how much return on investment are these companies going to actually get?
And what we've seen through this AI trade is we've seen semiconductors benefit massively.
They've been one of the best performers over the last couple of years.
Nvidia has been that poster child for share market returns and is really the bellwether for the semiconductor and AI trade.
And then it got pushed into some of the hyperscalers who have done exceptionally well.
They're investing a lot of their money in building out their data center capacity, investing in large language models to really offer and monetize this type of revolutionary technology.
Where we see one of the strongest opportunities at the moment is actually within the AI infrastructure space.
And I'll come to that in a second, but I want to address what we've seen with some of the concerns right now.
At the moment, people are a little bit worried that are we actually going to see
a return on that investment.
You've seen a lot of these big AI players actually go to the debt market to try and raise some money, not just from the equity market, but through bond issuance.
And what you've started to see is a lot of these credit default spreads, which is essentially just a barometer for how risky that these particular bonds are, they've started to rise a little bit.
So
People are starting to question, well, does this AI narrative still hold intact, right?
Our view is a fundamental yes.
Fundamentally, nothing has changed.
If anything, AI companies have become a little bit cheaper because companies are now being scrutinized for absolute perfection in terms of execution.
But there is still a lot of money being invested in this.
And in fact, if you normalize the amount of investment,
into AI to something that was like the dot com boom or the railroad build out in the early 1900s.
It's actually still quite small.
So my argument is we're actually not spending enough, which is why our view for 2027 and this whole theme of the next leg of the AI trade, it's let's look into the picks and shovels.
Let's look into not the huge returns we've seen in semiconductors, although we have a lot of conviction that a lot of semi-names actually look very cheap right now.
Not so much in the hyperscalers in the MAG7, but also down the track into the AI infrastructure build-out.
We're talking about the companies investing in building these data centers.
And there's been a bit of volatility around some of these data centers.
We had Next DC in Australia falling over 40% in a single day.
We had Firmus who had a failed IPO now looking to listen to NASDAQ.
So I'm not so worried about the particular data centers themselves.
Where I'd be looking as an investor is the companies actually building the data centers, looking at the cooling and the heating systems, the actual infrastructure of those data centers.
And not just that, but also what's going to power this, right?
We've seen a lot of investors here at GlobalX ask about copper mining and copper miners because the amount of intensive computational power
to, you know, essentially power a chat GPT query or a chord query, or to build these large language models and to have these multiple tokens.
It involves power.
And we're not going to get there based on fossil fuels and natural gas.
We are going to get there on renewable energy.
So we've seen a lot of people position in things like copper mining, because every gigawatt of data that's produced from one of these data centers requires about 65,000 tons of copper.
And we're also seeing people look for highly dependable and highly operational energy, such as nuclear power and uranium miners.
That's been two other areas that people have been very interested in.
And I feel like when you're blending all those three, that's where people are looking for that next leg in the AI trade.
It's not just about that.
but it's also about the broad commodities, right?
So we've spoken about copper, we've spoken about uranium and nuclear energy providers, but we're actually really bullish on this being a commodity super cycle.
you've seen a lot of investors start to position in looking for diversifiers, which commodities naturally fills that gap.
But also, there's been this de-globalization trade, right?
It's a complete opposite of what happened post the GFC, where in the 2000s, you had China joining the WTO, and it was all built about, you know, mutual relations, broader globalization.
Now, a lot of companies and a lot of countries, sorry, are building up their strategic reserves.
to essentially own these assets that's very critical for the AI build-out.
Think about your strategic and critical minerals and metals.
The whole tariffs that were going to be posed on copper led to a huge stockpiling of copper in the US because the US knows that it's going to be very critical for this build-out.
And we get a lot of questions asking, well, is it these Chinese low cost models that are going to win?
Is it the US models that are going to continue to lead?
Our view is it actually makes sense to own the entire system, the entire pie, because both of these providers are going to essentially
have what's called a Javon's paradox, where as costs get lower and lower, you get a mass amount of adoption.
And that's what we're seeing now.
