Q3 '25 Asset Allocation Viewpoints Webinar: AI Momentum and Tactical Asset Allocation

Christina Noonan

All right. Hello everyone, and thank you for joining us for AI Momentum and Tactical Asset Allocation, a quarterly Asset Allocation Viewpoints webcast brought to you by T. Rowe Price's Multi-Asset Division. I'm Christina Noonan, a portfolio manager within Multi- Asset, and I'll be your host for today's discussion. Today we're going to talk about recent trade policy action, market sentiment, and, of course, implications of AI.

But before we begin, a couple housekeeping items to note. In your video player, please use the right-side panel to submit questions and take our post-event survey. On the left, you'll find more information on Global Asset Allocation Viewpoints, and at the top of your screen, you'll find the link to register for our next webcast in October. Slides will be available to download with the replay of today's session as well.

If you joined us in April, you'll recognize that we are filming from our new office in Baltimore's Harbor Point. Our new studio is not quite ready yet, so today's session is recorded in advance, but we look forward to coming to you live again soon.

So with that, with me, as always, is Sébastien Page, who you may recognize from Bloomberg, CNBC, and LinkedIn. He is the chief investment officer and head of Global Multi-Asset, overseeing an investment team dedicated to multi-asset portfolios. He is a member of the Asset Allocation Committee responsible for tactical investment decisions and an author of two books, “Beyond Diversification” and, most recently, “The Psychology of Leadership.” Thank you, as always, for being with us today, Sébastien.

Sébastien Page

OK, Christina, I have to say—it's not because he's here—Dom Rizzo is one of my favorite guests on this webinar. And I prepared something. He doesn't know anything about this.

Dom Rizzo

Oh man, I do not know anything about this.

Sébastien Page

I went—just because we're going to talk about AI today—I went on Perplexity and I asked, “What would Dom Rizzo say about AI and capital markets today?” And I have a paragraph of what AI, based on anything about Dom Rizzo out there…

Dom Rizzo

Yeah, I bet you it’s pretty good.

Sébastien Page

…thinks he’s going to say. Now, I'm not going to say what it is. I'm just going to check at the end if it matches.

Dom Rizzo

Yeah, I bet you it is.

Sébastien Page

This will illustrate the power of AI, and maybe it’ll illustrate if Dom Rizzo is replaceable. I’m just kidding, you are un-replaceable.

Dom Rizzo

Yeah. No, it's replaceable based on my past thoughts, but then the question is my future thoughts. That's where I have to come in. So.

Sébastien Page

Right, right.

Christina Noonan

OK, well, this sounds exciting. So our special guest today, Dom Rizzo. He is the portfolio manager of the Global Technology Equity Strategy and has been with T. Rowe Price since 2015.

Dom Rizzo

Ten years now. It's crazy.

Christina Noonan

And you last joined this webinar two summers ago, so thank you so much for coming back and sharing your thoughts.

Dom Rizzo

Thank you for having me. It’s great to see everybody again.

Christina Noonan

Okay. Of course. All right. So as I mentioned, we have a lot to cover today, from AI to market data to trade discussions. So in this session, we're going to touch on how we're navigating this environment and our thoughts on tactical asset allocation. So with that, let's start with you, Sébastien. Quite a lot has changed since we've met last, especially around trade. We saw markets sell off post-Liberation Day, only to rebound sharply back almost to all-time highs. How are you and the Asset Allocation Committee thinking about your risk positioning today?

Sébastien Page

And what a round trip it's been. You know how I start my day every day? I exercise, and I put on headphones, and I listen to the financial media, usually Bloomberg or CNBC.

And I got to say, like, what do you think is the number one theme that I hear every day constantly this year? Of course, it's tariffs. And I have a bit of tariff fatigue, if you will. I'm going to make the point for a moment—and this might sound contrarian—that the importance of tariffs is vastly overstated.

Look, they matter, but let me make the point: The total value of all imported goods subject to tariffs represents about 9% of GDP. And sometimes the market trades like tariffs drive the entire U.S. economy. But what about the other 91% of the economy? Service—OK, I understand there are some indirect effects—but still services consumptions drives about 45% of GDP and about 70%, seven zero, of GDP growth.

I don't know about you, but like when I get my haircut, there's no tariff involved. So, services drives 70% of the GDP growth. It's not just about tariffs, even though every morning in my head, my headphones, it's all about tariffs. Why don't we hear more about the health of the consumer? Unemployment is low. Household net worth is at an all-time high.

Wages are growing faster than inflation. That's a good thing. So, if wages are growing at 4% and inflation's growing at 3%—-right now, 2.5%—that is supportive for the consumer. Consumer spending is expected to grow slower than last year, but still close to 4% this year. So if you think about tariffs, right, even at 10%–15% of imported goods, they may slow down the growth, but they don't have to cause a recession is my point. And we ought to talk about something else than tariffs, right? And today we're going to talk mostly about AI.

