00:00We're continuing to see the bifurcation in software, so you have the two neighborhoods,
00:04the application software names, so the bellwethers there would be the salesforce.coms and the
00:09adobes, and then you've got the infrastructure neighborhood, which by and large would include
00:14the hyperscalers, but I'll keep those separate because they're kind of a beast of their own
00:18in terms of how we think about them in the AI lifecycle and the value chain.
00:23But the infrastructure names continue to get bid up, and that's functionally because it's
00:27the most immediate and we think the most correct way to express the veritable AI winners in
00:36the software stock.
00:37I think you sort of were dancing around this, but software has sort of been deigned as this
00:42redheaded stepchild of tech, if you will, as the frontier labs absolutely express their
00:50might in terms of the type of innovation that they're putting out there and disrupting existing
00:56business models.
00:57You know, software companies, you know, arguably in retrospect, we're looking fat and happy,
01:02and now, you know, those bills are coming due and we're seeing a sort of different cadence
01:08of both innovation, operational improvement.
01:12And so infrastructure is the place to be because infrastructure software generally...
01:17What's an example of infrastructure software?
01:19It would be a company like a Datadog or a Cloudflare, right?
01:22So it's not terribly household names in the sense that we're used to, like we are in the
01:28application realm.
01:29So it's essentially companies that are enabling or providing scaffolding or visibility around
01:36the types of AI systems that you're an average organization is erupting.
01:40Does Datadog, which I know...
01:42Oh, the New York Yankees blue light for Anne-Marie Horton.
01:45The Datadog, which I know nothing about, do they provide services to, say, Microsoft?
01:52No, what they provide is effectively the equivalent of an MRI scan for your entire IT architectural
01:59topology, right?
02:00And that's important when...
02:01Why do I feel like I'm at Weill Cornell right now?
02:04It hurts here, Dr. Belani.
02:06It's just your renaissance woman with all my analogies.
02:09And so what the value they provide in the context they provide for AI is AI systems in
02:17the way they're constructed, in the novel capabilities that they're providing, you know, brick to brick.
02:22These are just very sophisticated systems and pieces of technology, right?
02:28And they're ever-changing.
02:29And we were offline talking about how the frontier is changing so dramatically that six months
02:34from now, we could be in yet another brave new world as it relates to, you know, technological
02:40proficiency with respect to models and what the labs are doing.
02:44And so what Datadog provides is, hey, you're going to build out these systems of agents that
02:49are going to run workflows for you.
02:51Well, guess what?
02:52There's a million things that could go wrong.
02:54Agents could hallucinate.
02:55The agents could spin up a wrong answer.
02:57They could outright fail.
02:59And so Datadog is that MRI scan where your innards are effectively becoming so complex
03:05that your points of failure are just becoming infinitely more or diffuse.
03:09And that's sort of the value they provide, ultimately.
03:12So Datadog infrastructure software, I get that.
03:16And the stock's up 88% year to date.
03:22Wow.
03:23What else is a Datadog comp, would you say?
03:26You know, we're very positively aligned on Cloudflare, right?
03:31Cloudflare.
03:32That's also in a similar echelon of outperformance.
03:36Yep.
03:37Call here is Cloudflare is basically the backbone of the internet, right?
03:41It's the easy pass equivalent for organizations to run their business on the internet, right?
03:49Okay.
03:49And the way they play in the AI theme is you've got these agents.
03:56They've got to run somewhere.
03:57I've got a non-in acquaintance with Matthew Prince going way back to some fun days in Davos.
04:02What is Matthew Prince doing differently at Cloudflare versus everyone else?
04:07Oh, he's tinkering with a lot of interesting ideas as it relates to what the agentic internet
04:11is going to look like.
04:12You know, one of the founding sort of principles around his view, and we tend to agree because
04:18a lot of the data points are supportive of this, is, you know, there's human-generated
04:22internet traffic, but hold the phone.
04:24There's going to be an absolute parabolic explosion in agentic internet traffic.
04:30So, you know, six months from now, 12 months from now, if I'm booking a family vacation,
04:35agentic commerce.
04:35I do not want you to answer this question, but in your head, I don't want to get you in
04:41compliance issues, are you sorting out winners and losers of the big people, the hyperscalers?
04:47Like in your head, don't, Heath Terry will put me in the timeout chair, but in the Fatima
04:54mind space, do you sort out winners and losers of the big people?
04:59Absolutely.
05:01There is a gradient, but in the infrastructure realm, I'd directionally say there's
05:05when we kind of cross over the Rubicon, the infrastructure software names are generally
05:11going to be beneficiaries of the computing shift, right?
05:15So that's why I bring up Datadog.
05:17That's why I bring up Cloudflare.
05:18We didn't really touch on the beast of a complex that is the cybersecurity universe, but as sure
05:24as death and taxes are, cybersecurity investments and budgets are going to follow what is going
05:29to end up being a much more nefarious cyber attack environment as it relates to AI technology.
05:37So we tend to also be bullish there.
Comments