I use AI in my everyday business, I use it in my personal business, and we're seeing it throughout the entire ecosystem, not just on the efficiency side of things, but also on the monetization.
And that's going to be key.
That is really key right now, which companies can actually monetize AI.
The other thing a lot of people ask us is, well, Marc, you've got these massive AI companies coming to the market.
Is that a bubble in itself?
We firmly believe that if you look at valuations and you look at earnings, we are not in a bubble because
These companies that are coming to market right now, we're talking about the likes of Anthropic, who's just essentially announced they want to come to market with a close to $2 trillion US IPO.
We've got the likes of OpenAI, which runs ChartGPT coming to markets.
And this is what's happening.
You get privates that are staying longer, and now they're coming to the market in a much bigger fashion.
And people are wondering, well, what does this mean for markets?
Our view is it doesn't matter who's going to win in this trade.
It's best to own the underlying theme that you're trying to get sectoral exposure to.
And these companies, while they might not be profitable right now, they are generating a massive amount of revenue.
They have enterprise agreements with clients.
retail and consumer-led businesses that are generating mass amounts of revenue.
Governments are using these tools.
There is mass integration.
This is a full ecosystem, which is why our belief is that with this next leg in AI, it really is going to be about those next order beneficiaries and how can investors position themselves in those areas.
And it's not just going to be large cap, we're seeing a lot of broadening happening right now within the AI space as well.
So think about going down the mid and small cap land as well.
A lot of people don't like being positioned in mid and small cap companies because high interest rates, and that's what we've seen.
We've seen massive bond yields rising, the highest they've been since 2002 in the US, the highest they've been in Australia since 2011.
Because one of the things about AI, and this is what's going to happen, even though I said costs are coming down, according to Javon's paradox, it's inflationary.
When you think about the mass amount of build-out that needs to happen, there's going to be inflationary pressures, which doesn't pose a very good tailwind for a lot of these small and mid-cap companies.
But if we start to see that broadening happening, and it's not just in the magnificent seven companies, we still like them.
If you think about it, those mag seven companies are not just seven companies.
They have hundreds of subsidiaries under them.
So you're actually getting a nice diversified portfolio with those big seven companies.
But then think about all the other mid and small cap layers that could benefit from AI adoption, from AI monetization, from feeding into some of these big, big players as part of their AI stack.
So when you do think about that next leg in AI, I think it's really important to not just think about the semiconductor companies and the memory names that have done so well.
It's not just about the big hyperscale of the companies.
It's what's going to be that next bottleneck.
And that next bottleneck for us is around the energy infrastructure, the grid build out, building out these huge data centers, and really servicing the needs of where all this CapEx is going.
We're talking about hundreds of billions of dollars spent per year.
Follow the money, right?
That's what we're telling a lot of our clients to see where is that money going?
And it's going into those next order beneficiaries.
To close things off, I think when investors are positioning for 2027, a lot of people might be concerned with valuations, but our view is valuations still look okay.
When you look at it from a normalised earnings perspective, Australia seems very expensive.
which is why I feel like here in Australia, we're going to see people prioritizing income.
We don't really have an AI ecosystem over here.
We're seeing people position in Australia as an almost anti AI trade.
People find the banks quite expensive.
So we really like the idea of being invested in banking credit.
But when you're looking for growth, you have to look overseas and you can't be underweight AI in this environment because that is what will in our belief will be continue to drive the market to get over these clearing events.
whether it's the midterm elections, whether it's earnings, whether it's some of the concerns around AI and some of the IPOs coming to market, people need to maintain a growth posture.
But the key is have a valuation discipline.
So leading into 2027,
Our advice is to position for growth, but you don't need to pay an arm and a leg for it, which is why we're seeing a lot of investors look at things like growth at a reasonable price as a particular strategy.
You still can get exposure to some of these AI names, which actually are trading at pretty good valuations if you look at some of these names, but you don't have to
stretch yourself and look for frothiness in the market because that could lead to some drawdowns.
So be agnostic to countries, just remember the AI trade is more than just the US, there's huge amounts happening in Asia, huge amounts happening in Japan over there as well and even parts of Europe as well.