So they matter, but their importance is often overstated. Recession risks are low. What would be the two most important factors for markets right now. It's more traditional stuff: valuations and earnings. Last quarter, earnings were up 13% year over year. That's very good, and we could get another 9%–10% according to consensus over the next 12 months. But that earnings growth, it kind of looks priced in if you compare it to the valuation. So that's the conundrum we're in. When valuations are expensive, there is upside in markets. If your horizon is six to 12 months, valuations don't matter as much as you might think.

And with stocks making all-time highs, you often get upside still. However, there's also higher—there’s a technical term—blowup risk. There's more fragility when valuations are above, say, 20 on the S&P 500 and forward P/E ratio. So in the Asset Allocation Committee, we still remain close to neutral on stocks versus bonds. We actually have a small underweight stocks, but close to neutral.

And we take a more diversified approach to stocks because market concentration is at an all-time high. Under the hood, we’re long value stocks, we’re long international stocks. So that's how we're playing this. But tariff matters. They matter, but their importance is overstated is my point.

Christina Noonan

I'd say you sound pretty optimistic for being close to neutral. Are there maybe signs or levels in markets where you would maybe look to step back in, if values like valuations? Maybe if we see sell-off?

Sébastien Page

Yeah, this is a good question, right? When you're concerned about valuations but you're generally optimistic that you're not getting a recession. And there are positives. I talked about the consumer. We'll talk about other positive factors.

Dom Rizzo

AI…

Sébastien Page

AI being one of those.

Dom Rizzo

We read each other's minds, Sébastien!

Sébastien Page

So, what do you do?

I mean, things look expensive. I think you do two separate things. One is you broaden the portfolio because not everything's expensive when you look at stocks. When you look at value stocks, non-U.S. stocks, despite the rally, still look good. So you diversify. You broaden the portfolio. And yeah, I think all else equal, it's OK to wait for a pullback to go back in.

I don't want to give you a percentage, like at 10% we’ll add more stocks, but a significant pullback, and it'll depend on what's happening at the time. But I still think buying the dip sounds simplistic, but it's worked. And if we don't get a recession and if there's a sentiment-driven geopolitical headline-driven panic, it might be another opportunity to actually add, not subtract, to risk.

Christina Noonan

OK. Makes sense. So, let's hear from Dom. We touched on AI, a huge driver of markets so far this year, past few years as well. AI has the potential to reshape macroeconomic dynamics in ways we haven't seen before, particularly through influence on labor markets, potential for deflation. And you've described AI as the biggest productivity enhancer since electricity. Can you expand on those productivity gains that you foresee?

Dom Rizzo

Well, first off, I have to say thank you to the Multi-Asset team because I always said AI has the potential to be the biggest productivity enhancer for the global economy since electricity. It's a mouthful, but we get there in the end. And someone said we have to we have to look at the numbers here, Dom. You can't say this without numbers to back it up.

And so if you go look at the numbers, electricity added roughly 1% a year to global GDP via productivity for 32 years. So if you think about GDP, there's three parts that really make it up, right? There's capital, there's labor, and there's productivity. Productivity is how much can you do without adding more people, basically. And 1% a year for 32 years actually looks pedestrian compared to what I think AI can do.

And so you can actually go look at the productivity statistics now, right? And go look at the productivity statistics from the day ChatGPT launched to today. We're running above trend. Productivity normally would add collectively 3%–4% to GDP over that time period. We’re north of 4%, closer to 5% now. So we're already looking at that kind of 1%-plus range because of, incrementally because of AI.

And when I think about it personally, right? I mean, Sébastien knows this. I have two young kids at home. I have a two-year-old and a four-month year old. And the amount of time that I use ChatGPT to check on: Am I doing this right? Am I doing that right? Can we improve potty training. All these things.

It has certainly added productivity to our household, but it's added so much productivity to my work as well. And so when I look at the numbers, I say 1% for 32 years, that actually will be quite pedestrian in comparison with what AI can do. Not only that, I think AI is delivering this to us at such a reasonable cost.

So let's think about the pro version of ChatGPT is $200 a month. But it may make an individual worker 20%, 30% more productive. So what is that worth to a company? Tens of thousands, hundreds of thousands of dollars potentially if you can get that person doing much more in less periods of time. And so the consumer surplus from ChatGPT and AI tools like it are really something like we haven't seen before.

Now there's two dynamics on the macro side of AI that are always worth discussing as well. And we didn't really bring up tariffs and what that means for inflation, but that's the other thing. Everyone says tariffs, and then what does that mean for inflation, right? And I think that's why tariffs get that airtime that it does. But AI is both inflationary and deflationary. And what do I mean by that? In the short term, AI capex is inflationary. It's just math. We're spending all this money. We're building out power, we’re building out data centers, we're trying to get chips, supply/demand are not in equilibrium, right? Everyone's buying. Everyone's buying. That's inflationary.

Over the medium to long term, as the productivity starts to kick in, that's deflationary for the economy. Tech is a natural deflationary force on the global economy, and I think AI will be like that over time.

