Brian Levitt:
Welcome, this is Rethink Markets. It's a series from the Greater Possibilities Podcast. I'm Brian Levitt. Jodi Phillips is here as always. Ready to get going, Jodi?
Jodi Phillips:
I sure am, Brian. So today I'm really looking forward to this conversation. We have Ashley Oerth, who's a senior investment strategist at Invesco, and we're going to have a conversation on artificial intelligence and what she believes the future is likely to hold for investors.
Brian Levitt:
Good. It's in the news. I think AI's not a fad at this point.
Jodi Phillips:
No, no. Maybe AI even wrote this podcast.
Brian Levitt:
Yeah, we're talking about it slowing down. There's scary letters coming out. But yes, we wrote this podcast or at least the outline for this podcast. So this is pure Brian Levitt and Jodi Phillips.
Jodi Phillips:
Yeah, yeah. You get what you get with this one. So anyway, Brian, how would you describe how investors are feeling about AI?
Brian Levitt:
Yeah, I mean probably conflicted.
Jodi Phillips:
Yeah, you sound conflicted with that answer a little bit also, but-
Brian Levitt:
No, look, I'm excited and we're all understanding more the benefits of all of this. I'm excited about the potential of it, what it can mean really for all facets of the global economy. But I think there are also investors are struggling to understand how we get there from here. Is it a mania? Are we slowing it down? That's what I mean by conflicted.
Jodi Phillips:
Well, sure. No, that makes total sense. And you used a very interesting word that I don't think I've heard you use before, although I've certainly read it in headlines, "mania." What do you think about that assessment?
Brian Levitt:
No, I don't think it's a mania. I mean I'm not going to be the first strategist to note this, Jodi, but earnings have been so robust this year, and the market hasn't advanced at that same pace. So that's not a mania, right? A mania is usually when the market is advancing far beyond the fundamentals.
Jodi Phillips:
Yeah. And of course that combination of elements you're talking about before that brought valuations down.
Brian Levitt:
Right. No, exactly. Price to forward earnings in the high teens. Maybe that's a little bit high compared to the history of the S&P 500 index, but it's not high compared to the last five years, or the last even 10 years. And that's when the S&P has increasingly looked more like a growth index given the concentration of it. So yeah, price to forward earnings in the high teens, to me, that's not mania.
Jodi Phillips:
No. And of course, I mean, look, we know that fear sells, dramatic headlines and whatnot. But we do also know that there is a substantial amount of investment that's being made right now in this AI space without having a full understanding of when the returns on that investment might fully arrive.
Brian Levitt:
Yeah, biggest debate. Biggest debate we have. What's the bullish side of that? The demand is only just starting to get going and it's growing at rates that we haven't seen before. This isn't a software upgrade, this is AI embedded in all we do. And so the question that investors really need to contemplate is when, if ever, does the supply of chips, memory, compute catch up to that demand? That's the big debate.
Jodi Phillips:
It is. And of course, another element of that debate is maybe we all slow it down as Anthropic's Dario Amodei recommended very recently.
Brian Levitt:
Or we all go extinct, that'll slow it down.
Jodi Phillips:
Yes.
Brian Levitt:
Or maybe not.
Jodi Phillips:
That'll slow demand down for a lot of things. A lot of things.
Brian Levitt:
We joke. We joke. Anyway, these are great questions to ask of Ashley. That's why you and I phoned friends, right? I mean we have our ideas, but there's always people around... That's why I love working at Invesco. There's always people around the firm who we dial up, either figuratively or literally, and ask them the questions. So I'm looking forward to that.
Jodi Phillips:
And we are going to get to Ashley in just a minute or two, but first we're going to just stick with you as the expert right now, Brian, and talk about
Narrator:
Trending Conversations
Jodi Phillips:
So Brian, what is it? I mean AI clearly, but there's a lot of other things in the world that are top of mind for investors right now. So what are you hearing about?
