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: Hey, everyone. Welcome back to Tech Talk. Of course, the monthly web series, where myself, alongside Marcello Montanari and Rob Cavallo, get together to talk all things technology and healthcare related. If you're new to the program, Marcello and I spent the last few weeks visiting cities like Toronto, Ottawa, Montreal, talking a lot about the series and so if you caught us in one of those three cities and are tuning in now, we welcome you.

Jordan: For those newer, like I mentioned, we get together monthly, Marcello and Rob, two of the most experienced growth managers we have at RBC GAM, best known for their work on strategies like the PH&N U.S. Growth Fund, the RBC Life Science and Technology Fund, and the RBC Global Technology Fund. So we get together monthly just to kind of talk all things technology, healthcare related, keep you up to date and informed on what's going on within the space.

Jordan: So, gentlemen, welcome back. It's good to have you again.

Marcello: Good to be here.

Jordan: Okay, so, I mean, it's always the case. There's always plenty to talk about. I'm not sure we'll get through everything I've jotted down today, but if not, we'll pick it back up next month. Maybe a good starting point, though, would be to talk about some of the news that Anthropic released not too long ago. They came out reporting some revenue growth figures, substantial growth. In fact, I've got a little chart up on one of my screens here, appears to be basically 2X growth between February and April, they reported roughly $30 billion in revenue. Very broad, open-ended question - would love to just get your thoughts on what this news means to you. Presumably it's positive. But we'd love to just get your thoughts on that news from Anthropic.

Marcello: Yeah. It's hard to describe how unique this is. In all my years doing this, I've never seen anything like this. The growth rate at Anthropic is spectacular. Kind of at the end of February I think it was, they announced that they were at a revenue run rate of like $19 billion.

And then just the other day, they said they were on a revenue run rate of $30 billion. So that means that they added $11 billion on an annualized rate. So kind of think about that as almost an extra billion dollars more than they maybe they were expecting it, but an extra billion dollars in the month of March versus February.

So when you annualize that billion, it comes out of $11 extra billion dollars. So that's why I said it's not totally a billion, but this is mind boggling. Clearly it's been driven by Claude Code, the coding agent, which has really taken the developer community by storm, and these are power users, right?

So they're basically using Claude Code to program whatever it is that they need to program and when you think about it, they're writing like large programs for the most part that use a lot of tokens. And a lot of people don't understand this, but the way a large language model works is basically it's kind of guessing the next word.

So what happens is, it starts from the process and goes forward, guesses the next word, and then it goes back again, guesses the next word, goes back, and it goes around in a cycle. So as the programs get bigger, it actually uses more and more tokens. From my point of view, I believe that what's driving this revenue run rate increase is just, you know, the developer community just using it wherever they can to code their programs.

And the programs are obviously getting bigger. On top of that, they can use agents and run multiple projects at the same time. So they're basically burning through a lot of tokens, and they get charged. There's a consumption element usually within an enterprise. There's a base fee with a certain amount of tokens.

Once you go over that, there's an overage charge. So clearly Anthropic’s coding agent is being used extensively. That's a great thing for the industry. They're burning a lot of tokens, and that's why you get these phenomenal numbers.

Jordan: I'm curious, some of the other models, are they reporting any revenue figures or is it just Anthropic at the moment?

Marcello: Well, every now and then you get some numbers out of OpenAI and so on, so forth. But it's just Anthropic’s the one that really kind of caught everyone's attention. It's clearly like, as we've spoken in the past month, everything's kind of moving around an ecosystem. So you've got the Anthropic ecosystem, you have the OpenAI ecosystem.

You might have a Grok ecosystem. And obviously there's the Google ecosystem, and there's different players in each of those. There's some overlap, but for the most part, there's different players in each of the ecosystems and the models, as they get better, they kind of one up each other as we go along, and so you'll see one ecosystem rise at the top.

Usually it's kind of rewarded in the stock market. As one rises, one kind of falls off. If we go back a year ago, Google was viewed as being kind of “a has been”. Suddenly the developers, with Gemini three, Nano Banana, and Antigravity, like the Google models, were viewed as being really powerful. And they rose to the top.

As a result, Google was like the number one performing Mag Seven stock last year. So right now, what's happened is Google's come down a little bit. Anthropic is kind of the top, Open AI is kind of viewed as right now, they're the guys who can't seem to get everything straight.

But they're going to come out with a model shortly that's all based on the Blackwell chip. Every indication is it's going to be a very powerful model. We've already seen some early indications of it because the OpenAI Codex coding engine is being used extensively by developers, and they're raving about it.

So it's going to be pretty promising, I think. So we might see a kind of a juggling of the order again. And maybe that leads to different just stock price reactions.

Rob: I think what's important to kind of really put in front of investors attention here is the big debate: are we on all the spend? I think when you look at the growth rate that you're seeing at Anthropic, the growth rate you're seeing at OpenAI - what you're seeing from all the different cloud vendors and such, you are starting to see that ROI really materialize.

