Jordan Wong - Portfolio Specialist

Marcello Montanari, CFA - Managing Director & Senior Portfolio Manager, North American Equities

Robert Cavallo, CFA - Managing Director & Senior Portfolio Manager, North American Equities

Jordan: All right. Hi everyone. Welcome back to Tech Talk. If this is your first time tuning in, my name is Jordan Wong, and I'm here with my usual suspects, Marcello Montanari, Rob Cavallo, both managing directors, senior portfolio managers on the RBC GAM North American equity team. Marcello and Rob manage a number of growth focused strategies. Notably RBC Life Science and Technology Fund and RBC Global Technology Fund.

Given how things change and evolve very rapidly in the technology and innovation space, this is why we get together monthly, to talk all things technology and healthcare related. Gentlemen, thank you for being back once again. And a big thank you to you, Marcello. I know you're on vacation. You are dialing in from Aruba, so we really do appreciate that.

Okay. So, you know, before we get into this month's episode, I mean, we are now four months and change into the year. Obviously, AI is the dominant theme playing out across the U.S. equity market. Marcello, you have the stat but you know, it's basically 35, 40 businesses that are driving all of the upward earnings revisions within the S&P 500 right now.

All of those stocks are AI levered, AI associated businesses, mostly on the tech hardware, semiconductor memory front. But there's a number of industrial businesses as well that are very involved in building this out. So, this thematic remains very present. That being said, there's a lot of volatility underneath the surface. You know, the semiconductors and the memory stocks are doing really well.

At the same time, your average software business has been highly disrupted. As AI now sort of puts a big question mark on software business models and the longer term growth outlook. So if we kind of jump into today's episode, I'd like to talk a little bit about the recent earnings period, notably Meta and Google.

So they released and on the day that those two businesses released, their stock prices diverged quite a bit. Google has gone up, but Meta has gone down. So, maybe we could start off getting your thoughts on what is driving that price action. Was there anything from the recent earnings period that has maybe altered or solidified your views behind these two businesses?

Marcello: Yeah, sure. Just to clarify something, I think there's 40 names that account - this is before the last quarter, we haven't really run the calculation since the most recent quarter, but 40 names are responsible for half of the earnings growth and then the other 460 or whatever are the other half.

So it just shows the magnitude. It's also a question of the market cap size of some of these names as well. So yeah. So Google and Meta definitely kind of diverged. They both had really strong quarters. There was nothing wrong with them. In fact, at the top level, Meta actually grew faster than Google did.

I think Meta grew at around 29% and I think Google was like 22. Meta's daily active users continues to rise, everything is moving in the right direction. I think the biggest difference is that Meta uses AI for its own purposes. For the most part, it's all about delivering the right ad to the right person at the right time. So all the benefits that Meta sees from AI is kind of embedded within the core business. And it's really kind of hard to untangle it.

Whereas Google, because they've got this hyperscale business where they're actually selling compute, they're selling access to GPUs and everything as part of their hyperscale offering.

That product line is broken out. It's in their GCP product line. And that had just a massive rise in revenues and performance. And the outlook is just amazing. And it's visible. So because it's broken out, whereas Meta is all kind of embedded. On top of that, Google showed that there was margin expansion and for technology investors, you're always looking for that combination of revenue growth and margin expansion.

So you have this really big business, which is costing a lot of money to roll out. But you can actually see the benefit to that particular line item on their income statement. And you can see that their margins - I think the margins went from 18% a year ago, gross margins to 33% this year.

So you're seeing the upward trajectory. And that's what investors like. Meta, on the other hand, they showed some margin degradation just because they've hired a whole bunch of AI talent, very expensive AI talent. On top of that, there's depreciation that's working through. And then it's not broken out the way it is in Google's case, so you've got good revenue growth, in fact a little bit better than Google. But the margins compressing a little bit. And that's the biggest differences.

There's other things as well. Meta just lost a case in New Mexico. So that's weighing on the stock a little bit too. But going forward, we're still positive on both names. We're overweight both of the names. We think that Meta's new model is really quite powerful.