Brian Levitt:
The most is interest rates, so 5% on the 10-year, but I'll also add that to 4.75 or so on a two-year, what the Fed's ultimately going to do after their first-rate hike. So that 5% I think is really on people's minds, give or take a few basis points or two on a given day, about what's been going on with the 10-year treasury rate.
Jodi Phillips:
All right. Well, we talked about mania earlier, another buzzword for this conversation might be breaking point.
Brian Levitt:
Breaking point. Is this the breaking point?
Jodi Phillips:
Mm-hmm. So is it?
Brian Levitt:
No, it's hyperbole. That's hyperbole.
Jodi Phillips:
I figured you'd say that, but I had to ask.
Brian Levitt:
Breaking point of what I think is the question. Look, I'll date us a little bit, Jodi, but you and I are children of the '80s and '90s, right?
Jodi Phillips:
The good old days.
Brian Levitt:
Yeah, the good old days. Over those couple of decades, the 10-year rate was always above the nominal growth rate. So that only changed after the financial crisis when rates were persistently below the nominal growth rate of the economy. So yeah.
Jodi Phillips:
Yeah, so hyperbole.
Brian Levitt:
Hyperbole.
Jodi Phillips:
But bringing it a little more into the here and now, I mean the US grew 6.6% last quarter in nominal terms, and the 10-year was still below that nominal growth.
Brian Levitt:
Yeah.
Jodi Phillips:
So important to keep that in mind, but of course we can't expect 6.6% to be the trend rate.
Brian Levitt:
No, that's a fair point. That's a fair point. But what is the trend growth rate over the next 10 years, and shouldn't the 10-year address that? So if it's 2.5% Real growth seems reasonable, that would be up a bit, but we'll give some AI boost. Maybe 2.5% Inflation seems like a reasonable target. So there's your 5%. That seems reasonable. I mean I know Trump said it should be one. It's not going to trade like a credit, people need to get paid something that outpaces inflation, so it's not going to be one. But 5% seems reasonable.
Jodi Phillips:
All right, fair. So you view this more as a reset of rates in a better growth backdrop?
Brian Levitt:
Yeah, it's a reset like it's a better growth backdrop. But there's also bond issuances from the hyperscalers, so better nominal growth, more supply of bonds, that seems to be the story.
Jodi Phillips:
All right. And not a breaking point?
Brian Levitt:
No. No, like what do we mean by breaking point? So think of it this way. Jodi, I presume you have a fixed rate mortgage?
Jodi Phillips:
Yes.
Brian Levitt:
Or did you pay off the home? Did the Phillips family pay off the home?
Jodi Phillips:
Let's just presume there's a fixed rate mortgage in this conversation.
Brian Levitt:
Same with the Levitt family, unfortunately. So higher rates doesn't break that for us, we've got our fixed rate mortgages. Doesn't change it. It makes housing slow down, but for the Americans who own homes, for the most part it doesn't change it. Same thing with businesses. The businesses have done a pretty good job of locking in low rates. So investment grade and high yield spreads are tight because the businesses aren't over levered.
The other thing people talk about, maybe 5%, is it a breaking point? Maybe they mean equity valuations have to come down? But like we already talked about, it seems like that's already happened. Why isn't the market advancing as much as earnings have advanced? Is it possible that the markets have already priced for an environment where rates were going to go up? So yeah, that's not a breaking point. And we record this with the S&P 500 a couple of percent from an all time high, spreads tight. I'm not going breaking point.
Jodi Phillips:
Good. Well, now this brings us to our conversation with Ashley on AI. And so we asked her multiple questions and got really in depth into what she thinks about the current atmosphere and different concerns that you hear, and so let's get to it.
Narrator:
Phone a Friend.
Jodi Phillips:
First, we asked Ashley if AI is living up to her expectations, and here's what she had to say.
