00:00 Speaker A
that that uncertainty you’re talking about there, Andrew, in terms of this data center financing, at least for some names, you would say, okay, that that’s that’s uncertainty, that’s not, that’s not real fundamental risk right now.
00:15 Speaker B
No, I I don’t think it’s a fundamental risk. I I just think when you you take the coreweaves and you know, neo clouds and take Oracle sort of a latecomer and that doesn’t really have as much room for to issue to issue debt. Neither does core weve necessarily and they don’t have the free cash flow really to keep up with the others. I would, you know, lump in of course Open AI and and Anthropic as not having the free cash flow to like sort of meet their data center ambitions. Um, that’s those are all legitimate concerns, but I don’t think it’s going to slow anything down from the legacy hyperscalers who do have ample free cash flow, and they’ve got tons of room on the balance sheet to issue debt if they choose to go that route. So scaling laws still apply. We expect the the trend to AI CapX trend, which is looking like 50%, you know, kind of growth year over year. Um, I don’t know if you come up in you come up with a higher number if at any point they those numbers start to move higher. Um, I think this, you know, little uncertainty pot hole that we’re in, um evaporates and and then you do get the tiny bit of multiple acceleration, but it’s really about earnings growth and and earnings growth in networking equipment. I think is is set to inflect higher.
01:53 Speaker A
And before I let you go, Andrew, uh let’s just get to some picks. Are you laying out just there because I was going to ask you, you know, why is Arista a buy?
02:04 Speaker B
Okay, so, um, the scale up, right, which is what’s going on now before like scale out, building more data centers, putting, you know, in servers and server racks and so forth. Scale up is about adding compute. It’s um basically adding, you know, to the XPU clusters, so to speak, and inside each um rack and then attaching, connecting the racks to each other. That leads to a higher attach rate. It drives exponential growth in in switch equipment and uh that’s what Arista does. Arista makes ethernet-based switching equipment and when people are looking for an alternative to GPUs, right? They they want to diversify away from Nvidia because they don’t want to be captive obviously and the higher prices and so forth. The TPU architecture for for example. Um, you’re going to see more of that to drive inferencing costs lower, but that just drives um exponential growth in networking uh switch equipment. And as you go from Nvidia-based, you know, Infiniband to ethernet, I think that all lands at Arista and Arista also has two and a half billion dollars of deferred revenue which will be coming off over the next 18 months, two years.
03:16 Speaker A
Yeah, one more I want to squeeze in here Andrew. Coherent. Now that stock has soared about 100% in 2025. Is it still a buy here, Andrew?
03:35 Speaker B
Absolutely. Yeah, I mean, the coherent, Lumentum, Fabrinet, these are optical transceiver companies. They’re important components, you know, as you you slowly have to move away from, you know, copper-based um equipment. And I think that the, you know, move from 800 gigs to 1.6T to 3.2, it’s all happening much faster than anybody had expected. And there’s an opportunity not after, you know, after we get through this scale up architecture, there’s scale across, which is after that. So we’re going to be connecting data centers with each other and that’s all going to require uh optical equipment and I expect coherent to be the one that’s closest to being able to meet the surge in demand. Um so that would be our top pick.