Do AI Moats Exist?
In this episode of Motley Fool Hidden Gems Investing, Motley Fool contributors Travis Hoium, Lou Whiteman, and Matt Frankel discuss:
- AI moats.
- Fragile competitive advantage.
- Valuing AI stocks.
- Metrics to watch.
- Stock opportunities.
To catch full episodes of all The Motley Fool's free podcasts, check out our podcast center. When you're ready to invest, check out this top 10 list of stocks to buy.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
A full transcript is below.
Should you buy stock in Modine Manufacturing right now?
Before you buy stock in Modine Manufacturing, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Modine Manufacturing wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $379,123!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,396,103!*
Now, it’s worth noting Stock Advisor’s total average return is 933% — a market-crushing outperformance compared to 212% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
*Stock Advisor returns as of September 30, 2026.
This podcast was recorded on Sept. 16, 2026.
Travis Hoium: Do moats 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. It's 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 moats really exist in artificial intelligence? If you look at the history of investing, finding companies with moats is a phenomenal way to make money. In the '80s, moats were driven by supply and big brands. 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 2000s and the 2010s. But Lou, do moats exist in the world of artificial intelligence because it seems like if one exists for five minutes and then another model is released, and everybody moves over to ChatGPT, then Muse is released. Now everybody's trying out Muse. It seems like these are incredibly fragile moats or do you see it differently?
Lou Whiteman: 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 moats in finance is just balderdash. I don't think moats really exist. If they exist, they're very rare, and as an investor, they're really only recognizable in hindsight. There are plenty of examples. There was a famous 2014 book about the value of moats. You look through, there's 20 or so companies that are mentioned in that. Half of them are out of business now. 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 moats in AI, almost by definition, because none of these companies even have pricing power. How can you have a moat? How can you have this idea that 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: Matt, do you see it the same way?
Matt Frankel: I'm a little less pessimistic about the moats in AI than Lou, but the moats definitely appear fragile. What I mean by that is a couple years ago, ChatGPT clearly had the first-mover advantage. A year ago, they had a 79% market share when it came to just query volume. That was at the start of 2025. Now it's down to 54%. That's 25 percentage points of market share that's gone. The same time, Claude went from roughly 1%-9%. The capability is not the moat. That's one place where Lou and I will clearly agree. All of the frontier AI models are within a little a narrow band when it comes to capability.
There's a couple of true moats that I see. I don't know if anyone's captured them completely yet. One is capturing enterprise workflow, being ingrained in workflow. It's the same reason that some of the software companies that could be disrupted by AI are still growing their revenue faster or not seeing too much customer turn. It's because they're embedded in workflow. Claude Code is a big example of that. I use Claude Cowork 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. Why? Not because they are pricing power, not because 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. Distribution is a big competitive advantage. I don't know how durable of a moat it is, but that seems to be something that could have some staying power here.
Lou Whiteman: It's the same reason we all use Bing.
Travis Hoium: Bing it up. Well, Lou, that does bring me to my other question, which is, when we look at this space, what seems so different than in the past, whether you're looking at consumer packaged goods or retailers or even just the last 20 years in technology, that was all companies. The companies that ended up winning ended up beating the competitors that they had, but then they weren't competing with also all the other companies. The banana company isn't competing with the cereal company. But now we have all of these tech companies, all of them in some way competing with each other in the future of AI. You have Anthropic takes the lead early this year, and then suddenly out of nowhere, Meta, who we thought was done and dusted, is now potentially as the most compelling product release of 2026. It seems part of the challenge from an economic investing standpoint is that they're all chasing after the same park. In that world, margins and moats seem really tough to build.
Lou Whiteman: These are all companies that built their existing business basically not monopolies, but being the dominant force in an industry. This is not something they've dealt with. But I do think it's interesting. I made the joke about Bing, but I'm serious with that because, yes, right now Gemini is the thing for consumers because none of it is really that good. We're just using what's there. If this Meta Muse product is good, that's a reason to use something different. I don't think Gemini will have a moat at all then if there's actually a compelling case, just like Chrome, despite its disadvantage, took over the browser wars just because it was a better browser.
