Do AI Moats Exist?

Travis Hoium (0:02)

Do motes exist in AI? Motley Fool Hidden Gems Investing starts now.

Welcome to Motley Fool Hidden Gems Investing. I'm Travis Hoium, joined today by Lou Whiteman and Matt Frankel. Guys, it's kind of an interesting time of the year because we're at the end of the calendar third quarter. Earning season hasn't begun yet, but there is a lot of moving and shaking in the world of artificial intelligence. A couple of IPOs that are potentially coming in the next month to six months with Anthropic likely to be first. I wanted to talk about if we're looking at some of these companies, let's take a big picture approach. One of the questions that I have is, do motes really exist in artificial intelligence? If you look at the history of investing, finding companies with motes is a phenomenal way to make money. In the 80s, motes were driven by supply and big brands. Just buy the biggest best brand you possibly can. The retailer with the most customers, you're going to do phenomenally well as an investor. Aggregators won the demand platforms, Amazon, Netflix, Uber in the 2000s and the 2010s.

But Lou, do motes exist in the world of artificial intelligence? Because it seems like one exists for five minutes, and then another model is released, and everybody moves over to ChatGPT, and then, oh, Muse is released. So now everybody is trying out Muse. It seems like these are incredibly fragile motes, or do you see it differently?

Lou Whiteman (1:27)

So Travis, you should have given me a trigger warning here, because this is a hill I'll die on. I'm a terrible person to ask this in part, because I think 99% of the discussion of motes in finance is just balderdash. I don't think motes really exist. If they exist, they're very rare, and as an investor, they're really only recognizable in hindsight. I don't think in real time, there are plenty of examples. There was a famous 2014 book about the value of motes, and you look through, there's 20 or so companies that are mentioned in that.

Half of them are out of business now. Okay, so, you know, everything... I mean, all we're really talking about is competitive advantage that gets to a point where you can't overcome it. I don't see any motes in AI, almost by definition, because none of these companies even have pricing power. How can you have a mote? How can you have this idea that, you know, like you are, you cannot be conquered when you can't even control the terms with your customer on pricing?

I don't see any permanent competitive advantages here, and I don't anticipate that happening anytime soon.

Travis Hoium (2:32)

Matt, do you see it the same way?

Matt Frankel (2:34)

I'm a little less, you know, pessimistic about the motes in AI than Lou. I mean, but the motes definitely appear fragile. So, I mean, what I mean by that is a couple of years ago, when ChatGPT clearly had the first mover advantage. A year ago, they had a 79% market share when it came to just, you know, query volume. That was to start 2025 Now it's down to 54%. That's 25 percentage points of market share that's gone. In the same time, Claude went from roughly 1% to 9%.

And the capability is not the mote. That's one place where Lou and I will clearly agree. All of the frontier AI models are within a little, you know, a narrow band when it comes to capability.

There's a couple of true motes that I see. I don't know if anyone's captured them completely yet.

One is capturing enterprise workflow, like being ingrained in workflow. It's the same reason that some of the software companies that, quote, could be disrupted by AI are still growing their revenue faster, not seeing too much customer churn. It's because they're embedded in workflow. And Claude Code is a big example of that. I use Claude Code work every day. I'm not going to switch that to a more capable model because I already depend on it.

The other one is distribution that I see. Think of Gemini. Gemini usage is up 450% over the past year. And why? Not because they have pricing power, or not because of anything of that, because Gemini is ingrained into everything Google does. It's being force-fed to everyone who uses Google Maps, who uses Gmail, who uses Google Search. So distribution is a big competitive advantage. I don't know how durable of a mode it is, but that seems to be something that could have some staying power here.

Lou Whiteman (4:24)

Yeah, as impressive. It's the same reason we all use Bing, right?