Ashley Oerth:
I guess first, it depends on who you ask when you ask how this is living up to expectations. To me, if you look out there, AI still has a lot of doubters, a lot of skeptics, but really I think it's beating expectations. So if we measure by adoption, for instance, it's progressing as one of the fastest technologies ever adopted. We're looking at somewhere around two billion, and some figures pushed as high as three billion individual users of AI globally. So really quite big numbers. Business surveys similarly, at least in the US, point to around one in five businesses are using AI, somewhere around 80% of large corporates. So it's really going through a massive adoption story, and I would say it's still very much early days.
If we measure by economic impact, I think it's a bit early to tell. But at least for some of the studies that I've seen, it suggests that there is a sort of productivity story that's beginning to build here, especially with industries with high exposure to AI. Really, it'll probably take a few years before we know the true direction of where this lands, but at least the direction of travel I think is going in the right way.
So I guess the one place where maybe we have some question marks is really around the monetization of these tools. And here the data is a bit more tricky to work through, primarily because a lot of the monetization is happening at the private side, and it makes it a bit more challenging to really understand precisely what the demand looks like, versus how much is being say subsidized by these AI labs versus really its users that are paying to make use of this. But I think overall, these companies, or what they're reporting in the media at least, huge eye-popping numbers in terms of their annualized revenue growth. So the inklings are there to say that this is also making money.
So broadly I'd say yes, AI is progressing as I would hope it to be. I think that for the level of capital being deployed, this is a very positive story. And it's showing up in earnings as well, especially if you look at the tech sector, those beneficiaries, the AI infrastructure boom, this is probably the single biggest story in markets, at least if we're looking at the sectoral breakdown of performance.
Jodi Phillips:
Next, we asked her to give us a little bit of context into the level of investment that's being made compared to the future potential for returns on that capital. So like trying to get to the point of when it might make sense to get concerned about that, and here are her thoughts.
Ashley Oerth:
This is probably, I think, the biggest question facing the AI theme. So as I mentioned, the sort of AI infrastructure boom, it's carried on. This is really all about building up the sort of computing capacity to power AI. And we have a good sense of the level of spend that we're seeing, so say sizing of data centers, the overall pipeline, but ultimately all the supply we're standing up it's about meeting what we think should be the explosive demand for computing power. And so far it has been, but forecasting that into the future I think is challenging. We don't quite know where that lands. We have comparatively less insight into the monetization trends, as I mentioned, how lucrative adoption has been and how much compute we'll need in the future to power these models. We just know it'll be a lot. So if we were to freeze today at the current levels of demand and say supply continues, it's built out, that this infrastructure boom plays out, then sure, return on capital likely to miss the mark. But that's of course not the story that we're seeing.
The way I like to frame this is that this computing demand, it's driven by two forces. So number one, there's the adoption of AI, which I touched on. And number two is that the technology itself continues to improve. So if we think about these forces, as adoption grows we need more computing power to meet that use, to meet that economic adoption. On the other hand, capabilities are still improving and cost to access those capabilities are also falling at the same time. So the demand for compute, it's growing because intelligence is becoming cheaper, and when we can do more with more intelligence and we pay less, we use a lot more of that. So there is your computing demand story. Now of course, as I mentioned, forecasting where this ends up is extremely challenging, and hence why we have so much angst around stories like DeepSeek last year, the recent talk around an intentional slowdown of AI progress, which I'm sure we'll touch on, and the other potential shocks that we could name for shocks to computing demand.
Jodi Phillips:
Next, we asked Ashley about agentic AI, and whether it's possible that the growth of AI agents could mean that supply constraints just never catch up to demand. And here's that discussion.