Maybe I just have too high of a standard of moat. If we're just talking about do some companies have competitive advantages in a moment? Of course, they do. That's how capitalism works. But this idea that there is something to invest around, I can trust this company because they have a competitive advantage that's strong enough that I don't have to worry about it, that is so rare in business. Just the definition of capitalism is that we see that these are rare. I just think we spend too much time to talk about moats. As an industry, it's a very young industry, and to your point, two years ago looked very different than today. I don't even really know who has a durable competitive advantage right now.
Travis Hoium: With those questions and durable competitive advantages in mind, we're gonna talk about valuations in just a moment. You're listening to Motley Fool Hidden Gems Investing.
Welcome back to the show. The second question I have for you guys today is, how do you value these AI companies? The context for this is we had the SpaceX IPO, so that includes XAI and all the stuff that was involved in that. We found out in their perspective or in their S-1 that they're actually mostly an AI company, not actually mostly a space company, which is, I think, a little bit ironic. But Matt, as you look at these companies, we've seen market share ebb and flow. We've seen margins and pricing ebb and flow. With that in mind, how do you value some of these companies, especially as they come to market? When the Anthropic S-1 becomes public, how do you look at that and say, Here's where I would price this company to see a value, or where I would maybe stay away?
Matt Frankel: The short answer is, I don't know how to value these companies, and neither does anybody else. We haven't seen a public S-1. Anthropic uses different revenue calculations that they're revealing to the public. They're using gross revenue, not net revenue, just for example. It's tough to value a business when revenue is growing at a run rate from $9 billion at the start of this year to $65 billion as of the end of July. That's Anthropic. But like I said we haven't seen the public S-1. We don't know what accounting they use because we haven't seen the public S-1. We don't know what percentage of their revenue is enterprise versus consumer. That makes a big difference in the valuation.
Enterprise revenue is generally higher margin and stickier. How much money will they need? How much money are they making? We can only speculate how much capex Anthropic is going to need this year. We can only really speculate how much they're losing right now. Other than to say lots and lots. These play into the valuation; there's simply no comparison. We don't have a basis for comparison to a public company that is growing this fast, that is at this scale already. SpaceX isn't growing this fast and at this scale. I'll be looking at the S-1 very closely. Once we have real numbers, I might have a better answer for you as to how to value Anthropic and OpenAI when we eventually get theirs, but it'll be really interesting. That's one of the S-1 I'm really going to be paying attention to
Travis Hoium: Matt, is there one number that you would look at? I want to ask Lou about comparisons to the '90s. But in the '90s, you would look at metrics like eyeballs. We're not doing that anymore, but is revenue the number that you would look at and say, "They're growing revenue at 800% year-over-year, and it's at a $65 billion run rate, so therefore put a reasonable multiple on that, and that's where I would maybe be interested." Is that a good number to look at, or is there something on profitability?
Matt Frankel: Going back to the moat conversation, revenue retention is a number that I would really look at. How well are these companies retaining revenue? Is someone paying $20 this month for Claude, $20 next month for ChatGPT, and so on? Are they turning revenue really fast? I want to see if the unit economics are getting better. As we said in the moat discussion, none of these companies have established pricing power yet. As these companies scale exponentially and have to build out their compute and things like that, are the unit economics getting better or worse? You can see that by just looking at the gross margin trend over time. Between revenue retention and gross margin trends, you'll get a really clear picture of the real situation with these businesses.
