Wedbush's Ives: Still less than 10-15% through AI revolution [5:00-10:04]

There could be a company you've never heard of now, where a year or two from now they could be competitors to a Snowflake and others. And that's why competition to spend, to innovate, and this AI arms race continues to play out. But also, you know, I heard your conversation before, for the first time in 30 years, the US is ahead of the child when it comes to tech. And that is just such an important dynamic that's going on here in terms of this innovation why you're going to continue to see more and more global hours going on, you know, US tech stocks. Dan, it's all very well, I suppose, for the AI companies to say there is so much demand and we see this only increasing. It's all very well to say that when people are getting it for free or at lower cost than they will in the longer term. But with Axios reporting, corporately, to starting to question whether they're actually getting value for the amount that they are spending as businesses on AI, to the extent the Verge says Microsoft canceled most of its Claude code licenses, Uber's chief operating officer saying AI costs are getting harder to justify. Is there a potential headwind there on the optimism around future earnings for these providers of AI platforms? Yeah, look, are there going to be companies that could slow down at one point in terms of just giving the cost? That's possible. But right now, I can tell you from all the checks that we do, talking to enterprises, CIOs, product meant just across the board, it's accelerating, which is ultimately driving catbacks, it's driving the data layer, it's driving the infrastructure. And I think when you look at ultimately the use cases, which are going to be software and data layer driven, that's going to give companies the ability in terms of the monetization side, which is why I continue to think it's just tug of war. But right now, it's showing, this is an acceleration, I think that's what we saw this week, I think a lot of the investors, you know, the bears, when they're in hibernation mode, you know, they don't see AI in the spreadsheets. You look around the corner, I think what you saw in Dell, we saw in Snowflake, I think what you're seeing in some of these other names, extremely important. Dan, I know you've got a call on some of the cyber security names from Palo Alto Networks to Crowdestrike. And I want to talk about how the disruption could impact some of these. I can see that Palo Alto apparently used my thoughts as we talk about the technology from Anthropic. It feels like it could be disruptive, but they've been using it to find vulnerabilities of their own system. Apparently, found 75 vulnerabilities in their own products in one month versus typically five to ten. Do you see this disruptive tech as supportive, helpful for their revenue or potential disrupted on the track? Look, I mean, cyber security, I know this is part of the AI ghost trade, and you go back call it in March, but we strongly disagree with it because the surface area is increasing. For every person, there's four, five, seven agents. That's just more demand for cyber security, Crowdestrike, Palo Alto, kind of front and center is the two leaders. And I think we believe from everything budgets could go from 5% to 10% in terms of over the next few years in terms of cyber security. And that's a good example of the derivative beneficiaries, and you cannot get caught up sometimes in negative narratives. I mean, go back a year ago, alphabet, I mean, you know, the one in cab driver was bearish on alphabet. Now, today, look at it. So it's very easy to caught up in narratives, and I just think that creates the opportunities if you're able to see around corners in terms of AI revolution. Damn, I'm bound to ask you a question based on rather glib, but bear with me on it. How intelligent is artificial intelligence at this stage, given that humans, working at Amazon, were able to outsmart the leaderboard software to such an extent Amazon has pulled this leaderboard, which was tracking their AI usage. The AI platform, unable to distinguish when it was being used by these humans for tasks that were useful and productive, or being used pointlessly, surely, that's a fairly simple thing for it to be able to work out. Yeah, look, I think it depends, look, you're talking about the worst AI model you'll use is today, because tomorrow, it will be better. So that's the thing the data makes just that much smarter. That's why ultimately you start to have what are essentially going to be rocket signs type of models over the coming years. That's how you get to physical AI. That's how you get into these use cases that are cracking code and a prize is typically what, 40, 50% of data can never be accessed from AI. Now it's now it's accessing and that's why, you know, companies that say, okay, we're just not going to play in this game's too expensive. That would be like someone back in the day being like, typewriter, this computer thing's