So have a regional diversification in your portfolio, think about those next order beneficiaries in AI because our belief is that this AI theme is one to stay
People might be a little bit nervous around what's going on in the current market at the moment.
Our view is that's only a reason to increase positioning in some of these stocks.
There has been a lot of leverage that has been unwinded, especially I'm talking about the likes of Korea, where you saw mass amounts going into Samsung and SK Hynix.
That is a bit of healthy consolidation that needed to happen.
So don't be don't be nervous around AI.
You know, it's perfectly normal for revolutionary technology to have these concerns and to almost have that hurdle rate go higher and higher and higher.
But these are going to be massive, massive companies.
And this is a trend that I can see continuing for a number of decades.
This isn't a short term thing.
We've got a lot of conviction in this trend, but just think about where that next leg will be.
And our view is think of the beneficiaries, think of that next order catalyst and be
exposed to a diverse set of that AI ecosystem when you're looking for AI exposure.
So with that, I'll hand it back to Darren to go through a few questions, but hopefully that gives you a bit of the lay of the land in terms of where we are in AI, where we're positioned now in terms of valuations and what we could look for as some of the key building blocks for putting portfolios together and having exposure to AI for the year ahead.
Thank you very much, Marc.
And I think just to echo one of the points you've made, at investment markets, we've also seen sort of two really divergent themes emerge over the last sort of 12, 24 months between that split between income markets.
and income markets becoming progressively more and more of interest to investors and then external growth, US tech, AI trade.
And it seems to be two very distinct things and trends that are happening at the moment.
Yeah, I think it's almost like a little bit of a barbell approach that we're seeing, Darren.
So with that, I feel
Especially with the capital gains tax changes that have happened here in Australia, people are really starting to realize that growth investing is essentially not where you want to be over here.
You're going to get punitively taxed on that.
There's a lot of small and medium sized companies that might not choose to list you anymore.
People are moving overseas, but the income yield and the income profile of Australia is still very attractive.
So from a domestic side of things, we have the resource companies, which I think are actually
going to be the key driver of the markets.
And that is on the back of the AI trade, because that's where all the amount of money is getting spent on, especially the likes of BHP, Rio, Glencore, just listed here in Australia.
That's a copper trade right there.
But then also, think about income producing assets.
So Australians' dividend yield is one of the lowest we've seen in a couple of decades.
We've seen bond yields now rise.
Banking bonds are now paying more than banking equities and people concerned about the housing and credit market talking about a 12 to 15% drawdown in housing prices.
So it's not a reason to be, you know, anti Australia, but it doesn't paint a great picture.
So barbell it with Australia for income, think about bank credit, think about manufacturing yield through things like covered calls.
That's what we're seeing a lot of interest in the moment through clients.
And then to your point, Darren.
Position for growth and you have to go overseas for growth because the global share market has grown their earnings at around about 11 to 12% per annum versus Australia, which is 4% per annum over the last 20, 25 years.
Share prices follow earnings.
And if you want to be positioned for earnings, Australia doesn't have that, especially when it comes to AI.
You need to have exposure to the US market, to the Asian markets as well, because that's where all these AI companies are going to be.
And if you're competing for capital, you have to be in it to win it.
So that's where we're seeing, have your core allocation, your typical broad-based low-cost indexes, and have your barbells, your two ends of income generation, and then your AI growth tilts as well.
Very much so.
It doesn't present a great picture, Marc, of the Australian market or the opportunities here domestically.
However, that's maybe a topic we'll park away from this presentation for today.
I did, however, want to ask about 2027 and maybe sort of a consideration from your point of view around a
what keeps you up at night?
What would you be worried about that maybe black swan type event in 2027 that may disrupt or impact the AI trade?
If you were a contrarian, what would you be looking out for?
Yeah it's a great question Darren because it's very tough to view these left tail risks or these black swan events
I would probably want to see, the biggest thing would be that companies can't monetize AI and they actually are not generating any earnings.