So those are the different macro dynamics that I'm thinking in. But overall I'd say compared with two years ago when we came on and we talked about AI, I am more convinced than ever that AI has the potential to transform the global economy, like electricity, the railroads, the internet. But actually, I'm becoming more convinced it may be more powerful than all those different trends.

Sébastien Page

So here's what Perplexity thought you would say.

Dom Rizzo

Oh, yeah. I want to hear.

Sébastien Page

It’s…maybe I didn’t ask the question right, but it's actually pretty close. So what would Dom Rizzo say today about markets and AI? And it responded—this is as of this recording just an hour ago. I said, ”Assume you're Dom Rizzo.” I'd emphasize—so this is Perplexity speaking as if it's Dom Rizzo.

Dom Rizzo

Yeah.

“I’d emphasize we're still in the early innings of a long-term productivity boom, with AI use cases rapidly expanding across sectors.” So, there's more. But it sounds like you, let’s just say.

Dom Rizzo

I think it's, I think these, I mean…. Frankly speaking, right, if you're doing a webinar like this and not using AI to prepare, I actually don't think you're doing your job, right? You need to use AI to sort through all this information, summarize what you're going to say, make it better. I mean, I put everything we said today through Perplexity.

Christina Noonan

I thought the version you sent looked better.

Sébastien Page

Now, let me get very meta here. The next question I asked Perplexity is, How will Dom react when he hears this?”

Dom Rizzo

Wow. How well does it know me?

Sébastien Page

It  predicted that you would say that's spot on.

Dom Rizzo

I don't think I quite said that, but that's pretty good. And it makes me feel better about the private investment that T. Rowe Price did in Perplexity as a result.

Sébastien Page

Look, this brings up the issue of AI uses existing information. And we're in the prediction business at the end of the day. So it's a prediction tool. But you, as an analyst, a portfolio manager, you have a different value add than just meshing together existing information, right? In other words, I can't replace you.

Dom Rizzo

Yet.

Sébastien Page

On this webinar with Perplexity.

Dom Rizzo

Not yet. No. Look so just for the everyone listening: How does AI work?

We'll just do 10 seconds on it. Sébastien's heard me say this a million times, but it's worth going through it. AI is a parallel computing technology. Most technology is serial, which means you do it one at a time, OK? And so if you gave a serial processor a CPU from the likes of Intel, or someone like that, the task of reading “A Tale of Two Cities” by Charles Dickens and said, hey, predict the next word that Charles Dickens would say in the book or count how many “the”s Charles Dickens would say in the book. Let's use the counting example. It would just count through, “It was the best at times. It was the worst of times.” And it would go count every sentence. The parallel processor, the GPU from NVIDIA or AMD, would rip the book into a thousand pages and pull out all of the “the”s simultaneously.

So, what Sébastien is saying is, how do we get the answer for this perplexity? We went and we studied the whole history of the internet, and we ripped it up into a thousand pages. And we're predicting what I'm going to say based on everything that we used to say, right? It's all stuff in the past. Now, the question is: With these new tools, these new reasoning tools, right—the old 3 model from open AI and taking more time to think about the answer—can AI actually come up with new ideas? Can it solve new physics problems, right? I think we're getting closer to that stage where it's not just a prediction tool but it's actually a tool that comes up with—

Sébastien Page

It's creative.

Dom Rizzo

It's creative in nature. It's not just generating based on past, but it's, but it's reasoning to new theories and potentially things like new physics, right? So this is where that reasoning technology is so, so powerful. And the great thing from my perspective, as the global technology strategy manager, is reasoning is so compute intensive. This just means even more GPUs.

That's all. Reasoning is so compute intensive. Orders of magnitude to compared to what they would call single shot. Think about the difference between using ChatGPT at the beginning with ChatGPT 3 for three and a half or 4 versus the old 3 model, which is the one that thinks.

Sébastien Page

More GPUs and more power.

Dom Rizzo

More GPUs and more power. Those are the two.

Christina Noonan

Extremely interesting insights, and we'll continue to talk more about AI, but did want to come back really quick to trade. We've received a lot of questions about this ahead of time and obviously been in the headlines as Sébastien touched on, but you sound a bit sanguine on trade. Can you elaborate a little more on what makes you less concerned and maybe why you think we have reached the peak uncertainty on trade?

Sébastien Page

As we're recording this, I'm starting to doubt whether we've actually reached the peak uncertainty. But what's interesting is the markets are just less reactive to trade. Earlier I said that the value of imported goods subject to tariffs represents about 9% of GDP. Of course, there are indirect effects and supply chain effects. But mathematically, that's 91% of the economy not directly affected by tariffs.

And remember we're getting 10%–15% tariffs on that 9% of GDP. So, the projections right now are for $300 billion in tariff revenues, so that's 1% of the $30 trillion GDP. So, they matter. And maybe it's controversial to say that they don't matter as much as you hear in the media every day. But recently markets do appear to be moving away from extreme tariff sensitivity.