Ashley Oerth:
Yeah, I think this is also another good question, and I think that this agentic AI story is really an important one. So if you think about the history of AI, we've gone from models that were Q&A, we could play around with them, maybe write some poems, fool around a bit with just what it could give us in terms of creativity or maybe it would give us nonsense. It was more fun rather than productive. But now with the age of agentic AI, we have these models that go through various reasoning steps and are able to go off and give themselves a task list for how to accomplish the objective you've given them. So maybe they need to build a program or build a workbook and do some actual analysis and come back to you. And that's what agentic AI is really all about, is how we can really impart some agency really into what these models can do, that they can actually carry out this extra work and get to the ultimately better outcomes.
Now, in terms of the supply constraints versus catching up to the level of demand, I mentioned this already a little bit but AI is interesting really because ultimately what we're trying to put a price on is intelligence. So suppose we take a cheap, low performing, really plain AI model, maybe those earlier generations I mentioned before. Say what your mobile phone can run when it's not connected to the internet, maybe you're editing an image or something like that. Realistically, it's quite limited and you'll probably not value it very highly. Now, if we turn to the stuff of what we're talking about today of what's going on in the data center story, these more expensive kinds of AI that run on these sort of dedicated computing centers, these are much more powerful. And here we have the sort of complex reasoning models, the agentic AI, all kinds of impressive achievements. And as we explore, right now, and implement all of the potential use cases of what this intelligence can do for us, we're still going through this process where we're hard at work of seeing where this is economically useful, where it's profitable. And as we play out in a story, when its price changes, when the price of intelligence falls, when it becomes more accessible, we see more uses of it.
So as we progress through this AI story, we keep revising really three things. Number one, what is AI capable of? Number two, what we think it will be capable of. So in other words, the progress has continued, these models have gotten better. And then number three, how efficiently we can actually achieve those capabilities. So in other words, falling costs. So in the process of trying to really price intelligence and determine the demand of it, it's increasingly more powerful, it's increasingly cheaper. And at some point, I imagine we'll reach some kind of equilibrium where there's a price that is reasonable for how much intelligence we're getting versus the actual price of that computing power that goes into serving that. But we're not there. We keep revising it, as I mentioned, models keep getting better and more powerful. And even at the cutting edge, seems expensive at the time when it's first announced, but then just months later, those costs have tended to fall dramatically. Take the first instance of ChatGPT. Its costs fell by about 99.7%. That's really, I think, highlighting just how much more accessible this gets. And that's been playing out time and time again with future generations of models.
So really when we're trying to price how expensive AI is to access versus its abilities, it's really difficult to price intelligence versus the kind of marginal cost of compute. But we keep moving the goalposts because computers are getting more efficient, models are getting more efficient and they're getting more powerful. So it's really, I think it's a long way of answering a question to say we don't know.
Jodi Phillips:
Then we asked Ashley to put her investment strategist hat on and talk about where she thinks the focus should be right now between cloud compute providers, service providers, chip manufacturers, downstream beneficiaries, and here's how she thinks about that.
Ashley Oerth:
Well, I do think that there is the short answer there, that yes, it is all of the above. But it really is about picking and choosing at this moment. Where in the early days, the AI trade seems like everything went up, whereas today there's more selectivity that's out there. So first it was hyperscalers were the place to be. Now those hyperscalers are going through things like they're loading up on debt and the cost of capital is shifting against them. And once you get to a certain size, it's difficult to keep growing at the same growth rates. So they've become less stellar in the story and I think subject to more investor scrutiny, but still they are spending, they are investing, and that's where the downstream beneficiaries, at least insofar as the supply chain side for what goes into a data center, that's still, I think, quite healthy.
So if you think of it year-to-date perspective, it's been a phenomenal story. But as you mentioned, there's been some trials along the way. I think that the sort of rerating that we've gone through in recent months is not to say that this AI story is over or dead in the water or reversing, I think it's more about taking a breather after a phenomenal growth rate. But I still very much see this kind of structural story in play, that we remain compute constrained, that there's a narrow set of companies that are really able to deliver what goes into data centers, the chips, the networking equipment, the power generation equipment. So really, I think that supply chain can continue to benefit. But equally, we're understanding how to use this technology better and better. So we're seeing more implementations, and I suspect that we'll see some companies that just get it right, that they'll have AI-driven products or services, and we'll see some increasing dispersion as a result of that outside of the tech sector, outside of the data center story.