If we start to see earnings go backwards and we start to see people pull back money from spending on AI, that would be a concern for me.
And it seems like all the big companies have really, with the exception of probably Apple, Apple's probably the only one Darren, I would say that has said, you know what, we're not going to spend a crazy amount of CapEx on AI and we'll see who's right.
And no one knows what the answer is going to be.
We'll see who gets punished in the end for people spending too much or not spending enough.
But I think that the key thing to monitor Darren and what I would be concerned about with the AI trade is if these companies are unable to monetize their AI build out, that would be a concern.
So leading into 2027, I still think there's going to be positive government support for this.
I think there's going to be a lot of spending.
but monitor earnings and see if a lot of these companies actually start to actually build out meaningful streams of revenue from their AI build out.
Can that translate into earnings?
We're going to have to wait to find out.
So far, the answer is yes.
Q3 earnings in the US still look pretty good.
We've had a little bit of wobbles along the way.
People are a little bit concerned here and there.
But overall, there's just such a diverse set of companies.
And the question we get a lot, Darren,
You've seen a lot of people concerned around these AI agents essentially hacking into tools and getting access to data that builds a case for this broader ecosystem of things like cybersecurity.
Because if you have a look at things like cybersecurity, which fell about 30 or 40% because of this AI disruption.
If you look at the interviews of the leading chief information and security officers, most of their budget is going into AI.
They're almost trying to build up defense in their own systems so that they can't get hacked.
I wouldn't be concerned about the hacking element.
I'd be concerned about can these cybersecurity companies and all these other companies not monetize this AI thing.
I think that's going to be the key for investors and not just on an individual retail level or consumer level.
It needs to be driven on the enterprise level as well.
There's a massive amount of market over there.
So follow the money, but also monitor the earnings.
That is what will keep me up at night.
And then maybe one follow on question from that, Marc, just referring back to the follow, follow the money comment.
If you were to look at the wider ecosystem, so it is quite complex, there's lots of different parts, lots of different components, lots of different companies.
Is there one particular sector of that sort of broader thematic or ecosystem you think is best place for 2027?
Yeah, I mean, I did speak about it earlier, Darren, and I'll double down on it because it really is the AI infrastructure layer.
It's the development layer.
So we've got an ETF which trades on the ticket code AINF.
It's been one of the most popular thematics so far this year.
It hasn't shot the lights out like semiconductors has done, but that's because some of these semi-names were the ones being highly leveraged and the ones being bought as single stocks.
A lot of people aren't really referring to some of these names like Train Technologies, which has a lot of the cooling systems for AI.
You've got companies who are actually doing the physical build out.
That would be my pick for what I'm seeing.
And that's this whole idea of broadening this AI pie.
Cybersecurity was a call that we had at the start of this year, which has actually been really good.
We feel that was oversold.
We still like semiconductors, don't get me wrong, but I still feel this is going to broaden out into things like, you know, even healthcare to things like humanoid robotics.
There's been a lot of talk around, well, what does this mean from a physical AI standpoint?
And that's what we're referring is how can this AI be applied to multiple different areas?
So when you're positioning for AI, we like to say that a lot of people prefer the lazy man's approach or the lazy woman's approach to AI by investing in a diversified AI ETF like our GXAI fund.
But then if you're trying to play those individual granular areas, AI infrastructure, Darren, is where I want to be positioned because I'm following the money, I'm following the capex of where people are spending, and that money is going into the build out.
So these companies who have the build up, they have really strong pipelines, right?
And they don't really care about the return on investment from these mega cap tech companies.
What will concern me and what will break that basis if we start to see companies push back on spending because they have targets around earnings and targets around buybacks or something else.
So that would be the one thing that would derail my view in terms of high conviction calls for 2027.
But otherwise I think that's where a lot of the money and a lot of the returns has yet to be materialized from the market just yet.
And I think, Marc, if you think back to a lot of the sort of exuberant periods in Australian history and gold mining and so on, it's always the picks and shovels is where the majority of the money is eventually made.
It is.
