Jordan Wong:

Hi everyone and welcome back to Tech Talk. My name is Jordan Wong. I'm a portfolio analyst on the Investment Products Advisory Group at RBC GAM. And of course, as always, I'm joined by Marcello Montanari and Rob Cavallo, both portfolio managers on our Toronto-based North American Equity team, and responsible for strategies like RBC Life Science Advisor Technology Fund and RBC Global Technology Fund, as well as a number of other growth-oriented investment strategies. Now if you're new to the program, we get together once a month, to talk all things technology and healthcare related. Highly innovative, dynamic and fast-paced sectors. So we think it's just a great way to keep you all up to date and informed on what's going on across these sectors.

Marcello, Rob, welcome back.

All right, so it's kind of funny, right? Like I think,  we've talked about this for a little while but we figured at some point we would get to a state where it would be time to talk about whether or not we are in a tech-related, AI-fueled bubble similar to that of the dot-com era in the late 1990s and early 2000s. So that's really where I want to kind of focus today's episode on. Of course, if we have time at the end, I think it'd be great to touch on recent antitrust settlement between Apple and Google and what the implications are for both of those businesses. But maybe if we kind of start on this AI bubble related topic. There have been growing calls from investors, economists, business people that we are maybe entering a bit of a tech-related bubble. I think, based off of a lot of what I'm reading, a lot of comparisons made to the dot-com bubble in terms of really elevated valuations, a high degree of market concentration across the top 10 stocks in the S&P 500 Index. At the same time though, we're seeing a tremendous commitment to captial expenditures (CapEx), revenue growth, and earnings growth that far outpaces what we saw during the dot-com bubble. And so maybe to kick us off, walk us through what you're thinking in terms of, whether or not we are in a bubbles, what are some of the factors that you're looking at? And if you can compare and contrast where we are today and where we were in the early 2000s, I think that would help us better understand where we're sitting at.

Marcello Montanari:

Thanks for having us. Maybe I’ll start. Are we actually going to do this today? It’s like today, I guess, because this will be posted later. Today’s what, September 10th? Oracle last night just reported that they have signed contract values totaling $455 billion. That’s billion with a B. Microsoft just signed a deal with a small kind of NEO cloud called Nebius for $10 billion. TSMC, I think, just put up some pretty good monthly numbers, Rob, if I’m not mistaken.

So, I mean, the technology seems to be moving forward. I know that Sam Altman was on a news program mentioning that he thought we might be in a bubble. I think it got misconstrued as “we are in a bubble.” He came back later and made some clarifying comments, saying he was really more focused on the exuberance of markets, which is important to us, obviously. But when he brought it back, he said the technology itself is not a bubble. He called it probably the biggest thing that’s happened in his life, technology-wise. It’s definitely on par with, if not exceeding, the development of the semiconductor, the Internet, and mobile—if not surpassing them.

I was around during the first tech bubble of the Internet, and what I can say is that this is very different. Coming back to Altman’s point, the technology itself is not in a bubble, but there is some exuberance. The tech is very real, and it’s still early. The usage appears to be exploding. For example, Sundar Pichai from Google mentioned on their last conference call that in May, they measured usage by the number of tokens processed. Google used 480 trillion tokens in May, and by the end of July, that number had jumped to 980 trillion tokens per month. So, you can see the growth.

If you follow the sector at all and listen to the people deeply involved in it, every time capacity comes online, it gets used up. This usage is obviously fueling CapEx. I’m sure Rob will have some points on this, but CapEx is being raised every quarter. For example, Oracle raised its CapEx by at least another $7 or $8 billion for this year. The CapEx is real.

What is different about this period versus the tech bubble of 2000 is that the CapEx is underpinned by companies with real revenues. These are the biggest companies on Earth with real revenues, real earnings, and real cash flow. They’re the ones predominantly driving all of this, at least at the moment. That’s in contrast to the 2000 bubble, where many new companies were funded by the market but didn’t have self-sustaining businesses. Today, the funding is in very strong hands. To me, it looks like we’re still in the early innings of this. I don’t see any signs that this will change, though Rob might add some additional colour on that.

On the valuation issue—having lived through the 2000 bubble—it’s very different today. For sure, valuations in some areas are getting stretched. But before I dive into that, let me address something else. When we talk about a bubble or exuberance, what tends to happen in a new ecosystem is a process coined by Michael Mauboussin as "fill and kill." Essentially, you get speciation, where all these companies—or species—put their hands up and say, “Hey, I’m an AI player.” Many of these are just pretenders. They’re not going to play a meaningful role, but they take advantage of the exuberance, get recognized as AI players, and attract capital. Some of these stocks even become meme stocks and have big runs.

I want to differentiate here—that’s not the game we’re playing. We’re focused on the bigger names that we believe have a true claim on this technology cycle. Eventually, these "fill and kill" cycles will lead to a culling. The market won’t support all of these players. So, is there a bubble emerging on the fringes of this ecosystem? I’d say yes, probably. But when it comes to the biggest, central players, I think we’re still okay.