Jodi Phillips:
Finally, we asked her about Dario Amodei and other tech leaders who are advocating for an intentional slowdown of the pace of AI development, and what the market implications could be of that. And here are her thoughts.
Ashley Oerth:
It's I think a very interesting story, and one that we've seen come up time and time again over the past few years. So as I mentioned, AI keeps getting better, and now that it's gotten to a point where it's so good that we're starting to ask questions of maybe slowing the rate of development here, and that's what this letter is all about. But if we back up, really this comes back to security concerns that it spawned. That as AI has improved, these labs have been experimenting with just what it can do, trying to lock these AIs in a box where they can't access the internet and give them a task, oftentimes an impossible task, and seeing what they do with it.
And this is where this story really came out of, is that we have these models that we're actually able to break out of the sort of containment that we built for them, and they were going off and doing things that we did not expect, that we did not program them for. That they went off and, for instance, hacked other companies and were breaking into other tools or other databases that they should not have had access to. And that's really, I think, the crux of this story. Is that the behavior that they're exhibiting is different than what we would expect and what we would want, what the parlance of the industry would call alignment. These models are looking misaligned, and that's what this essay that Dario Amodei put out about is really about an intentional slowdown of the pace of development so we can get a better handle on this.
Now, like I mentioned, this has happened before. These concerns, these cautions have been flagged in the past, but what's different about this moment is that it's a variety of tech leaders that have agreed to this. So Sam Altman of OpenAI, Elon Musk of SpaceX, Satya Nadella of Microsoft, they've all agreed to the broad framework of trying to slow the pace of development, try to embed human reviewers into these companies, these AI labs to try to make sure that the story isn't going off the rails, to develop AI ultimately more safely.
Now for markets, why markets care, and really what this is a market story about is does this affect compute demand? Does this mean that we slow those ever-expanding horizons? Do we stop doing that? And therefore, do we need less computing power? Personally, I'm not so concerned about that. I think that there's already a very healthy pipeline of demand because, again, adoption is still early days. It's still growing, we're still learning how to use this technology. And even if we froze progress today, it's still wildly powerful and I think that we're just still trying to get our heads around how to implement this, how to use it, how to expand it into businesses, into the economy. And that's really, I think, what the story is all about, at least insofar as the AI investment boom goes.
Jodi Phillips:
Thank you so much, Ashley, for joining us. And now we're going to move on to...
Narrator:
I Disagree.
Jodi Phillips:
And this is where Brian chooses a quote that's been in the news lately, or maybe from decades ago, you never know, but Brian disagrees with it. So what is it this time?
Brian Levitt:
Yeah, and it's not going to be disagreeing with Ashley.
Jodi Phillips:
Certainly not.
Brian Levitt:
Certainly not. Okay, so I'm going to go with Ray Dario. They've been in the news a lot lately, and he said recently, "I am confident that the government's financial condition is at an inflection point, because if this is not dealt with now, the debts will build up to levels where they can't be managed without great trauma."
Jodi Phillips:
All right, so you disagree?
Brian Levitt:
I disagree, vehemently.
Jodi Phillips:
Very good. So this is Ray Dalio, correct?
Brian Levitt:
What did I say, Dario?
Jodi Phillips:
Yep, Dario. Dario Amodei's on our mind, I know.
Brian Levitt:
Oh my God. Yes, Ray Dalio.
Jodi Phillips:
All right, all right. But now we know who we're talking about. But why do you disagree?
Brian Levitt:
Well, and since I'm one-take Levitt, we're not editing that out. We're keeping that.
Jodi Phillips:
Yeah.
