The Royce Roundtable: Risks and opportunities in AI—and elsewhere

Portfolio Manager Joe Hintz and Assistant Portfolio Manager Tim Hipskind join Royce CEO and Co-CIO Chris Clark and Co-CIO Francis Gannon to discuss AI’s impact on the US small-cap market and why earnings strength can sustain recent US small-cap leadership.

What were some of the factors that drove small-cap performance both in 3Q25 and off the 4/8/25 low?

Francis Gannon: I think that the two biggest factors for small-cap performance were rates coming down and healthy earnings. We’ve been discussing what looks very much like the end of a nearly two-year earnings recession. For a lot of small-cap companies, including many of our own holdings, earnings in general were better in the second quarter, which definitely helped small-caps—and micro-caps—to do well on both an absolute and relative basis for both 3Q25 and off the low on 4/8/25.

The Russell 2000 Index advanced 12.4% in 3Q25, the Russell Microcap Index rose 17.0%; both were ahead of the large-cap Russell 1000 Index, which was up 8.0% for the same period. From the 4/8/25 market low through 9/30/25, the Russell 2000 climbed 39.9%, the Microcap Index rose 53.8%, and the Russell 1000 was up 35.2%.¹ Needless to say, those were impressive numbers regardless of asset class, but we’re pleased with how well small- and micro-cap stocks have fared since the market began to recover in early April.

Joe Hintz: Picking up on that point, the only issue is that the rally has so far mostly benefited lower-quality small-caps—which is consistent historically with previous early small-cap runs off of a of market bottom. When we look at the top performers in the Russell 2000, we’re seeing companies in biotech or Bitcoin mining; we’re seeing a lot of non-earners, a lot of more capital-intensive businesses or companies carrying large debt loads. So, it’s been a bit tougher but we expect this to shift as the rally matures, but it’s still been a little frustrating.

Tim Hipskind: US small-caps came out of a very uneasy moment in early April—that followed a bearish first quarter—in which the market believed that small-caps would be more negatively affected by tariffs. When investors got a look at the big board Trump brought out on “Liberation Day” to show all the tariff rates, there was a lot of fear. People were wondering, “Is he really going to potentially crush the economy to reorganize world trade?” Of course, a lot of what was on that board did not end up sticking or continue to be delayed. As that anxiety began to unwind, some of the optimism of having a business-favorable president returned, which helped clear the way for better small-cap performance.

I think another important factor in the quarter that drove market performance including small-caps was how AI reasoning models really grabbed hold of people’s attention. I’m specifically thinking of o3 from ChatGPT, which got a lot of acceptance in the quarter. I think it really opened people’s eyes in terms what sort of tasks these models will be capable of, and the associated inference that will be needed, and all of the downstream implications of that. That obviously drove a lot of the large names in AI, but it also impacted many companies downstream of that, including companies involved in data center construction.

How have you been thinking about ways you—or the companies you hold—can begin to quantify the downstream benefits of AI?

Tim Hipskind: For a few reasons, companies have a tough time saying anything like “Our AI benefit was 10 cents to EPS.” We do think that the benefits will eventually accrue to a much broader array of companies in productivity and efficiency beyond those building out AI infrastructure. We’ve also seen companies reassess their initial estimates of 30%-40% increases in productivity, with 15% becoming a more common data point.

So, we’ve had this first phase in which businesses are figuring out how to get all the data in place in order to benefit from AI. Next comes rounds of testing to determine if a company is getting consistently positive outcomes. I think as investors we’re still a little early in terms of being able to measure these benefits because the companies themselves are still learning about how AI can be used in productive ways.

Joe Hintz: I agree with Tim on all of that. So far, the biggest benefits have been on the coding side, specifically with some of the large-cap tech companies. With some of those businesses, I think you can probably start putting some actual numbers together to measure its impact.

As investors, we’re focused on both the risks and the opportunities around AI, and it’s become part of our research and due diligence to ask management teams how they’re using or planning on using it. For most of the companies we’ve been talking to, AI remains very much on the exploratory side, so it’s difficult to put actual numbers to it right now. No one really knows yet how much AI will save these companies in costs, how much it’ll boost productivity, and/or how much of those benefits will be kept versus passed on to customers.

Going back to recent performance for a moment, has the strength of small-cap performance over the last six months been surprising?

Chris Clark: Not really. We’ve certainly been waiting for a long time for small-cap to reassert its leadership after an extended period of underperformance, and it’s been gratifying to see small- and micro-cap stocks do as well as they have even as some of our strategies have been challenged to keep pace with the early phase of the upswing.

In fact, out of all the important elements we’ve been discussing in terms of small-cap’s recent robust performance, I think the Fed's September rate cut was the most important, especially in light of how well the lower-quality factors Joe detailed have performed so far. There’s an expectation that smaller businesses that finance with variable rates were going to have the most discernible benefit from a rate cut—which is precisely what we’ve been seeing. Everyone else is on a somewhat similar playing field with regard to the overall economy, as well as industry and sector dynamics.

Looking back at previous cycles, we can see a transition from the initial phase that benefits more levered, lower-quality businesses, before strong financial and operational fundamentals start to matter more, especially if vibrant economic activity is sustained. Prior small-cap bull runs usually see a leadership shift toward higher-quality businesses—for example, those with low or no debt, healthy cash flows, and positive earnings.

Small-caps have had a few outperformance head fakes over the last several years—short-term runs that lasted for a few months before large-cap recovered leadership. Do you see anything that leads you to think small-cap can sustain its recent outperformance?