The S&P 500 has rallied 25% since the April 8 bottom. Valuations are back into expensive territory. What happened? A few things. Yes, maybe we have reached stop-getting-worse moment with tariffs. We'll see. Second, there's no sign of austerity. For a moment, it looked like we were going to really cut back deficits. It doesn't look like this. That has its own issues for yields and for inflation. But from the market's perspective right now, no sign of austerity, Big Beautiful Bill and so on. Recent inflation prints have been soft. That's also good—keeps the Fed engaged into cutting. And earnings have shown resilience. So, I'll just say this. Maybe the world is not falling apart after all.

Christina Noonan

OK. Makes sense. And let's hear from you, Dom. How is trade impacting technology? What are you watching closely as the trade tariff landscape continues to evolve?

Dom Rizzo

Well, what is the trade war? In a way, it’s really a reordering of the geopolitical world order. If you take a step back, right, and you study rising economies and hegemony, you have to control your supply of essential goods.

And in technology, where what is the supply of essential goods that probably matters the most? It's semiconductors, right? All roads lead to semiconductors. They're strategically vital for national security. They're the backbone of all this AI revolution. And, oh, yeah, we manufacture 90% of our leading edge in Asia. That's, geopolitically, not sustainable, right? And so when I look at all the different trade policies and tariffs, I look at what's probably going to happen is a major reshoring of the semiconductor industry, particularly at the leading edge.

And we're already seeing signs of it.

So TSMC has really doubled down on their Arizona production facility. And they manufacture the leading-edge chips for every fabless semiconductor company—AMD, Qualcomm, you name it, Apple. Apple is the biggest fabless semiconductor company in many ways, right? They manufacture all of Apple's chips. All of those companies want production of leading-edge silicon in Arizona.

NVIDIA, on their last call, Jensen talked about being able to build an entire supercomputer in the Americas in 2025. It's amazing. Right? And by the end of 2026, you should totally, totally see that happening.

Sébastien Page

Including the rare earths?

Dom Rizzo

Rare earths.

Sébastien Page

Do you need rare earths?

Dom Rizzo

Well, you need rare earths for everything. The thing about rare earths is they're actually not that rare.

It's the refining of rare earths that's very rare. And so I think what you've seen with recent transactions by the U.S. government is really a dedication to doing the refining of some of the rare earths in the Americas. But that's an area that going to take the longest. When I look at the entire supply chain of what do you have to move to build a full supercomputer in the United States, that’s going to take some time.

When I take a step back and I say, wow, have tariffs really impacted AI demand? The answer is absolutely not. We saw last quarter was an incredible quarter because that was the quarter that the trade restrictions really came in and hit some of the U.S. semiconductor companies in full force.

Hey, you can't sell these type of chips to China anymore. NVIDIA took almost a $5 billion write-off last quarter. AMD took almost $1 billion write-off last quarter, and they both still beat and raise. There's been no company that hasn't taken a $5 billion write-off in U.S. American corporate history and hasn't gone bankrupt, according to Jensen.

This is what Jensen told me on our callback. I'm going to assume he's right. If anyone wants to fact check him, they can. Last thing I'd say is, have you looked at that Big Beautiful Bill? What does that have in it? It has a lot of incentive for building, right? 100% bonus depreciation. A lot of things on the regulatory side.

So when I look at the combination of tariffs and the Big Beautiful Bill, I say, okay, we're probably going to see an even faster acceleration of some of this AI capex. And now that we have some level of certainty, maybe we start to see that flow through into the industrial economy, into things like autos. PMIs have been below 50 forever. Maybe we see those start to come back too. And so you have the AI remaining really strong and the industrial economy potentially having a comeback in the same time.

Christina Noonan

That's perfect segue; I was going to head to macro economy next. Certainly tariffs, Big Beautiful Bill impacting, but also a lot of crosscurrents on the economy. We're seeing slowing growth, decent labor market, tame inflation but potential for inflection higher, because how are you and the AAC thinking about all those crosscurrents on the macroeconomy and the forces that may be coming or already hitting the economy?

Sébastien Page

It's interesting on growth because we saw growth expectations crater in April, but they're actually coming back up right now. And the Atlanta Fed GDP now looks pretty good. So, it's even hard to make this statement. I think growth is slowing, but it's not a statement, a pound-the-table kind of statement. And we could re-inflect back up with fiscal spending and animal spirits sooner than later.

So our strategy focuses on the interplay between all these factors, plus what we think is priced in the markets. As I said, we're close to neutral on risk and  we're long diversification. We're fading market concentration, and we're positioned for upside risk in inflation. You know, Christina, the second thing I hear the most in my headphones when I exercise every morning, listening to financial media? What do you think it is?

So, tariffs is number one; too much tariff talk or too much talk as if tariffs drove the entire economy. What's the second thing? Fed talk. Fed talk. Yeah. All right. That's also I think right now and overrated factor for markets. I'm trying to be controversial here. Look, we had 550 basis points of Fed hikes, and house prices went up 12%. 550 basis points of Fed hikes, and house prices went up 12%. Rate sensitivity in housing is low because a majority of homeowners still have a long-term mortgage below 4%. Where's your mortgage?