Brian Levitt:
So I heard the inflection-
Jodi Phillips:
This is how you know AI didn't write it.
Brian Levitt:
Or record it. So I heard the inflection point was when the boomers all turned 65. Okay? That was going to be the inflection point.
Jodi Phillips:
That was. Well, now the oldest baby boomers are 80.
Brian Levitt:
Yeah, and the median's 71. So that wasn't the inflection point.
Jodi Phillips:
No. I mean, if anything interest rates were historically low when most of the boomers were 65.
Brian Levitt:
Yeah, dollar was strong too. Okay, so not that.
Jodi Phillips:
All right. So why does Dalio think that this is an inflection point?
Brian Levitt:
Dalio, not Dario. Right, Jodi?
Jodi Phillips:
Correct.
Brian Levitt:
You got it right. You got it right. Well, I think because the country's running very large deficits during good times, we're bringing in about five trillion in revenue, we're spending seven trillion. Okay, so that's usually reserved for a recession, and it's happening at a time when interest rates are rising, so that's why it's getting a lot of attention.
Jodi Phillips:
Okay. But the US is now, for example, paying more to service the debt than paying to fund the defense department. So yeah, I'll ask again, why wouldn't that be an inflection point? That seems big.
Brian Levitt:
Yeah, those are inconvenient truths, right? Is there anything to suggest that it must be dealt with now? We already talked about rates moving up with stronger nominal growth. I guess the question is, what is Dalio asking investors to do? Because it's a pretty bold comment to say, "We're at an inflection point." What does that mean? What are we supposed to do? I mean he's made his money, are we supposed to take our chips off the table because there's an inflection point coming?
Jodi Phillips:
Yeah, yeah. Seems like a risk for investors to think that way.
Brian Levitt:
Yeah. Let me ask you this: is it more likely that the US will sit idly by if there's a huge debt challenge and it's going to cause irreparable damage? Or is it more likely that the US will take steps to stave off the challenge? You might say, "What are those?" I mean we've seen the Treasury get involved a bit to not a lot of effect, but Congress could act. We're not just going to sit there and let the crisis happen. There's a lot of tools at our disposal to address a crisis.
Jodi Phillips:
Understood. And as you also like to say, yeah, the debt's risen from 20 trillion to 40 trillion over the past 20 years, but during that time, household net worth has climbed from 85 trillion to 180 trillion.
Brian Levitt:
Yeah, I'm glad you listened. And maybe it's that 180 trillion that's driving all of this growth and causing interest rates to go higher. So no, I don't think this is our comeuppance. I'm glad you listened to me though, Jodi. Thank you.
Jodi Phillips:
Always. All right, so now we're going to put Brian on...
Narrator:
The Hot Seat.
Jodi Phillips:
We just got our first Federal Reserve rate hike in three years, Brian. Are you concerned?
Brian Levitt:
No, it's not an over-levered economy. We talked about that with our mortgage rates. Also comforted that inflation expectations in the bond market are contained, and also comforted that the two-year treasury is up at 4.75, so that's already priced in additional rate hikes.
Jodi Phillips:
All right. Oil prices have spiked again, so what's driving that?
Brian Levitt:
Yeah, we were doing a decent job as the... Well, not we, I don't drill for oil. But the world was doing a decent job making up for those 15 million barrels per day that were going through the Strait of Hormuz. Some of that makeup were the five million of barrels coming from the Saudi East-West Pipeline. Unfortunately, that was hit by a drone in September, so that's offline for a bit. Fortunately, there are estimates that it could take days or weeks to get repaired, but importantly, not months or years. But that's what drove it.
Jodi Phillips:
Okay. So next question, there's a lot of negative crosscurrents, the rates, oil, war, AI fears. So why is the US stock market still near an all-time high?
Brian Levitt:
Earnings. And you also make a good point about being near an all-time high. I think people have been feeling like there's been a huge correction in the market, it's just not the case. And historically, a wall of worry isn't the worst thing for stocks. You just named a pretty big wall, so we're climbing it.