Joe Hintz: Earnings are the key. Stocks always follow earnings. As Frank mentioned, the last two years have been tough for a lot of small-caps. First, more than 40% of the companies in the Russell 2000 currently have no earnings²—and many of those that do had not been keeping pace with large-caps until just recently. There was also a lot of pull forward in earnings power within small-caps, especially during COVID. And while we stay away from companies with high debt loads, there’s typically a lot of floating rate debt broadly within the asset class, and those stocks were penalized when rates were rising.

On balance, then, if we looked at most of those head fakes over the last couple of years, we’d find that small-caps didn’t have the broader earnings strength follow through across most of the asset class. This period feels different. I think we're heading into a phase where earnings growth is going to be stronger for small caps.

Chris Clark: I agree. It’s important to emphasize that rising rates and inflation were massive headwinds for small-caps coming out of COVID, despite small-cap’s history of outperforming inflation. Certainly, the rate story has improved, and while inflation remains an issue, the rate of change has stabilized to the point where it’s no longer such a powerful headwind.

How critical do you think policy uncertainty is in terms of the health both for the market and the economy?

Joe Hintz: Uncertainty is the enemy of capital spending. We all need to adjust and get as comfortable as possible, because it’s hard to tell whether current policy uncertainty is going to go away, especially given how rapidly things change from week to week and even day-to-day.

Do you have any concerns about the state of the US economy?

Tim Hipskind: I think the unraveling of the AI trade might be the biggest risk. To provide a kind of rough estimate: if we take the biggest publicly traded AI players, along with some of the smaller ones that have raised money, they account for roughly US$10-US$11 trillion of market cap added from the AI trade alone. If we put a 20 times multiple on that, it implies around US$700 billion of operating profit to the S&P 500’s earnings—and that amounts to a huge uptick.³ By some estimates, open AI will need to raise another US$115 billion. So a lot of benefit is being underwritten by the market, ahead of significant liquidity needs and questions about the payback on AI investments. If for any reason investor appetite to fund these endeavors dries up, it's obviously a huge risk.

The top 10% of spenders account for an increasing share of overall consumption, and that top 10% cohort is widely exposed to AI stocks. If there’s any weakness in that segment, it’s going to hit these richer consumers hard and will eventually trickle down to blunt consumer spending levels. That’s my biggest concern—the liquidity for these AI companies. As long as they can keep raising money and telling people they're going to create digital god, then they'll probably be able to continue to raise the hundreds of billions of dollars they need. But this is unlike any other capital build out we've ever seen, just in terms of dollars, so if the music stops, that’s a significant issue.

Joe Hintz: What we really need to see is a broadening out of the AI benefit. To the extent that we can start to see more companies benefiting from AI and a broadening out of the AI story within all market cap ranges, as opposed to a handful of companies that are benefiting from this massive AI buildout, getting into the actual usage of AI and turning that usage into strong benefits, whether it's for the consumer or the company, hopefully both—going forward, that is the question.

2025 was supposed to be the year of Agentic AI. Agents are getting built out, but at a slower pace than people had hoped for. As long as that is only a delay, rather than something that’s not going to actually happen, then it’s ultimately a positive. But that is a big question. We haven’t yet seen how AI actually gets disseminated throughout the economy and how those benefits will be distributed. That’s both a risk and an opportunity.

Are tariffs still a concern?

Joe Hintz: Yes, I think tariffs are one of the biggest uncertainties. We still don’t know the extent to which they’re going to impact the economy; will they turn into a nice revenue stream for the US government or are there negative impacts coming that we just haven’t seen yet? Our concern is that it’s more the latter than the former except that we have no way to call how and when they will make a meaningful impact, whether positive or negative.

There’s also the question of whether companies are waiting for policy certainty to start increasing prices and passing them on to their customers: Have they been running through lower priced inventory before they begin moving higher priced, tariff-affected inventory? There’s just so much uncertainty around the economic impact and how durable that impact may be. At the same time, it feels like the market is pricing in the idea that everything is going perfectly smoothly—and to the extent that it doesn't, that seems to be a pretty significant risk to us. This point of view hasn’t worked as well as low-quality or high growth small-cap investing off the April 8th bottom, but we still think that the potential for a correction and/or increased volatility is out there.

What is your long-term outlook for small-caps?

Francis Gannon: I feel pretty confident that AI will broaden out to other areas of the market. I think that investors are going to realize that the people and companies who provide or create AI are one thing—but companies and people who’ll be the beneficiaries are equally if not more important.

Chris has been saying, “AI is coming to a small-cap company near you,” and I think that—as Joe and Tim have detailed—there are great opportunities as well as near-term risks. Thinking about the long term, as we always do, I see the opportunities outweighing the risks, but, to Joe’s point, there’s certainly the possibility for more volatility if investors get impatient about the progress. But we’re already seeing companies in areas like energy and construction that are benefiting from the AI data center buildout in addition to more familiar industries like semiconductors and semiconductor equipment.

Joe Hintz: Given both our long-term investment horizon and our emphasis on quality, we’re looking over companies involved in AI for what we think are the most attractive secular opportunities that might have outrageously inexpensive multiples on them. With every investment, we are constantly asking questions both from an opportunity and a risk perspective, particularly on the risk side, when we’re looking at new opportunities. We are seeing a lot of names that have gotten incredibly beaten down.

We’re also starting to see some green shoots in non-AI exposed tech and in certain consumer areas where we focus on companies that can potentially benefit from inflation and tariffs. Management’s ability to run a relatively defensive business model with the ability to push through price increases is a pretty strong benefit in our view.

Tim Hipskind: We’re cautiously optimistic. Much of the market looks expensive for our discipline, which focuses on companies with superior balance sheets, sustainable returns on invested capital, and strong levels of free cash flow from operations, among other attributes. But we also added four new names in 3Q25. Given everything we’ve been talking about, we think that investors who’ve made a bundle of money in large-cap AI stocks may want to preserve their gains with some upside growth.