And it's based on, well, it's one of the first models that's going to come out on Blackwell's, the new chips from Nvidia. And so far, the early returns on it, it still needs to be basically fine tuned and stuff, it's looking like it's going to be pretty powerful.

I don't think it's going to be as powerful as GPT, but it's going to be up there. And remember, Meta just needs the model to be good enough for its own purposes. And so that's kind of the major differences. So we've had a divergence in performance. And we think as the new model comes out - we think if the model is strong and powerful, and remember that Meta is basically putting an AI window on top of all of its applications. And remember, they have 3.56 billion daily active users. Those models are going to get pretty good pretty fast.

So, we're still pretty positive long term on Meta for sure.

Jordan: That's excellent. Thank you. Rob, maybe you know, we talked about memory stocks either last month or a few months ago. They've been on an incredible tear. You know, the demand for memory is up here. The supply, you can't see it on the screen, but my hand is basically sort of touching the floor. That delta has led to significant price action.

And what is a very small group of stocks that produce memory? But they've taken a life of their own since we last spoke. Like these businesses have basically doubled again in the last month. So I think it's worth revisiting. What are your thoughts on what's going on now?

And is it still the same answer or the same approach as when we last talked about it or just the fact that these stocks have doubled? Plus once again, change things at all.

Rob: Yeah. Before I jump into that, I would just add one further point to the Meta Google discussion, and it kind of ties back into memory. You know, Meta got dinged for raising their CapEx. Google did it too. And the vast majority of the reason for the CapEx going higher is because of higher memory costs. So it's not necessarily greater volume of GPU capacity or TPU capacity.

It's just that higher memory cost. And Google, if you look at the actual leasing rates right now for GPU capacity as a measure of demand pricing is rising quite dramatically. So Google was able to pass on these memory costs. And to Marcello's point, Meta doesn't have that same capability. They have to absorb it. Another point I would add is, you know, in that whole discussion of 30 or 40 stocks driving earnings revisions, Micron's right at the top of the list.

And I don't know what the number is today, but it's a vast majority of the contribution of that 30-40 stocks is actually Micron whose numbers have just dramatically increased so far this year and a lot of it is due to the same factors we've talked about the past - we're still in an acute shortage phase and that's not going to be resolved anytime soon.

So when you look forward to estimates for this year, you look forward to estimates into next year. Like clearly there's probably upside pressure. Some of that is going to be limited by capacity restraints around what could actually be built. But if anything, pricing should remain strong. Demand is being driven by AI.

So we feel good about where the numbers situation and where the numbers set up into next year. The bigger question that we're battling with right now is trying to figure out the goalposts of an upside and the downside scenario because it's not the numbers that are going to come through first. It's going to be the first signs of supply shortages easing and then the market trying to grasp with what's the real demand in that next environment where supply doesn't drive the type of pricing that we've seen over the past six, 12 months.

So what's difficult is the stocks are going to start pricing that turn, the next inflection point before you actually see the numbers. We feel good that that inflection point is probably still more of a second half 2027 event at best. And even taking that point of view, you have to make the call that 2028, 2029 is going to be a down year.

And these hyperscalers are not going to continue to order as much memory capacity as they can. And actually, even if as supply comes back into equilibrium, you're going to still continue to see growth in the 2028, 2029. Which brings me back to a point. Is it cyclical? Is it structural? What we're seeing in this cycle for memory.

And I lean towards it being more structural. But clearly memory to some degree is still commodity. So there's a cyclical aspect to it. The way we're playing it is memory - Micron has been our favorite name in the group forever. It's the one name that we have that's a pure memory type of company. We have other names that we treat in sympathy, like in the copper currencies, but in terms of pure memory, Micron is our play because of our comfort with management, with our comfort with the balance sheet, with the fact that we'll be able to start buyback tremendous amount of cash.

So we feel like some protection there. And we are clearly nervous of the timing around when the memory trade starts to price the next stage. And we're watching very closely for things like spot pricing and things like lead time starting to plateau. But for now, we don't see that being a risk. And we'll stay invested in Micron.