And I think especially now, and it ties into this whole thing that we're seeing around broad-based commodities as well.
I think people are going to prioritize these hard assets and these real assets, right?
So they do provide a really good defensive part of portfolios.
When you think about precious metals like gold, and you mentioned gold mining, Darren, you do have some idiosyncratic risk or company specific risks with gold miners.
They haven't really been involved in the AI trade, but people have been playing gold from a debasement trade, from inflationary pressures.
So I almost like blending gold because we actually are quite constructive on gold, even though the high yields have been a little bit of a headwind for it.
We actually like pairing that with the AI trade, but you can play the AI trade through these miners, is what you were saying.
And it's not mass exuberance that we're building out all this supply for the demand not to come.
We think we're going to be in structural deficits for some of these commodities.
And that was the key difference with what happened in the 2000s of the previous commodity boom that we saw.
was there was too much supply and the actual build-out never and the actual demand never came.
Here we have a very fundamental different approach and I'm always conscious of using that word Darren where this time is different but we have a very very polarised and fragmented geopolitical landscape
that's going to almost see the benefit and the need for people to earn these critical resources, which of natural benefit will be some of these mining companies.
So it could be a golden era for investors.
It's just about how you're positioning that.
And at the moment in Australia, everyone wants some of the miners.
And we think that you just need to think more broadly around natural resources, around hard assets, around broader based part of AI to monetize some of this potential.
Well, ultimately, Marc, I think one of the truisms of markets is that supply and demand is a very difficult thing to fight against.
It is.
And when you're thinking about your, and I'm going back to my economics, you know, university courses around where that equilibrium point is.
Things can move up and down the supply and demand curve right now.
And I think a lot of people are building up their strategic resources.
And it's a real shame in Australia because we do have those resources.
We're on a really strong stockpile of uranium reserves.
Lithium miners, copper miners, we really have a lot of the tools at our disposal over here, but we haven't materialized it as much as we can, excuse the pun.
So as long as there's to be consistent demand from some of these huge, huge players in the market,
And it's also well supported from the government and from a fiscal perspective.
I can't see that that supply to overshoot at the moment, especially when you've got tariffs from Trump coming in on some of these strategic reserves, people are building up their national resilience.
You're seeing it in the likes of defense spending as well with people building out their defense budgets in terms of the targets that they're trying to hit, because we are going to live in a very polarized world, a very fragmented world.
There's elections going on all across the world.
People can't really agree.
Do I want to be left-leaning?
Do I want to be right-leaning?
I feel like this polarization is consistent with what you're seeing in the world.
And one of the best forms of defense is to actually ignore some of that and focus on these key mega trends.
So going back to your supply and demand side, the demand is being driven from AI.
The demand is being driven from technology.
Supply needs to catch up and it's going to take a long time to catch up in terms of the build out that we need to serve as some of the AI demand.
That's very true Marc and thank you for those insights.
We appreciate your time today as always illuminating and I think of value to the audience and our investors.
Thank you.
My pleasure Darren.
Thanks again for having me and if anyone wants to find out more jump onto our website or feel free to contact us through your local manager.
I'm also available on LinkedIn if anyone wants to send me a message but thanks again Darren.
Thank you Marc.
The next phase of the AI trade may look different from the last. Presented by Marc Jocum, Senior ETF Strategist at Global X, and hosted by Darren Connolly, this video looks at how investors might think about AI exposure in 2027 beyond the headline names.
Marc begins with the recent market volatility around the AI trade, including questions about return on investment, rising credit spreads and the scrutiny companies now face over their spending. He explains why he believes the underlying story may still be intact, and why the infrastructure supporting the AI build-out is an area of interest. He also discusses the broader commodities theme, the potential for AI benefits to spread to mid and small cap companies, and the importance of valuation discipline and regional diversification.
In the discussion with Darren, Marc compares the Australian and global share markets, suggesting a barbell approach that pairs income exposure with growth exposure overseas. He also covers what could challenge the AI trade in 2027, why he favours the AI infrastructure layer, and how gold and hard assets might fit alongside an AI allocation.