Lou Whiteman: Back to your original question, how do I value these companies? The answer is less. Like with Matt, I don't know how much less because we haven't actually seen the numbers. I don't think either of the companies seem to want to put forward the audited numbers, which says something, but less than what we're hearing. I'll tell you why. It's not to say they're without value. Both companies have massive amounts of value, but we are talking about numbers that would place them among the top 10 most valuable companies in the world. I will buy the argument that one day they could be. I will accept that. I might even agree. But my guess is all of these companies will not be among the most valuable in the future. Some will fail. Some will be seriously second best. I am not willing to assume mega growth for generations to come from any of these people when I don't even know who the winner is. There's a there there. This isn't eToys level hype, but we never tried to make eToys a trillion-dollar company.
Travis Hoium: Even some of those. You go back to Pets.com. We talked about this on the show a few months ago, but market cap was shockingly small for as much mind share as it got from the dot-com bubble. These are some of the biggest companies in the world. Lou, final word on this, is this an area where it's just OK to stay away for a while in any of these AI-related names until we get better answers on what the competitive advantages, as you might call them, are going to be in the industry?
Lou Whiteman: Anything is a good deal at a price, and maybe if we one day see these numbers and they're better than I think they are, then maybe I'll reconsider. But right now I don't think we know enough. Matt was talking about it, but look, apparently, what they're doing right now is they are reporting revenue collected, not revenue that goes to them. If I am generating $10 billion in revenue, but $6 billion if it's going out the door to affiliates and stuff like that.
Travis Hoium: Let's explain what that means if you're not familiar with this. For example, Uber is a good example. If you pay $20 for an Uber ride, Uber does not count that as $20 in revenue that is $20 in what they call gross bookings. The take rate, which is usually about 25%, is going to be Uber's revenue, so $5 in revenue, $20 in gross bookings. That's the way that they would report it. What you're saying is that Anthropic, at least from what we know right now, is saying we have $20 in revenue.
Lou Whiteman: From what we know, and again, it would be really nice to see those audited financials, wouldn't it? But until we know more, I don't think it's really a slam or a hot take to say not all companies are worth trillions of dollars. But right now we are being asked to assume that all of these companies are worth $1 trillion. I think we're allowed to be a little patient and let that simmer.
Travis Hoium: It'll be interesting to see with the two what the obligations are for these companies, because those are measured in the hundreds of billions of dollars in cash that they do not yet have. When we come back, I'm going to actually get some stocks out of Lou and Matt. Where are they interested in in the AI space today? You're listening to Motley Fool Hidden Gems Investing.
ADVERTISEMENT: You just found out that your sales team is at risk of missing quota. Don't panic. Just ask Rippling AI. Since it's built on your real time people and business data, Rippling AI can pull metrics from rippling and Salesforce into a meeting ready dashboard showing quota attainment, headcount plan, and monthly revenue to quota by region. In seconds, you'll see exactly what's behind your quota risk and fix it before it's missed. Question answered. Action taken, crisis averted. When you have critical business questions that need answers, don't just file a ticket and wait weeks for an outdated report. Describe what you need and have Rippling AI build it instantly from your live people and business data, whether it's a dashboard with detailed charts or automated workflows with the right triggers, conditions, and approvals. Ready to rule your business, head to rippling.ai/fool to get the only AI built to give you full visibility and take complex actions across your entire organization. Sign up for exclusive access today rippling.ai/fool.
ADVERTISEMENT: Legendary flavors from Legendary Xbox games are now at McDonald's. Crispy Golden Garlic and black pepper McNuggets. Cheesy potato bites from Fallout 76. Candy Crush McFlurry. The main mission can wait. Take a side mission to McDonald's. Available until the 29th of September from 11:00 A.M., while stocks last, subject to availability. Participation varies.
Travis Hoium: Welcome back to the show, guys, as you look for opportunities in the world of artificial intelligence. Today, what kind of stocks are you looking at right now?
