Host (Remy): Today's capital markets segment is brought to you by Alpaca. Wall Street remains glued to the boom, but a growing debate over revenue accounting and opening and anthropic has investors questioning what is real growth and what is portfolio posturing beyond the accounting noise a fierce war is being fought across the entire six layered AI tech stack from hardware and infra to end user applications joining us to break. All of this down for us regarding the AI economy and enterprise decision making is Eddie, co-founder of Category Pirates. Eddie, good morning. Thank you so much for joining us. Well, we're continuing to keep an eye on artificial intelligence, but for our viewers out there who may not be familiar with what you call a six-layer cake, give us a take on what those layers are and what you're focused on right now and why.
Eddie (Co-founder, Category Pirates): So, um, in the six layer cake, what you want to think about Remi is on one level of the spectrum is more hardware type stuff, so PowerShell, chips, things of that nature. And on the other end is more software, all the way down to the consumer apps with the models in the middle. And the other way to think about it too is more of a B2B spectrum, and then all the way to a B2C spectrum. And so, that's a simple way to think about it, and that in the last 4 eras of technology, what you've really seen is that usually the, the one side of the layers make wind early. So the hardware and the B2B side, and later on, it's the consumer side and the software that wins. Now, what's different about this time around is that in the AI section, the hardware might actually be the ultimate winner in the long run, because never before in history has software become a self-fulfilling product where AI can create software, which is traditionally where all the money's been made. So, this is a different phenomenon than the last 4 kind of eras of call it social and mobile, and the internet, and all the way back to uh databases.
Host (Remy): Yes, and Eddie, this week our eyes are on earnings as well as economic data, but key IPOs do remain in focus for us. And of course when we're talking about upcoming IPOs we're all focused on open AI as well as anthropic. So given the concerns about net revenue as well as how anthropic accounts grow sales. Direct comparisons may not be possible. So are these confusing sales numbers just tech executives trying to hype up their own portfolios, or how are you looking at these numbers as well as expectations?
Eddie (Co-founder, Category Pirates): Well, I, I, I think the, the net revenue, gross revenue issue was probably just a mistake in the reporting of it, and some confusion around people who are happy to, you know, profit from the volatility in the stock market with it. Um, I, I actually think that more than one thing can be true, is that both, uh, the upcoming IPOs of anthropic and open AI are gonna represent some of the most explosive growth in revenue that we've always ever seen in the history of business, and that that rocket ship may not last until they expand more in the six layers of the cake, Remy. So this is the part that's really interesting is that historically, you've been able to win and be successful in one layer of the cake. But the people who have kind of call it more spots on the bingo card do far better in the long run than not. And so you think about a company like Google, who participates, um, they're making, they're starting to make their own chips, they have their own cloud infrastructure, but they are also participating on the consumer hardware with Android, all the way down to YouTube, and search. And, and, and obviously with Gemini, they are participating in multiple layers of the cake, and that for the ultimate uh upstarts that come in, the open AIs and anthropic, they can't just stay in one layer on the model side, they're gonna need to expand to others, and that's where you see open AI bringing on a Johnny Ivy to try to do some kind of consumer hardware. And so, what you're gonna see, it's a bit like, you know, a boxer competing against a mixed martial artist. If you can't punch, kick, and wrestle at the same time, you're not gonna do well in a match.
Host (Remy): Well, that's an interesting analogy right there, so I do want to get your take on some of the software names speaking of volatility because earlier this year we added a new term to our lexicon which was Sask Apocalypse and uh some did predict legacy software giants like. Force would get crushed by artificial intelligence, but when we look at the year to date performance for some of those key names, that is not the case. So why is it so hard for some of the AI startups out there to actually steal real business away from the established tech names?