Jodi Phillips:
All right. Sounds good. So what would it take for you to get negative on stocks?
Brian Levitt:
Yeah, earnings begin to disappoint, analysts revise lower, hyperscalers scale back AI investment, bond market pushes back against AI issuance. Those are the types of things. None of those really seem to be happening.
Jodi Phillips:
Okay. Well, you are now off The Hot Seat. And we're going to conclude with our newest segment, Think Rethink. So Brian, tell me something that you believe people are thinking right now, but you believe they should rethink.
Brian Levitt:
Yeah. So can we think what we talked about earlier about manias?
Jodi Phillips:
Mm-hmm.
Brian Levitt:
People think the stock market advance is a mania.
Jodi Phillips:
Okay.
Brian Levitt:
Can we rethink and I'll tell you what a mania might look like?
Jodi Phillips:
Yes, help us rethink that.
Brian Levitt:
Okay. So stock market is up 127% since the last time we had a 20% decline. That was around 2022, coincidentally about the same time ChatGPT came out. So we're up 127% over 3.9 years. In the 1990s, we advanced over 500% over 10 years without a 20% decline.
Jodi Phillips:
Oh, okay.
Brian Levitt:
So just by definition, I mean that pales in comparison…
Jodi Phillips:
The comparison.
Brian Levitt:
Yeah, it's just not a mania.
Jodi Phillips:
All right. So before we go, Brian, where can our listeners follow you for more insights?
Brian Levitt:
Please visit invesco.com/brianlevitt to read my latest commentaries. And of course you can follow me on LinkedIn and on X @brianlevitt. I'm hearing I am now being paid by the click, so please go do that.
Jodi Phillips:
Huh, all right. You have to tell me about that deal after the show.
Brian Levitt:
I don't think it's a good thing.
Jodi Phillips:
Oh no. Thanks for joining, we'll see you next time.
Brian Levitt:
Thank you.
Important information
You've been listening to Invesco's Greater Possibilities podcast, Rethink Markets.
The opinions expressed are those of the speakers, are based on current market conditions as of September 18, 2026, and are subject to change without notice. These opinions may differ from those of other Invesco investment professionals.
This does not constitute a recommendation of any investment strategy or product for a particular investor. Investors should consult a financial professional before making any investment decisions. Should this content contain any forward looking statements, understand that they are not guarantees of future results. They involve risks, uncertainties and assumptions. There can be no assurance that actual results will not differ materially from expectations.
All investing involves risk, including the risk of loss.
Past performance does not guarantee future results.
Investments cannot be made directly in an index.
References to specific companies aren’t buy/sell recommendations.
Artificial intelligence (AI) technology companies are sensitive to specific risks such as small markets, business cycle changes, economic growth, technological progress, obsolescence, and regulation. These companies may have limited products, markets, resources, or personnel, making their securities more volatile, especially for smaller start-ups. Rapid technological changes can adversely affect their results. AI companies often rely on patents, copyrights, trademarks, and trade secrets to protect their technology, but there's no guarantee these protections will be sufficient. Significant research and development (R&D) spending doesn’t ensure product or service success.
Many products and services offered in technology-related industries are subject to rapid obsolescence, which may lower the value of the issuers.
Fixed income investments are subject to credit risk of the issuer and the effects of changing interest rates. Interest rate risk refers to the risk that bond prices generally fall as interest rates rise and vice versa. An issuer may be unable to meet interest and/or principal payments, thereby causing its instruments to decrease in value and lowering the issuer’s credit rating.
In general, stock values fluctuate, sometimes widely, in response to activities specific to the company as well as general market, economic and political conditions.
Booming AI investment is illustrated by hyperscaler capital spending, which has more than tripled since 2022, based on Bloomberg consensus analyst forecasts as of March 2026. Hyperscalers include Amazon, Google, Meta, Microsoft, and Oracle.