Now, coming back to valuations: for some of these fringe players, I haven’t really looked at them because, to me, they just don’t make sense. The valuations are probably nonsensical. But for the names we’re focused on, while valuations have stretched a bit, they’re not crazy. They’re certainly not like the 2000 tech bubble. They’re based on real metrics—real earnings, real cash flow, and EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization). For younger companies, they might just be based on revenue, but at least there’s something tangible there.

Back in the 2000 bubble, companies were being valued on things like price per engineer, price per PP&E, price per fiber route mile, or price per page view. It was all these crazy metrics that made no sense. It’s a very different situation now.

So, yeah, on valuations, I’ll just leave it there for now. I’m sure Rob will have some thoughts to add as well.

 Jordan Wong:

And sorry, maybe just before we move on to Rob. Earlier you'd mentioned CapEx. Are there concerns at all on the ROI-related to CapEx? We're reading that CapEx is totally close to $600 billion. AI-related revenue is still a very small fraction of that. Is this just something that we expect to pay a dividend later on?

Marcello Montanari:

Yeah, I mean, for the most part, there are some companies that are already using AI and machine learning to improve their operations. These are the usual suspects—the companies spending the most money. Whether it’s Meta or Google, they’re using AI to drive all their search algorithms and everything that serves up ads at the right time to the right person. All of that is being powered by AI and machine learning. These companies are already starting to see results from this.

Microsoft and Amazon Web Services (AWS) are also leveraging AI. For example, AWS is using it to offer services, and they’re starting to see an upward inflection in revenues just in the last quarter. For some of these names, the spending is quite significant, and the cost to create these models—especially at the frontier—has been quite large.

It’s important to note that this takes time. If you build a dam to generate power, you can’t expect an ROI before the dam is fully operational. I kind of liken it to something like that. In many cases, it’s a "build it now, and they will come" sort of scenario. I hate to use that phrase, but there is an expectation that businesses will eventually be built off of this technology.

So, to expect an ROI today for many of these efforts is a little misplaced.

Jordan Wong:

All right. No, that's great. And Rob, maybe just jumping over to you. Kind of related to this AI bubble question. You cover the semis. We just had quarterly earnings from Nvidia and Broadcom, two key players in the AI infrastructure build out. There were some pretty significant takeaways from these earning releases, particularly comments around committing $3 to $4 trillion in CapEx by 2030. So maybe just walk us through what you took away from both Nvidia and Broadcom earnings.

Rob Cavallo:

Yeah, thanks. While I answer that question, I’ll actually add a few points to the prior question about the bubble. For context, Sam Altman made those comments, and now it’s rumoured that they’ve actually increased their cash burn profile by $80 billion over the next four years. There are projections of about $35 billion in cash burn in 2027 and $45 billion in 2028. So, there’s a little bit of self-serving talk in what he’s saying.

But I also think it’s worth taking the lesson from 2000. Back then, there was a tech bubble and a telecom bubble. The tech bubble you can liken to the meme or metaverse bubble of 2021-2022. It didn’t have huge tentacles across the entire market—maybe it impacted the venture ecosystem, but not the broader market. The bigger problem in 2000 was the broadband and telecom bubble, which is more like the CapEx bubble people are talking about today. Big difference.

Now, you never say never—you never know you’re in a bubble until after the fact. So, take this with a grain of salt. But if you look at Cisco in 2000, it was the poster child of that era, trading well above 100 times forward multiples. The market was essentially saying there was a very long tail of growth ahead, with no foreseeable problems for Cisco’s business. Today, you look at Nvidia. It’s trading at about 40 times calendar 2025 earnings, and around 27 or 28 times next year’s earnings. Despite that, there’s still a significant wall of worry about 2027 and future spending.

I think the market is okay with what 2026 spending will look like from a CapEx perspective. But for Nvidia, Broadcom, and others, there’s still this wall of worry. The evidence does suggest that while the growth rate is decelerating, the absolute dollars being spent are still increasing. Based on what we can see for next year, and probably 2027-2028, spending is still going higher. This is especially clear when you listen to comments from Nvidia, Broadcom, and Oracle, as you mentioned last night.

So, while there’s a wall of worry and different valuation levels, this may not be a bubble today. But if it is, it could get much crazier before we hit that first air pocket. Based on what we’re seeing today—product cycles, management team comments, and piecing together other industry information—we don’t see a break in spending yet for 2026 or 2027. Beyond that, visibility becomes more difficult, but for now, there’s no clear sign of a break in the CapEx cycle.

For example, Broadcom’s CEO just signed on to stay through the end of the decade, with his pay package completely tied to AI revenue. Broadcom is expected to generate around $20 billion in revenue this year, and more than $40 billion by 2026. His payout is based on Broadcom hitting $90 billion in revenue—more than doubling in three fiscal years, with the high end being close to a quadrupling of revenue, from $40 billion to $120 billion. So, this could grow much larger before we see any sort of downturn in the spending environment.