Eddie (Co-founder, Category Pirates): Well, I, I think what you're really seeing is, uh, the explosiveness of the growth in some really mission critical sectors of the economy. So, where AI has done fantastic, is where there's really kind of primarily numbers that are the, the medium that are being worked with, and there's not a human that's required to validate it at the end. So, you know, from a coding standpoint, you have um claw code, you have, uh, you know, codex. These are things that the AI can do recursively, like on a self-improving loop, and you hit compiler and you see if the code works or not. Whereas, when you talk about a company like Salesforce and the other SAS players, it's actually multiple levels to actually get anything done, right? So, when you have a company that isn't uh using CRM you have three kinds of things. You have data, you have the decision-makers, and you need um somebody's throat to choke in the event that somebody goes wrong, damage control. as it were, right? And that when it comes to actual big business decisions, say, should I raise my price of Bud Light? Should I increase my marketing of Kraft cheese? These are things that you cannot just make a decision without only the data, you have to include some humans, oftentimes many humans in the loop, and that there's a lot of damage control that has to be worried about. And so, in the course of business, where you're seeing AI take real root is when those 3 IDs are just far less of an issue in terms of the natural course of business, and AI can work much faster through it. When you have other parts of the business, and that's a huge part of the economy still, that's gonna just take some time that you will get great data, but humans are always gonna be in the loop in certain businesses and certain decision making. And you're gonna need to hold somebody accountable at the end of the day, and that's what's gonna slow things down in a way that's not as fast as coding or mathematics.
Host (Remy): Yes, and speaking of what you just mentioned, damage control, and we can think of many instances in 2026 when we saw some of these incidents. But when an artificial intelligence system actually makes a costly mistake, whether we're talking about a public or private company, who actually takes the fall and tell us a little bit about the one throat to choke need and why humans will never be replaced in the actual boardroom.
Eddie (Co-founder, Category Pirates): Yeah, it's an important distinction because when it, when there's a difference between call a frontier lab, it's kind of an R&D space versus Uh, and engineering a product and running a corporation. Uh, if you remember, if you go back a couple of decades back, when Tylenol was laced with cyanide and 7 people died, what did James Burke do as the CEO of the company? He pulled every single bottle of Tylenol off preemptively, even though the FBI and the FDA said, you don't have to do all of that. And he did so because he knew that he was the throat to choke, and that he was gonna take accountability and ownership of it for the public safety and good. And that when it comes to this kind of a thing, this is a normal course of business. If you are Campbell's Soup, you make sure that your soup uh is safe for people to eat. If you are a General Motors, you make sure your cars are safe to drive. This is kind of the normal course of business, and that what some of the um uh situations that we had where AI seems to be running amok, that's usually because the sandbox and the safety measures weren't taken. Appropriately seriously enough. And in some cases, they were attempting to break it. It's almost as if like, imagine a car company that was like testing the airbag safety, and they were being more reckless and reckless without actually doing the dummies and everything else in the loop. You have to take precautions, you have to make sure that there are safety guards in place, and most importantly, you have to take accountability for when something goes wrong, just like any other product in the economy, you are held responsible for it, and I think that's what's coming happening here.
Host (Remy): And finally, Eddie, before I let you go, I do want to get your take on CapEx because that is something all of us on Main Street as well as Wall Street are paying attention to when it comes to some of these hyperscalers, and it goes without saying that big tech has been spending hundreds of billions on AI, especially data centers, but they can always rent that infrastructure out to other companies. So do you think that safety net gives the hyper scalers an advantage over, say, pure play AI startups?
Eddie (Co-founder, Category Pirates): 0, 100%. Um, 1 of the things that I, I was listening to in a podcast with Alexander Wang, the chief AI officer at Meta. One of the things he mentioned about what Zuckerberg said was that, hey, um, I'm not, I'm a little different than some of the other kind of, uh, AI executives and founders out here in that I, what I'm great at, I'm building, building teams, he's great at. Uh, number 2, he's very patient in terms of optimizing product relentlessly over and over again. And 3, I never have to raise cash again, right? So, Meta is a cash machine, Alphabet is a cash machine. If they appoint appoint those cash flows appropriately, they will be able to self-fund or generate debt uh in order to do so in a way that uh Open AI and anthropic really don't have that same luxury. Again, it goes back to being one single player in the six-layer cake. If you have multiple players, some of the layers generate cash to fund the other ones. If you just play in one part of the uh layer cake, you're not gonna do as well, and you're gonna be at the, at the behest of the capital markets. And that's why these companies are trying to get out as quickly as they can without having a busted IPO.
Host (Remy): Well, Eddie, I appreciate your time this morning. Thank you so much for joining us and thank you so much for sharing your perspective as well as all of your insights.
Eddie (Co-founder, Category Pirates): Thanks, Remy.