Dom Rizzo

I rent. I arb the difference from my landlord. My landlord, I think, has got like a 1 or 2%. So I arb the difference.

Sébastien Page

And he's passing you a decent one.

Dom Rizzo

Oh, yeah. Oh, we have a great deal.

Sébastien Page

So now we're at 4.5% fed funds rate. Everyone's saying the Fed is restrictive. The Fed is saying we're restrictive. But financial conditions are about as loose as they get in my mind. Stocks near all-time high, credit spreads near all-time tight. U.S. dollar down. Why? We've had massive fiscal stimulus. We're running recession levels of deficits, close to 7% of GDP.

So, I hear tariffs, I hear the Fed. But fiscal maybe right now matters more than the Fed. The Big Beautiful Bill is stimulative, especially given something you mentioned, Dom, which is now companies are going to be able to expense capital spending as well as new buildings. Like, if you're building a factory, you can expense all of it 100% right away, which this will spur investments hopefully and lead to lower tax bills for those corporations as well. So I think the Big Beautiful Bill is stimulative, but there's a cap. There's a ceiling to how stimulative it can be, and the governor to me is the 10-year yield. So, above 4.9% say financial conditions start to tighten and that leads to a pullback in investments and consumption. So we have a stimulative bill.

A lot has been said about we're just extending tax cuts. But I think no lower taxes on tips and overtime. And so I think it is stimulative, the 100% expensing of capex. So, bottom line is I think whether we get zero, one, or two cuts this year, it's not that important for S&P 500 earnings. AI is more important for S&P 500 earnings, for example. So, tariffs are not driving the entire economy.

So for advisors in our audience, next time your investors panic because there's a headline about tariffs going up, just remind them of this. And also the Fed is not driving the entire economy. Both matter, but fiscal policy, AI investments and productivity gains, consumer spending still solid, and many other factors can swamp the impact of these two most favored themes of the financial media.

As for markets, expectations, revaluations, and earnings ultimately matter most. Granted, all these things are connected, but those are what we need to look at. Those are the factors we need to look at. And one trade I like right now in the portfolio is hedged equity. So you can buy exposure or get exposure to hedged equity strategy. And, to me, when valuation is high, as I said earlier, you still have upside over six to 12 months. It's not like the world is going to fall apart because stocks are near all-time highs, but you also have higher blowup risk. So I'd be willing in a part of the portfolio to get exposure to hedged equity, because then I can get maybe 80% of the upside but half of the tail risk. Cut the blowup risk in half.

And if there is a huge pullback in stocks, then you can buy and you just cushion the blow. So at high valuation levels where I don't expect a recession, I kind of like this idea of having a hedged equity exposure.

Christina Noonan

Yeah, I think that makes a lot of sense in current environment. You mentioned the Fed cuts maybe not mattering as much, but what about—you probably hear it in your headphones when you're working out—but the Fed independence? That seems to be a key topic. Do you think that is maybe a black swan or key risk that markets may be underpricing? We've seen a few headlines recently of threats that have been pulled back, but it seems like it's still a risk.

Sébastien Page

It's a risk. I would say it's a very low probability, despite all the talk. We're close to the end of Powell's term. And President Trump has said several times that he doesn't want to fire Jerome Powell. But sometimes you, especially as we record this, you hear different things, and that happens a lot in the media these days. You hear one thing from one source and another thing from another.

I'm not too worried about this. However, it's not a zero probability, and it is very important for the central bank to remain independent. If you look at countries and time periods where central banks were not independent, it didn't end well, usually in hyperinflation. And so we don't want to get there. I don't think we'll get there.

Christina Noonan

OK. All right. Makes sense. All right, let’s come back to AI. So, Dom, as the leader of T. Rowe Price’s technology portfolio, broad question, but how are you navigating this seismic shift in technology?

Dom Rizzo

How do we navigate bubbles responsibly? That's what I always tell the team, right? That's my job, right? So, AI has the potential to be the biggest productivity enhancer for the global economy since electricity. Productivity-enhancing technologies come with speculative bubbles. Our job on the team is to navigate speculative bubbles responsibly. And what does that mean? That doesn't mean sitting on the sidelines the whole time. That means trying to capture the upside and blunt the downside. Look for historical analogies that can help us guide through the way. Use valuation as a barometer.

But just because some stocks look relatively expensive doesn't mean they can't be more expensive over time and have imagination and creativity in trying to create a strategy that can capture the upside and blunt the downside, right? That's what that's what we try to do. That's what we try to do in trying to find idiosyncratic ideas that follow our framework to do it.

Look, I don't think it's a surprise. I remain extremely bullish on the AI ecosystem and all that is going on there. And if you think about the best analogy, the best analogy in recent financial market history is clearly the internet, right? That is the best analogy for what's happening. And so what I would encourage all of our listeners to do is go look up the chart of the Nasdaq from the launch of Netscape through the whole bubble.

