David Belle argues that software stocks are oversold because investors misunderstand AI disruption. Companies with strong moats like CrowdStrike (proprietary threat data) and RELX (legal data courts must cite) will survive where others won't. The sell-off creates buying opportunities in fundamentally sound businesses.
Has everyone literally gone insane? There is so much pessimism over AI and software stocks right now that I literally cannot fucking believe my mind whenever I log onto X or even LinkedIn to see some of the drivel coming out of people's mouths. There was a piece just released over the weekend by a guy named Citrini, who, to be fair to him, is a very, very good analyst, and I don't necessarily have a good or bad opinion about the piece itself. What I do have a negative opinion on is people's perceptions and interpretation of how AI is actually gonna disrupt businesses. Now, don't get me wrong, there's definitely some businesses that are ripe for disruption by AI, but there's one simple fact as to why they're ripe for disruption. And these businesses would've been ripe for disruption whether AI came in or not. The primary factor here is that many businesses simply do not have a big enough moat. And towards the end of this video, I'm going to highlight one business which I think has a massive moat, which has been affected by Claude introducing some legal tools. And I just don't think it makes sense for the price to be where it is right now. But firstly, let's have a grasp of the lay of the land. You can see here that this software and services ETF has been absolutely smashed over the last 6 weeks. Claude seems to be rolling out tools, or Anthropic seems to be rolling out tools at a really, really fast pace. But there's some ignorance over the actual effects of rolling out these tools and what these businesses are specialized in. One example is CrowdStrike. Now, CrowdStrike has been hit just over the last few days because Anthropic said that they had released cybersecurity tools. What many are forgetting, and it's perhaps because they don't have domain knowledge, and neither do I, to be fair, but I go out and seek it. What they're forgetting is that a cybersecurity firm isn't just based around the specific front-end tools. No, what sets CrowdStrike apart is things like the size of their datasets, which Cloud simply doesn't have access to. They have billions, trillions of different data points to be able to facilitate better analysis of exploits, of different key security risks. And Claude simply cannot do this. And the other thing is that where does the point of responsibility lie if something does go wrong? In a court of law, you wouldn't be able to say, oh no, you know, Claude is the one who actually facilitated this exploit, maybe being more, I don't know, prescient or more observable. That just doesn't hold up. So something key that I think people are forgetting about with AI, is that there has to be a point of responsibility. And when you've got a human there, that tends to be what people are actually compensated for. Let's say you're head of security or head of risk. You're compensated highly because you're taking on the responsibility of doing a good job to ensure things don't go wrong. An LLM is not taking on any responsibility. And so you've got a key risk there, I guess legally, as to, okay, if something does go wrong, where do we actually go with this? Do we sue Anthropic? Do we sue OpenAI? Do we sue Google? It really, really becomes a legal sticky ground. One of the most interesting things about the kind of AI disruptive sell-off is that a lot of people just seem to completely ignore this domain knowledge. And that is a real, real problem for me. But that creates serious opportunity. I've just mentioned CrowdStrike. I think it's a strong value play. Maybe not right now because we're in the midst of a few different key narratives. You know, the, the, the Trump tariffs are coming back into play. We've got potential war in Iran kicking off. But I really, really think that opportunities like this or risks like this tend to create those opportunities. But one massive opportunity I think is in a pretty underspoken about company, which is actually LSE listed. and the company is RELX, and that's their ticker as well on the LSE. Now they own LexisNexis, and effectively they own all of the copyrights to any court cases. I think I've got that right anyway. Now this is me not understanding the domain knowledge, but I asked my lawyer mate, and so he's told me a few things, but effectively they are the source of truth for legal cases. Now the other interesting thing about this is I mentioned legal previously with regards to CrowdStrike. This is exactly the same case because in a court of law, you can't reference an LLM. What you can reference is that single point of truth, and that is what Relex does. Now, over the last couple of months, their share price has been battered purely because people are believing that AI can take over Relex's job. But what their ignorance comes down to here is that they've forgotten their key moat. They've forgotten that Relex is actually legally that key point of truth. And this is something that AI can't fight. And where I'm looking at opportunity now is really in, okay, where are these key moats to do with regulation, to do with, you know, AI not being able to overcome a task because it simply cannot. And I'm not talking about utility here. I'm talking about structure, legal, and regulation, which certainly are intertwined, but I think they are also considerably different. But above all of those three, I think the domain knowledge is absolutely key. See, people outside of the domain, as I keep saying, keep thinking that all of this knowledge can just be rolled up into one giant LLM and then produced out. When LLMs don't really work like that, they work like a sort of chain. They work on predicting the next words or the next task that should be done. And the other interesting thing is lots of people keep talking about OpenClaw. Well, no one has actually shipped anything to do with OpenClaw. They might be able to, you know, process their emails faster, or they might be able to do certain tasks faster. They're not actually making anything yet. Now, I think this will come in a few years. I think things are gonna absolutely amp up and go exponentially, even if you just look at something like Moore's Law. Definitely, you know, things are going to change, but I think this real hype at the moment over companies being completely displaced when actually they have a key moat, and obviously that's my thesis here, is really quite silly. But for investors and traders, it provides great opportunity. So CrowdStrike and Relx are two companies that I'm looking at at the moment.