But with a very strong eye on the door to figure out tactically when it's time to take some of the weighting out of the name and group.

Jordan: I'm curious like how significant could a rerating be for a business like Micron because it trades at a very low multiple from where we are today.

Rob: Yeah. Generally speaking, what's going to happen to Micron here is not so much on a PE multiple basis. It's just what is the earnings like. What's the real earnings number in outer year. So right now the street is saying they're going to do like $100 in earnings in 2027. And truthfully based on fiscal 27 which is August year-end, truthfully that number probably has upward pressure there, but they're doing over 80% gross margin right now. This is a type of business that at the low point of a cycle, can be negative margin. So the delta on what's going to drive this is not the P or price to book or any kind of metric.

It's going to be how, what's the direction of the intensity of the margin decline as the supply picture comes into balance. And then do the earnings go from $100 to $20? It's really more going to be on the earnings. You're not going to get rerating on a valuation perspective. It's going to be much more what's the margin and then the margin downside.

And then what does that mean from an earnings perspective.

Marcello: I'll hijack the meeting; closely associated with the whole memory issue is optical networking.

So this is a lot of the same trade, which is infrastructure and the optical links and the network optical networking names are moving along with a lot of them, the memory names as well. So we're playing a whole bunch of areas in the optical space. So we own Ciena optical networking.

We own Coherent, for example, in the components, Fabrinet in the manufacturing side as well as Celestica, there's Keysight and Keysight plays a key role because they do all the test and measurement equipment that goes around the implementation of all the gear and equipment. So, again, this is all part of the infrastructure stage of this.

And we're well represented in, and Corning obviously as well with fiber optic glass and everything. So we're well represented there. Sorry about that.

Jordan: No, of course that's the dry sandy air. Yeah. Maybe. Okay. So for the remaining part of the episode, I was thinking we could just maybe rapid fire through a few things that have popped up in the last few weeks. Okay. Maybe starting off Neo clouds. We're starting to hear about these more and more.

Maybe you could just talk a bit about what they are. What kind of where do they fit into the value chain? And you know, whether this has legs or this is a durable investment theme or not?

Marcello: I don't know about the ultimate durability. But Neo clouds play an important role because they're kind of infill at the moment there. Neo cloud, a lot of them actually repurpose Bitcoin miners to a certain extent. But not all of them, of course. But they're purpose built GPU factories or token factories.

And that's what they do. They basically just do GPUs, mainly for AI training and inference, whereas the hyperscalers do, you know, there's your big buffet. You have, they have just about every service under the sun from databases to SaaS hosting, you name it.

The hyperscalers do all of that. In terms of the customer base, the Neo clouds started off as doing the AI native type clients. But as the demand has been so large, what you're seeing is some of the hyperscale businesses now basically feeding off and going into the Neo clouds because they need the capacity themselves.

So that's what's basically been happening on that front. Typically the contracts on hyperscalers are more long term with more enterprise in the mix. And I don't want to say the hyperscalers don't do smaller clients because they do. But there's more enterprise in the mix, more long term contracts, more pricing that's established early on.

And the entirety of the contract, and the Neo clouds are more come as you go, a lot more short term pricing, spot pricing. And so but really the Neo clouds are just raw compute, GPUs when you need it on demand.

The hyperscalers are more, they've had more long lasting relationships. And again, there's now a crossover there. But the relationships, long term relationships, in terms of the end user customers, the enterprises has still rests more with the hyperscalers. But the Neo clouds are definitely playing a role here.

And we'll see how their businesses evolve over time. Some of them are pretty good position right now.

Jordan: Rob, we lost you there for a few minutes, but anything to add on Neo?

Rob: Yeah, sorry. And I apologize if I'm repeating anything.

Jordan: We can hear you. You are now horizontal, but-

Marcello: I think he froze up.

Jordan: He froze up. Okay. He's gone. Maybe he'll come back. We can circle back. Marcello. There was an article. OpenAI is thinking about building out a handset. So what are your thoughts on that? And does that threaten Apple at all?