That's something that's just going to continue to happen over the next 6, 12, 18 months. Just like for some other numbers here, OpenAI's advertising went to a $100 million annual run rate in two months. Amazon’s CEO, Andy Jassy, came out in his annual letter just today, and talked about being at a $15 billion AI cloud revenue run rate at the end of Q1.

For context, and this is not apples to apples because AWS cloud growth when that started, it's a different context of where the market was and all that. But for perspective, AI cloud is 260 times bigger than where AWS cloud was in for its time period from birth to today.

Just to put it in context, there are real revenue being generated, real ROI being generated. Hopefully this puts some of that bear argument. It's not going to all put it to bed in one fell swoop, but as these kind of build on top of each other and stack on top of each other, it'll hopefully start to address that ROI concern that's overhang this whole space for a very long time now.

Jordan: Yeah, good point, Rob. Actually, you just got me to thinking. We talked, last episode, around NVIDIA earnings; pretty standout quarter. Of course, the market reaction was negative. In fact, on the stock, a lot of those concerns are around visibility earnings, and visibility over the next call it a year and a half, or two years.

Does this number from Anthropic kind of quell some of the market's concerns around that, like how do you interpret this information from the point of view of a business like NVIDIA?

Rob: Yeah, they just pulled a one off. When I say the biggest thing with NVIDIA is, yes, the durability that's been - because the market wants to see real ROI to get comfort around the CapEx data at, and growing at, a certain level over the next several years.

This is one data point. You know, one data point in itself isn't going to solve that and completely put the barricades to rest. But, this is clearly a positive one, just showing real returns and real stellar growth that, you know, we already talked about. But you know, NVIDIA is generally more associated with the OpenAI ecosystem.

What's important is this mythos model, this next gen of what Anthropic is doing is a separate announcement from this whole revenue discussion, a lot of that was trained on NVIDIA Blackwall chips. So it's expanding the breadth of NVIDIA offering beyond just the perception that it's an OpenAI vendor or a part of the OpenAI ecosystem.

So definitely a good data point for NVIDIA. Again, not in itself something that's going to address the overarching concern, but clearly, as more and more of these use cases and more and more tangible points of real cash is being generated, that's going to be good for the stock and kind of speak to our confidence.

Why? This spend on GPUs and the spend on overall AI infrastructure is going to continue to grow for the next several years, clearly not at the growth rate that we've seen in the past 1 or 2 years. That's just not sustainable. But, it just speaks to our conviction that this is still a growth area.

Jordan: That's excellent. On the topic of models, Meta released their latest model, the Muse Spark model. In the market, from what I'm reading, it seems to be well-received, but maybe Marcello, Rob, you could give us your thoughts on what this release means for businesses like Meta.

Marcello: Yeah. Again, I was talking about how we kind of have these ecosystems moving around and Meta originally had come out with kind of this Llama family of models that was showing promise initially, but kind of faltered compared to some of the other ones. And Zuckerberg basically overhauled the entire team, did a lot of hiring and acqui hires, has spent immense amounts of money, and has kind of completely reoriented the company around a whole new family of models and the first one that's come out and this is kind of like just the initial model.

There's going to be an entire family of them. This model, in terms of all the benchmarks that are commonly used for measuring the effectiveness and the power of the models is scoring really high. And it’s just kind of behind the Gemini 3.1 and the OpenAI 5.2 or whatever it is, I lose track of some of the versions.

Now there's so many, but it's right in there. It's right within the ballpark that it needs to be. So there's a couple of things to think about here. Number one, it depends what Meta wants to do with these models -  if it's just to really improve their algorithms for serving up ads and things, it probably doesn't need to be as powerful as the latest generation, for example, the Mythos model.

It just needs to be really good at improving Meta's business. But in terms of like getting consumers to start using it, there's no company on Earth that has the distribution that Meta does in terms of Instagram, Facebook, and all their properties. They all kind of have a little window at the top or bottom, depending on how you've got to set up.

That basically gives users the access to this new model, which is proving so far that it's going to be a contender. So again, Meta was being viewed as the company that couldn't get it straight and all of a sudden they're starting to ascend and, as I mentioned earlier, sometimes the stock reflect that.

Over the last couple of days, I think Meta is up just a good 10 to 12%. So we're starting to see it. So anyway, it's looking promising for Meta.

Rob: It would support just like to say to NVIDIA, Meta is either the largest or one of the or their second largest spender of NVIDIA GPU systems. One of the bear arguments was if Meta can't get their superintelligent lab right this year, is there a risk to Meta just really pulling it back on their spend and saying that they can't get to a lead model position?

Rob: And that would have been a threat to, you know, a big portion of NVIDIA revenues. And clearly a data point like this is positive from NVIDIA, it's a lower risk now that that spend is going to be at risk. You know, in the next 12 months.

Jordan: That's great. So maybe we'll shift gears a little bit. We've talked a bit about memory in the last few months; it's been a roller coaster year for these memory stocks. So, Rob, maybe just give us an update. What's kind of happening these days? These stocks have sold off quite a bit over the last few weeks.