This brings us back to your ROI question. If the ROI profile is there in three to five years, and this truly is transformational—both in terms of cost savings and new vertical or market expansion—spending could continue much longer than many think. It could outlast the predictions of naysayers and doomsday theorists.

We need to be smart and selective, staying away from areas that are more susceptible to competition or where the margin profile may never catch up to the revenue profile. That’s where we’re trying to be diligent and responsible. But as of now, we still see growth in spending for at least the next couple of years. Beyond that, we’ll have to wait and see and keep updating our view.

Jordan Wong:

Maybe just pivoting ever so slightly. I do want to talk about this Google/Apple antitrust court ruling. Maybe just walk us through the verdict there and what it means for both of these businesses since the verdict released? Both stocks shot up. Google's continued an upward trend. Apple's kind of fallen back down. So, maybe what's going on between these two and how are you feeling about this court ruling?

Marcello Montanari:

It’s more about the remedy stage here. The verdict had already been rendered that Google was running a kind of monopoly. So, this was the remedy stage, and you can look at the stock prices to see how it’s been interpreted. Essentially, as one person put it, this was expected to be a clearing event. There were some pretty draconian outcomes within the cone of uncertainty that had given people pause. Ultimately, the remedy was viewed as more of a slap on the wrist. As someone said, “a slap on the wrist with a feather.”

To summarize, Google doesn’t have to sell Chrome or divest from Android. Apple is allowed to sell access through their operating system, though the contracts have been whittled down to one-year terms. For now, it looks like Google will continue to hold onto that position. As a result, a significant headwind has been removed from the entire sector.

For Apple, looking ahead, they’re probably in a good position. At some point, they may be able to use their leverage to negotiate better terms with other search engines, whether it’s Google or someone else. That’s more of a second-derivative effect, though. At the moment, I don’t think anyone is in a position to displace Google. The amount of money Google pays Apple is just too significant. While I could be wrong, I don’t see any player displacing them in the short term.

As I’ve said to others, there’s a difference between the desire to take this business and the ability to do so. Google has all the infrastructure to carry this traffic and process it. As AI becomes more integrated, the processing demands are only increasing. That’s likely why, as I mentioned earlier, the token count at Google is exploding. There aren’t many players capable of handling this type of risk and scale. We’ll see how this plays out, but for now, Apple retains its gatekeeper status and can charge for it. Google, in turn, will continue with one-year rolling contracts. While they could be displaced at some point, that doesn’t seem imminent.

What’s more interesting are the second-derivative effects. Other players will likely start repositioning themselves in this new landscape. For example, it seems probable that OpenAI will eventually launch an advertising-based approach. They clearly need more revenue, and search and advertising are natural targets. OpenAI recently hired a key person from Google, so something is clearly in the works. That could be a bad headline for Google in the future.

Meanwhile, other players like Bing, OpenAI, and Anthropic are all likely to reposition themselves in this evolving environment. At some point, Google could lose some of its position within Apple. However, there was a recent article in Bloomberg suggesting Apple and Google are discussing a potential collaboration, possibly involving Gemini. While it might not be called Gemini, the idea is that Google could become the AI engine for Apple. This is because Apple’s own AI efforts appear to be faltering, marred by controversy and underwhelming results. Apple is in a great position as the gatekeeper, but their AI advancements haven’t been remarkable so far.

So, we’ll see how this all unfolds. The second-derivative effects will be particularly interesting to watch. For now, Google’s stock has continued to rise, while Apple’s has been coming off. That’s likely due to Apple’s recent unveiling of their new phones, which were a bit underwhelming and seem to have weighed on the stock.

Jordan Wong:

That's great. Rob, anything from you before we close out the episode?

Rob Cavallo:

No, I mean, I think you touched on it. There are obviously rumours about Apple pursuing some sort of big M&A to get themselves into the large language model game. But this situation clearly changes things and solidifies the idea that Apple can continue to provide the network or ecosystem while letting others compete within it.

You also mentioned Google and Apple potentially working together from an AI perspective. I think it’s important to flag the rumours around Siri possibly being powered by Google Gemini. If true, this could represent a significant upside opportunity for Google, especially now that this situation seems to be moving toward some sort of finality. 

Jordan Wong:

Well, thanks guys. That’s all I wanted to cover off today. Maybe we’ll end the episode off here. As always, really appreciate you sharing your insights with me, and all the people that tune into this program. For those of you watching, if you’d like to learn more about the strategies that Marcello and Rob manage, strategies like RBC Life Science & Technology Fund and RBC Global Technology Fund, it’s best to visit our website rbcgam.com. I will mention before we sign off, we are now on YouTube. So all of the episode we’re recording will be posted to the RBC YouTube page. We’ll make sure to put a link somewhere in the description down below in case that’s an easier way for you to stay up to speed and stay current with the episodes.