Quality companies, with low debt and solid returns, have historically traded at healthy premiums. But as investors flock to high-powered artificial-intelligence names, something interesting is happening, says Julie Biel, chief market strategist and portfolio manager at Kayne Anderson Rudnick Investment Management: Those traditional businesses have increasingly become available for more-reasonable prices. "I think that's a function of the excitement, particularly in small-caps, on more speculative companies and businesses that are directly tied to AI," says Biel. "Our nice little industrial companies, our nice little financial businesses, just don't have the same sex appeal."
Speaking with Barron's Advisor, Biel weighs in on the rebound in software-as-a-service (SaaS) stocks following fears of an AI-pocalypse. And she argues that companies using AI to improve their operations may end up being the winners, rather than the upstream AI companies selling the core models, computing infrastructure, or specific AI services. SharkNinja, anyone? And she says "good deadlock" resulting from the November election would be a dream scenario, at least for markets.
Long-term bond yields are at nearly 20-year highs, and the Fed is raising short-term rates for the first time in three years. What are the implications of this moment for stocks? It's an important moment because there is a recognition that things have changed. We have a very different kind of Fed. We should expect more volatility in interest rates as a result of this transition away from specific forward guidance, and we're not even getting reaction-function information [rules the Fed uses to react to economic changes]. That's important for stocks in that it probably makes people think a little differently about risk and about how many businesses they really want to own that don't yet have earnings. That's what we saw in 2022 after 2020 and 2021 when we had tech IPOs and investors pushing management teams to be unprofitable so they could grow as quickly as possible. Then all of a sudden it was like: Oh no, what if interest rates are going up? We saw a real reset as a result. I don't think this is going to be anywhere near as significant, but higher interest rates inherently make it harder for lower-quality businesses to do as well. This isn't a period where a rising tide lifts all boats. It really starts to matter if you have a differentiated business with a good balance sheet, because it creates differentiation. As an investor, you really have to roll up your sleeves and recognize that this is a more nuanced market than it had been.
Software-as-a-service stocks as a group have bounced back following fears that they were going to be eradicated by AI. What do you see as the lesson in that rebound? For many years, portfolio managers could rely on software names as businesses with high levels of recurring revenue. Even though they were a little bit on the pricey side, they had high returns on equity, durable earnings, and good growth. It's hard to find businesses that have a similar level of durability, so people felt comfortable owning a good amount of those stocks. It changes the dynamic when you're suddenly selling your software names because you now have more cyclicality in your business. So I think this kind of bounceback is a recognition of a couple things. It's so much easier now to produce software, but does that necessarily mean that companies will? I think that's a question mark.
You can make a case that there is going to be pricing pressure on a lot of the traditional SaaS names-for sure the ones that don't have very differentiated software or are not embedded in workflows. But it's also now faster for these companies to produce new software, new AI modules, what have you-and they have the distribution, which a lot of smaller upstart AI companies don't. That's not a trivial thing. Being able to actually call on your customers is critical when you're talking about enterprise sales. Not all of them are going to be successful. But I see in the software companies that I continue to own this interesting dynamic where they're excited about being able to internally generate new software.
A good example is a company called the Descartes Systems Group. You could think of this business like the Bloomberg for trade. They probably touch two-thirds of the packages shipping globally. The business historically bought modules to add on to their network, for things like customs processing and analyzing tariff impacts, and that's a lot of how they've been able to grow. For the first time this year, we're hearing the CEO talk about how with AI they feel more confident in their ability to develop software on their own. That really changes the dynamic for them: They don't have to integrate new software they've acquired. They don't have to take on a whole bunch of goodwill. They can purpose-build. I do think with some of these businesses, their reason for being is in question. There's some existential threat to some of the lower-quality software businesses that don't have good data or are not embedded in workflows. But for many software companies, this is an opportunity to protect what they have and to attack.
How do you hunt for disconnects between value and price? It's hard when you're a quality investor to find major disconnects in the market because quality companies generally screen really well: They tend to have high return on equity and low debt. So typically we're happy if we pay a fair price for the companies we like. What's interesting lately is that with a lot of the quality businesses we're finding, we're not having to pay as much of a premium as in the past. I think that's a function of the excitement, particularly in small-caps, on more speculative companies and businesses that are directly tied to AI. Our nice little industrial companies, our nice little financial businesses, just don't have the same sex appeal.
Can you name a couple of businesses that fall under that description? I'm thinking of a company like Andersen Group, which recently went public. All they do is tax advisory, where they're working with ultrahigh-net worth clients who want very specific, complex tax strategies. So they don't have any conflicts with audit. It's a very high-touch business, and it doesn't really have a lot of AI exposure or adjacency. It would be difficult to feel confident that AI could put something together that would stand up when it comes time to file.
I think of a company like iRadimed. When you're in an MRI, there can't be any metal in the room. But if you are a critical-care patient on life support, you still need to get those intravenous medications. So the hospital has a choice: They can either run superlong tubes out of the MRI suite, or they can buy an Iratimed pump, which has nonferrous metals that won't attract the MRI. IRadimed is completely alone in the market with this pump. Part of the reason is the technology is complicated and expensive, but the market isn't very big, so it's not very attractive to pursue. They've released this new pump and are seeing a lot of good activity as a result. It's this quiet little company, not really connected to AI. The growth isn't 50%, but we see good durability for the medium term in the low double digits.
How much AI-related investment opportunity has been identified and evaluated by the market, and how much is a wide-open frontier? I think most of it is still a wide-open frontier in terms of understanding how this technology is best used. What I notice in terms of our own AI usage at the firm is that there's a real learning curve for using these tools effectively. It's almost like learning another language. I'm starting to recognize that in places where it disappoints you, when it hallucinates for example. Yesterday I asked it to compare a company's financials, and it said, "The take rate of this business in FY '23 was 52 basis points, and it's since declined materially: Last quarter it was 53 basis points." I was like, what?
The minute you get something like that, you question the quality of the rest of the analysis. I personally think the tools are wonderful. They have specific use cases where it's easy when you get that kind of an error to fix it on the go. What's tough about it is that the companies selling AI technology themselves have the best use case right now, which is software development. They're like, "This technology is life changing." And yes, this technology is life changing for some. But not every business has the type of use case the AI is ideal for.
It's hard to gauge AI's long-term implications. Think of the dawn of the spreadsheet. If I told you then that it could do the calculations that would take your accountant a really long time, what do you think your outlook for accountants would have been in the next 20 years? It'd be easy to say there's going to be far fewer accountants in the future, but that didn't turn out to be true. AI technology is so much more powerful and meaningful than spreadsheets that I think it will be very hard to identify which sectors will be hit, which companies will be disrupted, and so on. It's still too early to do that because we don't have a great handle on what is the highest and best use of this technology.