OK? December 1994 is the launch of Netscape. November 30, 2022, is the launch of ChatGPT. Put them on top of each other. Tell me what that chart looks like. And it's amazing how much they sit on top of each other. And that would basically tell us, we use that as our guide, we may actually be in 1997 or 1998 from a potential performance perspective. Now, so many different things happened in 1999, so it's not a perfect analogy, right? Just talking about interest rates. Interest rates were lowered by the Fed at that period. Like, how much do Feds get? How much does the Fed actually lower interest rates throughout that time? So, many different things. But the chart of the Nasdaq of the launch of Netscape through 1999 and relative to the chart of ChatGPT through today would kind of imply we're in that 1997, 1998 time frame.

Sébastien Page

But, Dom, those companies of today that are involved in AI have more scale and better cash flow generation.

Dom Rizzo

They’re so much better. So, this is the part that makes me believe in the durability of this. So let's take ChatGPT versus Internet Explorer— a little trivia. How long do you think it took Internet Explorer to hit 100 million users?

Sébastien Page

Ten years.

Dom Rizzo

A little high.

Christina Noonan

Five years.

Dom Rizzo

So five years. Well, you saw my notes. How long do you think it took ChatGPT to hit 100 million users?

Sébastien Page

You saw the notes.

Christina Noonan

One year.

Sébastien Page

Maybe less than one year.

Dom Rizzo

Yeah, two months. So five years, two months. We're all digital natives today as well. Every single person works with a phone, the internet, Outlook. Whatever it is, we're all digital natives. So the time to distribute this technology is completely different than it was in the ‘90s, right? People thought that the Internet was a fad in the ‘90s. I don't know that many people who think AI is a fad today, right? Not only that, the profile of who spent money on AI versus the profile of who spent money on the internet is completely different.

This AI capex boom is being funded by the most profitable companies in corporate history: Microsoft, Google, Amazon. These companies are behemoths, right? We haven't seen anything like these. Hundreds of billions of dollars of free cash flow that they had to be able to allocate to spending on capex, unlike the ‘90s fiber buildout. Completely different. We were running internet fiber, not knowing who was going to use it, and we were using a lot of debt to do it.

So, if you remember, two years ago we talked about this actually, where were we in the AI cycle? And I think I said, maybe, you know, in inning one or two. We're a little bit further along now today, and….

Sébastien Page

Oh, by the way, I’ve got to interrupt you.

Dom Rizzo

Yes, please.

Sébastien Page

Perplexity predicted that you would say we're still in the early innings.

Dom Rizzo

Well, so I think we're still early to mid-innings, and so here is the piece that is starting to change. We are just starting to go from the cash flow driven part of the cycle to the capital markets-driven part of the cycle. And two waves—I know you've talked about the private credit wave on here in the past.

Imagine when the AI wave meets that private credit wave. All the money that is sitting on the sidelines ready to go give private credit lending to companies to go build out some of this infrastructure that hasn't really even started to hit yet. And I think that's coming. And if you think about financial innovation, capital markets theory, these things always come with those productivity speculative bubbles.

And we're getting that financial innovation, which is that private credit coupled with that productivity cycle. And so I'm looking at that cocktail and I'm saying wow, we have all the pieces here for this to really keep going potentially for a little bit of time.

So, I really focus how do we navigate this responsibly? And it comes to my framework: linchpin technologies innovating in secular growth markets with improving fundamentals and reasonable valuations.

And I know we're going to talk more about valuation, so I'm going to put that to the side. But these companies are absolute linchpins for the economy. We're innovating in potentially the biggest secular growth market that the world has ever seen: AI chip spend $45 billion in ‘23 to $500 billion in ‘28; 50%+ incremental operating margins. Where we sit today, many of these companies are accelerating into the back half of the year.

Revenue will grow faster over the next six months than it did over the past six months. And valuations, yes, at the higher end of historical averages, but in my opinion, not yet extreme. And we'll talk about tha a bit more. But I put it together, and I think this AI trade can continue.

Christina Noonan

So a lot of the AI optimism you're describing benefits U.S. markets, and we've talked a lot about U.S. equity markets. I want to dive into some of our Asset Allocation Committee decisions. So, one decision right now is that we're currently overweight these markets outside the U.S., which is a trade that had worked very well earlier this year. Some give back since then. Sébastien, can you talk a little bit about what could drive international markets higher from here?

Sébastien Page

Recently, an analyst sent me a relative valuation dashboard for non-U.S. international stocks relative to the U.S., and as I was looking at the numbers, I started thinking this dashboard is completely wrong. It was very surprising because international earnings growth, margins, and ROE were all substantially higher than for the U.S. Does that make sense to you? Would you question such a dashboard?

Dom Rizzo

I would totally question that dashboard.

Sébastien Page

Earnings growth were 2% higher, and this was for All Country World Index ex-U.S. compared to Russell 3000. Earnings growth at 2%higher. Margins were 5% higher, 20% margins internationally versus 15% in the U.S. ROE was 5% higher for international relative to the U.S. And guess what? Like, there's a little twist here.