Estimates of US business usage of AI sourced from the Organisation for Economic Co-operation and Development and the US Census Bureau.
Discussions about robust corporate earnings sourced from Bloomberg, L.P., as of June 30, 2026. Based on the operating earnings of the companies in the S&P 500 Index.
Comments about the price-to-forward earnings of the S&P 500 Index sourced from Bloomberg, L.P., as of September 17, 2026.
The 10-year US Treasury rate sourced from Bloomberg, L.P., as of September 17, 2026.
The US nominal growth rate sourced from the US Bureau of Economic Analysis and Bloomberg.
Comments about equity valuations sourced from Bloomberg, L.P., as of September 17, 2026. Based on price-to-forward earnings of the S&P 500 Index.
Discussion about spreads sourced from Bloomberg, L.P., as of September 17, 2026. Based on the option adjusted spread of the Bloomberg US Corporate Bond Index and Bloomberg US Corporate High Yield Bond Index.
The Bloomberg US Corporate Bond Index measures the investment grade, fixed-rate, taxable corporate bond market. It includes US dollar-denominated securities publicly issued by US and non-US industrial, utility, and financial issuers.
The Bloomberg US Corporate High Yield Bond Index measures the US dollar-denominated, high yield, fixed-rate corporate bond market.
Option-adjusted spread (OAS) is the yield spread that must be added to a benchmark yield curve to discount a security’s payments to match its market price, using a dynamic pricing model that accounts for embedded options.
Comments about the age of Baby Boomers sourced from the US Census Bureau, 2026.
Comments about interest rates and US dollar strength when boomers reached 65 sourced from Bloomberg L.P. Based on the 10-year US Treasury rate and the US Dollar Index from 2010 to 2025, respectively.
The US Dollar Index measures the value of the US dollar relative to the majority of its most significant trading partners.
Comments about the size of the US deficit compared to the Defense Department budget sourced from the Congressional Budget Office.
The growth of the US debt sourced from the US Treasury.
Data on US household net worth sourced from the Federal Reserve.
Comments that the economy is not over-levered sourced from Bloomberg L.P., Macrobond, and the Federal Reserve as of June 2026. Based on household debt to household net worth.
Comments about inflation expectations sourced from Bloomberg, L.P. as of September 17, 2026. Based on the 5-year US Treasury inflation breakeven.
Breakeven inflation is the difference in yield between a nominal Treasury security and a Treasury Inflation-Protected Security of the same maturity.
Comments about global oil supply sourced from the US Energy Administration.
Data on the advance of the US stock market sourced from Bloomberg, L.P., as of September 17, 2026. Based on the S&P 500 Index.
Estimates on individual and business use of AI based on various company and media-reported estimates. For illustrative purposes only.
The cost decline of the ChatGPT model sourced from the Stanford AI Index 2025 report, Epoch AI, and LMSys Chatbot Arena. From November 2022 to April 2025.
A basis point is one-hundredth of a percentage point.
The cost of capital is the opportunity cost of an investment or the rate of return that a company would otherwise be able to earn at the same risk level as the selected investment.
Dispersion measures the degree of uncertainty, and thus risk, associated with a particular security or investment portfolio.
Hyperscalers are large cloud service providers that can provide services such as computing and storage at enterprise scale.
An inflection point is an event that results in a significant positive or negative change in the progress of a company, industry, sector, economy, or geopolitical situation.
Price-to-forward earnings is a valuation ratio that compares a company's current share price to its expected earnings per share over the next 12 months or next fiscal year.
Agentic AI refers to artificial intelligence that can independently plan and carry out sequences of actions to achieve a goal, often using software tools, data sources, or other systems, with limited human supervision.
A spread is the difference between two related values, such as prices, rates, or yields.
Rethink Markets and the Greater Possibilities Podcast are brought to you by Invesco Distributors, Inc.