Marcello: I wish I had a hundred bucks for everybody who talked about building a handset.

There's so many companies have gone out there saying they're going to build a handset, including Facebook. We saw what happened to BlackBerry here in Canada, while it was a global player, obviously. But you know, that market has been pretty much solidified between the iOS and the Android ecosystems.

I suspect if they did come out with something, it would probably be based on Android itself. So we'll see. But I mean, I wouldn't be too worried about them coming out with a handset. I don't see that happening. You know, things can always change. And there's definitely a discontinuity in terms of the technology base with AI and everything, but I wouldn't be making any investment decisions based on that right now for sure.

Jordan: Okay, Rob, I think we've got you back. We'll try to - I gotta go right to you while we have you, we might lose you midway through. No worries if that's the case, but Hantavirus, reading a lot about that these days. It's in the news, of course. Have any of the big health care businesses been talking about this at all?

And is this maybe, you know, a bit of a revival for a business like Moderna?

Rob: So at this point, I would say that I wouldn't freak out from a big picture perspective. You've seen Moderna trade up with a little bit of sympathy just given their vaccine play. But it's very different at this point than other like Covid. So I wouldn't jump to that yet there's differences in transmissibility.

But for sure you've seen some excitement in the vaccine space. Moderna most notably haven't really seen any reaction in the Pfizer BioNTech yet. But you've seen at the very margin, some incremental interest. But I don't think this is something yet to freak over. Say norovirus picked up some steam recently as well.

Clearly things are monitored because, famous last words, I don't want to die on that hill, but I think, I don't think this becomes a bigger issue or a bigger vaccine opportunity, but I guess we'll see.

Jordan: Okay. Well, knock on wood that this stays isolated. Okay, maybe lastly, we can end off with the xAI Anthropic Partnership agreement. Maybe just talk to us a little bit about that, what your thoughts are and how that maybe impacts the potential SpaceX IPO at some point this year or early next.

Marcello: I don't know if extra details have come out since I've been away for a few days, but if anything, I mean, it hints that xAI does not have that much faith in its own model, which is Grok. It's always been a lagging model.

A lot of its information or its training is based on stuff that's on Twitter or X. So I don't know, I think there's a sign there that maybe Elon Musk is shifting the business to actually turning xAI into leaning more towards becoming a hyperscaler at some point rather than a frontier model provider. On top of that, if I'm not mistaken, they just made another announcement with Google to provide capacity to Google as well.

So xAI working with Anthropic, you know, Grok competes with Claude and all of it and to be working with them is - I don't know, I have to say there's a message in there. I could be wrong. There's a lot of-

He's saying that he's got great faith in it, but his actions speak otherwise. So I guess we'll have to see, and if it is a change in strategy, I mean, what does that mean for the SpaceX IPO? The business is not fully baked yet. It's if it's going to be changing at this late stage before an IPO.

But again, you know, I'm saying this based on some limited information here.

Rob: Yeah, I would just add that clearly they're looking to add some revenue to go for run rate for what they're going to show on their roadshow. And it could also be a signal like maybe Grok model isn't going to be more than an internal model.

So maybe Grok runs for the different Musk companies without necessarily being something that they're looking to roll out for enterprise or consumer more broadly. And then they can rent space and share some of the costs on ramping up additional data centre capacity. Yeah. It's unclear to say, but it definitely an interesting development, but not one that we had thought about being a real possibility prior to both sets of announcements.

Jordan: No, that's really helpful. Thank you both. Okay. So that was all I had on the agenda here. Unless there's anything that you think I missed or that we should talk about. Otherwise we can wrap things up here and Marcello, I'm sure you got better things to do while on vacation.

So we'll let you get back to what it is that you're up to. Thank you both for your time, as always, and everyone who tuned in. We really appreciate your support. We'll be back next month. And if in the meantime, you do want to learn more about products like the RBC Life Science and Technology Fund, or the RBC Global Technology Fund, we do encourage you to visit our website, RBC GAM.com.

All right. Well wrap things off here. Thanks everyone. Cheers.