Dom Rizzo

I hope so.

Sébastien Page

The dashboard was actually technically correct. And this is going to blow your mind. It was no data error.

It was just a different lens. What lens? These were all median numbers. So when I say median numbers, let me give you an example for margins. You rank all the companies in the index by margin, from highest margin to lowest margin. And you go to the middle, the 50th percentile. And that's your median number for the index, meaning half the companies in that universe have a higher margin than that number, half the companies have a lower margin than that number. And when you look at it from that lens, because small- and mid-cap and value stocks in general are actually performing well outside the U.S. from a fundamentals perspective, you get a completely different view.

So valuation medians are preferable also for international stocks. So, it's about 19 for the Russell 3000, and it's about 15 for international stocks. So, if I switch my lens from market cap-weighted index to median and I say I'm going to look at all the stocks and maybe go down the cap spectrum—we’re long international small cap, for example—I get a completely different relative valuation perspective and fundamental perspective.

So, it's one of the reasons, one of the ways to illustrate why we're tactically overweight international stocks with a focus on Europe and value and small- and mid-caps. International stocks, if you just take the EAFE Index—

Do you just say "EE-FA" or "EE-FEE"?

Dom Rizzo

"EE-FEE."

Christina Noonan

"EE-FEE."

Dom Rizzo

I was "EE-FEE" when I was in London, so that may be….

Sébastien Page

The next time we do a webinar live, we'll have a live voting to say if you say it "EE-FA" or "EE-FEE."

I guess, OK. EAFE has outperformed the S&P 500 by 12% year-to-date, You mentioned this. By the way, a lot of this is the currency effect because the U.S. dollar went down. But EAFE is down 158%, even after that rally over the last 10 years relative to the S&P 500. Yes, it's been a value trap, but here right now, value and momentum both agree. Those who do backtesting and quantitative research know that, in general, valuation works better when there's positive momentum. So if something is cheap and is starting to outperform, that's usually a better signal than if something is just cheap standalone.

So you want the intersection of these signals, you have it right now in international. The main catalyst to unlock the fundamental and valuation advantage of those stocks include positive interest rates. That's good for European banks that have historically operated in zero or negative interest rates, stimulus infrastructure, defense spending. We think Germany can increase its debt-to-GDP ratio by over 20% for the next 10 years, and you have dovish central bank policy.

European banks traded at two times book value before the financial crisis. They're now just at about one, so there's room here. It really sounds like I'm pounding the table. I'm making the case. It's not a very large trade in the portfolios, but if you look at a broadening of the markets—not taking away anything from the AI team theme and from…I've named, I've called these companies awesome, awesome mega-cap tech.

But just tactically right now, betting on a broadening of the markets, international is a good way to do it. Our portfolio managers, the stock pickers across our platform, also like international health care and industrial names.

Christina Noonan

And we've been overweight international value, which has been a very strong performer year-to-date, up 25%; S&P up around 7%. So, just to put a finer point on the specific areas.

Sébastien Page

I'm being carried, I'm being carried away with pounding the table making my case, because I hear Dom make the case about AI, and it's like so emphatic. Why not?

Dom Rizzo

Well, I think what's so great about our firm is that we can have different ideas. We can have different…there's no house view at T. Rowe Price, right? Every PM has their own view, and so that's great. And, frankly, the great thing about my strategy and when I look in the marketplace and I look at competitor strategies, they're all U.S. tech usually. I run a global technology strategy. I'm able to find ideas anywhere in the world: all market cap, all regions, small-cap, mid-cap, large-cap, mega-cap, U.S., Europe, Asia. Wherever it may be, as long as it fits my framework, I can go find them. So I have some of those types of ideas in my strategy as well.

Christina Noonan

Yeah. And you were just in China and Europe, and can you talk about some of the key takeaways there and the tone around technology outside the U.S. as well?

Dom Rizzo

I had a pretty incredible last few weeks on the road, Sébastien. I was on the road for three weeks. And I was in Spain, London, China, California, and then actually back to London for our international off-site. So I've been traveling a lot lately. And wow, did we get some incredible insights on that trip. So one of the most important insights, and probably the most important insight that that I've had in the past few months, is how dominant the U.S. AI ecosystem is.

So I'm not just talking about the hardware but the software and how all these great China companies really want to build on top of the U.S. AI ecosystem, and what does that mean for these kind of geopolitical tensions that we saw before. So, one reason I actually think we've actually seen geopolitical tensions come down is because everyone knows how important the U.S. AI ecosystem is to winning that AI race, right?

Sébastien Page

Yes. And CUDA.

Dom Rizzo

CUDA particularly. Yeah, CUDA particularly, from NVIDIA is the backbone for so many different robotics companies, AI companies. Wherever it is, whether it's Europe, Asia or the U.S., CUDA is such a strong software ecosystem. Now, I do think AMD has their own ecosystem around something called ROCm that's becoming more robust. You have to remember, the NVIDIA team has been building out that CUDA ecosystem for decades now. And it's a very, very powerful move. And, frankly speaking, it's very strategic for the United States.

