Cybersecurity in the Age of AI & How AI Can Fix the Workforce

0:00Joining me right here on the show on Delay Bastion, Palai is with me, senior client, portfolio manager at Piquetay Asset Management. I'm so glad you're with us because we're looking at what's going on with AI and is it shifting to something new. How is the shift? How would you describe it? Because we saw some of that from the GTC conference in particular. Absolutely, I think it's shifting towards these kind of advancements in tokens and how much token can be processed per watt. I'm sure you heard about that term, the token per watt. The advancements are really coming in all this sort of build-out in AI infrastructure, which is still at its early innings. We think there's still a lot more to go, clearly now with all the power you have in GPUs and all these advanced accelerators. You're hitting this memory wall, so great opportunity there to expand on what's coming next in terms of expansion. But I think we see three layers, if you want. There's the whole kind of AI infrastructure build-out with semiconductor advancements and the chip complexity. The second part is obviously the software space with the pullback and clearly some clear winners there.

1:05I mentioned to you earlier the cybersecurity space. And then finally I think what's sparring all these AI infrastructure build-out that big capex of 650 billion is ultimately to basically have all this software and what we call physically AI also being powered by that AI build-out. So we see these third and fourth layers of that S curve becoming more and more important and interesting in terms of advancements from that build-out. Because we were talking about shifting away, shifting from height to hard infrastructure and the applications for AI that go with this. What would be a great example of that? I think a great example, especially with the software pullback right now, is cybersecurity with what's happening in Japan, deeply in Iran. The fact that today one human equals one, sorry, in the past one human was one identity. Today one human is over 90 identities. So basically with anthropic and obviously all the new apps or the new AI labs if you want, Agent TKI is advancing so fast that it's very hard to protect the identity.

2:08So enterprises today need to really increase that spend, but also the products being sold by cybersecurity need to be more advanced to take it together to Agent TKI. I think that's one area that I like to highlight because that's not going anywhere but up in terms of demand. I think the second part also is the physical AI side, I think how the adoption of humanoids, of robot axes, even of what you see in the industrial application have advanced a lot. I think Morgan Stanley just had a report out that by 2050 there'll be over 1 billion humanoids. When you look at Amazon today, there are more than 1 million robots. Oh my goodness. So this is automation at a different level. So yeah, and you have an AI automation ETF. Tell us about that. So the AI and automation ETF is all about combining those three categories that I mentioned. It's really building out the automation and automation processes. So in LLM first, it's a general purpose technology.

3:10What you automate from a robotics or from a different AI process and monetize that, that's what we're looking at. And I think in AI and automation, what you find is that generally around the world you have an aging population. You have a shortage of labor, you have a labor cost that's going up. And the only way, especially when you're re-showing manufacturing to the US or building these big fabs out in Arizona for TSMC, you're willing to automate all these processes as much as possible. And so we're really like investing in that full value chain of the AI automation sort of theme. So I mean, obviously people have been worried about AI bubbles or check out a favor. And you know, any names and a lot of ETFs related to this have pulled back. Does this present an opportunity right now because you look at some of these names. Are they even sort of value names in your opinion at this point? This is a good point. I think the pullback has been very attractive first. I mentioned in the cybersecurity space, obviously in some enterprise software names as well,

4:11which have really been indiscriminately beaten down by the market. I think really looking at fundamentals, the long-term business model, their earnings potential, even their revenue growth, X growth, like how much are they generating today in terms of cash flow? That's super interesting for us. But I think what we look for is the long-term investment thesis. Where do you see this market volatility sort of reducing maybe evaluation and the stock price? But long-term when you have that pullback, which are the names that you want to be in? In a way, we're unconstrained, which is probably what I should mention from the beginning. We don't need to build a portfolio based on a benchmark. And that allows us to find these winners in the AI space as and when they become more relevant, obviously I mentioned memory in the space. Super attractive right now, right? No matter what's happening in the pullback, that's going to do well. There's a shortage of land, DRAM, and high performance, HBM.

5:11So those are the areas that we like to focus on. But ultimately, I think we're looking at where this AI curve is going in terms of really producing either software for automation, industrial or enterprise, but also on the physical AI side. What are you producing that's automating a whole industrial process or a whole consumer process? Yeah, and just quickly, I mean something bullish or some sort of tailwind, you were just mentioning certain sectors, particularly, that are doing better than others. I don't know whether it's supply chain or the revolution innovation, the applications actually being put to use finally monetization. What would you say is a great tailwind in your mind? I think a great tailwind right now is really the infrastructure buildout and that capex spend that's actually really to power a lot more of that software that's going to come, a lot more of those humanoids, robotaxies, all that technology.

6:13That's really the opportunity further down the line. But that AI buildout is really critical in order to be able to have all these like, we're in an AI era really, we've gone from cloud to AI. And I think in that AI era, we see multiple business models which will come. But at the same time, that capex is really important. Well, thank you. We'll keep an eye on Peabot, too. The ETF, thank you so much. Thank you. Thank you. Thank you so much. Thank you. Peabot, I'm glad you were with us today.