Matt Frankel: There are a lot of these AI infrastructure stocks that have cratered over the past month. Some of them did so after reporting great earnings, raising guidance, things like that. But even after that, there's a lot that are very expensive. I'm sure you and Lou would both agree that I'm not interested in a power company that's trading at 100 times revenue or something like that. But there are a few that are still really attractive investments. Modine Manufacturing is one that I have my eye on right now. Ticker symbol is MOD. If you've ever heard of Comfort Systems, that's a more popular pick-and-shovel play. They do almost the same thing, but the valuation just makes a whole lot more sense to me.
They specialize in thermal management, specifically cooling, and the Data Center business is not surprisingly the most exciting part of the company right now. Overall, their revenue grew 23% year-over-year in the latest fiscal year. The data center revenue grew 90%. They just got a $4 billion multi-year deal from one of the hyperscalers. We don't know which one, but it's for cooling products and data centers. It's a major problem. The more complex data centers get, the more cooling they're going to need. Unlike some of the other companies in the picks and shovels space that are really expensive, Modine trades for 22 times forward earnings. They expect 50%-70% annualized data center revenue growth over the next couple of years. Most of that's already booked. They're guiding for 20%-35% overall sales growth in the current 2027 fiscal year. More than 40% growth in adjusted EBITA, remember, there's a lot more to this business than data centers, which is if you're skeptical about how long the AI build-out is going to go on at this pace, this could be a good one to look at because it is a truly diversified business, even though obviously the rest of it isn't growing as fast as the data center side.
Travis Hoium: Sometimes boring is good. Lou, where are you seeing opportunities?
Lou Whiteman: I am less interested in pick-and-shovel game right now because we are late in the cycle. Picks and shovels sound great, and they are a clever way to get in on a trend after the primary beneficiaries have already benefited.
Travis Hoium: Second derivative, if you will.
Lou Whiteman: If the first derivative is too expensive, this is a backdoor way in. But right now, why buy the pick and shovels when the actual companies that they serve are trading at a discount to them? Matt's right. Modine is pretty cheap for a pick and shovel, but that 22 times earnings, compare that to Alphabet. Why not just buy Alphabet? As boring as my answer is, my answer would be the Nvidias, the Alphabets, the Microsofts, mostly because they have multiple ways to win. It's not an all-or-nothing AI bet. They all have other businesses. Even if AI isn't what we think or even if their AI isn't what they think it is.
Travis Hoium: It seems like the worst-case scenario is we just slow our spending down a little bit, and we have more cash.
Lou Whiteman: We still have this other great business. But look, here's an even easier way. People have been stressing out about there's too much concentration in the S&P 500, that it's too heavily weighted for these AI hyperscalers. But if you are looking today to invest in AI, that's an advantage, not a weakness. Just buy the S&P 500, and you'll get plenty of exposure to AI without having to take single-party risk not early in the trend. I think, honestly, the S&P 500 might be a better choice than anything.
Travis Hoium: Well, I'm excited to see. It could be in the next week or two that we get the S-1 from Anthropic, although it seems like it keeps getting pushed back, but then we'll have more to talk about what the numbers look like, what the potential moats or competitive advantages look like, but lots to discuss here.
As always, people on the program may have interest in the stocks they talk about, and The Motley Fool may have formal recommendations for or against, so don’t buy or sell stocks based solely on what you hear. All personal finance content follows The Motley Fool’s editorial standards and is not approved by advertisers. Advertisements are sponsored content and provided for informational purposes only. To see our full advertising disclosure, please check where they're shown off. For Lou Whiteman, Matt Frankel, and Dan Boyd behind the glass, I'm Travis Hoium. Thanks for listening. We'll see you here tomorrow.
Lou Whiteman has no position in any of the stocks mentioned. Matt Frankel, CFP® has positions in Amazon. Travis Hoium has positions in Alphabet and Uber Technologies. The Motley Fool has positions in and recommends Alphabet, Amazon, Comfort Systems USA, Meta Platforms, Microsoft, Modine Manufacturing, Netflix, and Nvidia. The Motley Fool recommends Uber Technologies. The Motley Fool has a disclosure policy.