Sébastien Page

And so we sell lower-end chips that use CUDA that kind of flood the ecosystem. And then everyone's dependent on CUDA.

Dom Rizzo

For China, particularly. And, frankly, if you look at the China ecosystem, one really interesting takeaway for me: Huawei is the behemoth in China, right?

And they—purely from a technology perspective—they make amazing phones, they make amazing chips, they make amazing networking equipment, and they do all this despite a lot of rules around what they can't import into the country. So, incredibly technologically competent organization. Yet also they're often a competitor to many of the companies within the China ecosystem. So, in some cases, you don't necessarily want to rely on Huawei. You may actually want to rely on NVIDIA, who has better technology anyway. So it's a really interesting dynamic right now that's playing out. So, all that to say is I think the U.S. has this unique place in the AI ecosystem, particularly because of our silicon design and our software design.

But there's going to be winners in all regions of the world because this technology is just too powerful, right? There will be great names defined in Asia, all throughout it, in Europe. I have plenty of European exposure across the strategy because there are these linchpin technologies wherever you look.

Sébastien Page

And, by the way, for our audience, Eric Veiel, our head of investments, interviewed Jensen Huang in his podcast, and it's available out there, so reach out if you can't find it. But it's excellent.

Dom Rizzo

There's a funny anecdote in that podcast, like Eric is talking about, oh, some PM at T. Rowe Price was using Perplexity with me and showing me how to ask the questions. And, I was that PM. So, but look, I think there are great companies wherever you look, and if I put them through my framework—linchpin technologies innovating in secondary growth markets, improving fundamentals, reasonable valuations—you can find that really, truly in any region in the world.

But those U.S. awesome companies—and we’re going to talk a bit a bit more about valuation, I hope, because it's so important when you talk about AI in the ecosystem—they truly are awesome. I think it's a great description.

Christina Noonan

I think you laid out a very compelling case for U.S. AI. Can you touch on the valuations? How does valuation fit into your overall investment framework?

Dom Rizzo

Valuation really matters. And I think it's particularly important for the technology portfolio manager to say that valuation really matters. Now, valuation’s spectrum. It's all often in the eye of the beholder. It's a bit like beauty. And it's a function of the quality and the growth of the business. And so let's just look at a couple of numbers here.

Why do I say the valuations remain relatively reasonable across tech? The global technology index today trades at 24 to 25 times earnings. Historically, it would trade kind of in the low- to mid-20s on average. But at the at the short-term peaks, it gets to 27 to 28 times. So we're actually below those short-term peaks that that we've seen, say in 2021 or another periods. If you look at bubbles, it gets even higher, right? In periods of bubbles you can see, you know, north of 30 times. S&P 500? almost 50% of the S&P 500 today, is technology and the expanded technology sector.

And what do we know about the technology in the expanded technology sector? They grow earnings faster, they have higher recurring revenue, they have higher ROE. You know, everyone talks about the K-shaped economy. In many ways. We should be talking about the K-shaped indices.

Sébastien Page

Yeah. And it’s so important because when we say the S&P at a price/earnings ratio of 22 and we say that's really high historically, yeah, but historically you haven't had 50% of the index in technology at high earnings.

Dom Rizzo

It’s high earnings, high-recurring, high-profitable companies. Now, if you put it in history, there are periods where the S&P 500 goes to higher multiples, right? So 1999, we saw 28, 29 times earnings on the S&P 500. And then the last point on this Mag Seven thing. The Mag Seven historically will grow earnings anywhere two to three times faster than the rest of the market.

So this year in 2025, the Mag Seven is going to grow earnings roughly 15%. The 493 are going to grow roughly 5%. Yet, they trade at a 50% premium to those other stocks. So on a PEG ratio, on a P/E-to-growth basis, they are actually cheaper than the 493. So, I think you have to look at valuation in a couple of different contexts to get kind of the best answer. But I look at it and say yes, it's on the higher end of history. But no, it's not the craziest that we've ever seen historically. And there are good reasons for some of this as well.

Christina Noonan

Great insights there and important that valuations do matter. But a lot of nuances there. Any final thoughts before we wrap up here?

Sébastien Page

As I said at the beginning of this, Dom is one of our favorite, maybe our favorite, guests on this list.

Dom Rizzo

That's positive thinking right there, Sébastien.

Sébastien Page

You absolutely deliver, Dom. It's fantastic. Thank you. And thank you everybody for listening in.

Christina Noonan

So, yeah, great note to end on. And that about wraps up today's Asset Allocation Viewpoints webcast, AI Momentum and Tactical Asset Allocation. It's been a great discussion, and I want to extend a big thank you to our panelists for joining us today.

Thanks again, Sébastien and Dom. Today's slides will be available for download once our replay is ready. Just follow your registration link to access the replay starting on Wednesday. We encourage you to join us for our final Asset Allocation Viewpoints Webinar of 2025 in October, so please click the link above your video player to register. So thank you for joining, and we hope to see you again soon.