Rob: So it really started about the middle of March when Micron reported they had an outstanding quarter and they blew away numbers on their guide for the next quarter. For context, they reported something like $19 billion, the guide was for $19 billion in Q3. Versus the consensus, which is that $11 billion gross margins were up above 80%. That really marked a short term peak for that stock and for other similar memory names.

Any concern that we're sort of peak margins for this cycle, some inkling that there's some spot pricing is starting to flatten out a little bit. And you know what that meant for memory pricing going forward. Then, there was this terrible quant paper that came out, which isn't really new news, but started to put some question marks out to whether there's going to be a destruction, a demand for memory.

And then you throw in the U.S.-Iran war and clearly not a good thing for a commodity type business and or for stocks, weren’t up as strongly as they were going into the event. So it was a confluence of all this that caused a severe downside in these stocks over the past 3 or 4 weeks.

It probably had gotten to a bit of an overshoot because clearly there was some short term spot pricing dynamic, but there is such a severe undersupply in memory that cannot be addressed before at least 2028. So I too was also a bit concerned about, you know, what that pricing dynamic would look like in the next six months.

But clearly the stocks kind of overshot. And we're really in kind of a once in a generation supercycle for memory. And like I said, they can't build capacity. The big three players Micron, Samsung and SK Hynix cannot build enough capacity, at least through the end of 2027 to address where the shortage is today.

I wouldn't expect memory to be down for long. I think the market will start to look through a bit more into sustainability in what might be peak margins, potentially. But the sustainability being much greater than just a quarter, as you typically see in a regular memory cycle.

Jordan: That's great, thanks, Rob. I'm just looking at the time here, I think we are coming up on time, but maybe we can end off with some thoughts on healthcare, particularly Eli Lilly. So last week or the week before that, they received approval for the oral version of Ozempic, the pill format. How important is that for the stock and do you anticipate, call it another leg up in terms of adoption, growth for medication.

Rob: So this was expected, it was actually supposed to come April 10th. That came about a week earlier. The new drug is going to be called Fundo. It will compete against Novo Nordisk, which already has an oral version of a goby on the market that came out early in 2026. So a bit of a, you know, first mover advantage for Novo and the prescription growth week over week has been very strong, well above consensus.

Going into it, Lilly (this is what I would say is expectations for the very near term) have come down a bit, like it's going to take a little bit of time for this to ramp. And it's going up against again, something that's already entrenched. But clearly having an oral on the market and now having two oral orals on the market, it's going to be important to expand the margin at least, around the edges.

There's definitely a portion of the population that is needle phobic to some degree, and don't need the types of weight loss that are associated with higher, more powerful injectable versions of the drug. And they can get to their weight goals with a lower doses that you can get from an oral form.

So, it's open in that part of the market. Internationally, it can be very big because there's much more needle phobia internationally and much more price sensitive market, even internationally. These orals will be able to address that. So, over the long term we see a nice opportunity for orals to become; 25% to 35% of the market is not going to overtake injectables for a lot of reasons that are take longer to explain than we have time for today. But it is a nice incremental for the stock now or a nice incremental for the overall market. I would say from a stock perspective, I think a lot of it is probably already discounted. It's not something that we're kind of rushing out to needing to buy more stock because of this.

But it's a nice check box that the dynamics in the OBC market remain strong. And that's going to be good for both Lilly and Novo. I think my bet is, longer term, Lilly is going to be better positioned because of a wider range of product offerings that they'll have available to address different areas of the market.

So my bet is still that Lilly is going to be better positioned longer term, but the stock is probably reflecting that at this point.

Jordan: I was going to say, is one better than the other?

Rob: At the margin. No; really from a marketing perspective, what Lilly will say or what Lilly will try to use commercially is that there are fashion and food requirements around the Wigo V, the Novo drug that is not present for the Lilly drug. So the drug is a little bit easier to actually administer. And in theory, that makes sense in practice.

Rob: I don't know if it's actually that much of an issue. The drug's going to be comparable efficacy wise. So really a lot of it is going to come down to direct to consumer marketing and really building campaigns around that. But like, for all intents and purposes, they'll split the market. I just think Lilly has a little bit of advantage, because I think the rest of their obesity portfolio is better, and there's some bundling and some packaging they can sort of do around all that.

That makes it a bit more powerful in reimbursed channels and the cash paid channel. Honestly, it's going to be very indistinguishable, in my opinion, between the two. Some of it will just come down to distribution capabilities and execution by each company.

Jordan: Excellent. That's great. Well, thanks, guys. Just looking at the clock here, and maybe this is a good place to wrap up this month's episode. Always appreciate you making the time to meet and speak with us, and looking forward to picking this back up next month. So, for those of you tuning in, if you'd like to learn more about strategies like RBC Life Science and Technology or the RBC Global Technology Fund, the PH&N U.S. Growth Fund – I do encourage you to visit our website, www.rbcgam.com.

Thanks, everyone for tuning in and we'll be back soon.

Rob: Thank you.

Marcello: Thanks. Bye for now.

Jordan: Cheers.