Nvidia is pointing to the next big AI trade

So a lot of the talk around Nvidia has been around circular financing. I think it's also interesting to talk about it in the context of what its portfolio is, right? Because it is pretty unusual for a large publicly traded company to invest in other publicly traded companies, including some of those that are its clients. Um, should people be following the Nvidia portfolio and investing in it for themselves? >> Well, absolutely, Julie. I think what's happening here is Nvidia is trying to support the ecosystem and they're investing in two ways. They're first trying to support their suppliers. So anywhere where they see any sort of bottleneck or something where they think that this could be something that uh will be really important to their uh their buildout uh they're investing there. So from that perspective those usually end up being very good opportunities. And then on the other side, they're trying to support customers or people that are buying their chips. And that's a little bit more debatable, right? So here is here companies that maybe are not able to secure financing on their own. Uh and with Nvidia support, they're going to be able to get better credit terms. So I think if you're going to be following what Nvidia is doing, I think the telling tale is a little stronger on the supplier side rather than the customer side. Um, and you know, incidentally, at least some of what Nvidia invests in is an overlap with what you invest in um, in your in your ETF. And I've actually created um, recreated the top 10 holdings in your portfolio if I had them right on our Alphasace platform here. A lot of them are these sort of optical components companies, but you also have the likes of ARM in there. You have Coherent, which we mentioned. We have Lmentum, which also uh, just reported. And also, if you take a look at the sort of year-to-ate performance for many of these, you know, even if they've had some more recent volatility, the year-to- date performance, as we're showing it here, has been pretty darn good. The the white one here is coherent, which has been really an outperformer um this year. So, how do you think about, as you say, the suppliers are where you want to be. Clearly, you're in the suppliers. Um, talk to me about that thesis, which is, you know, the sort of like picks and shovels have been popular for a while now. Where are we right now in that cycle? >> Well, Julie, what's been happening with the picks and shovels is the bottlenecks are shifting from some place to others. So, it used to be going into this year, memory was a big bottleneck. Going into next year, we think optics is going to have the same sort of dynamics is equally cyclical to memory and there is not enough supply. So, companies are going to have to rush to add supply, but they're not going to be able to add it just in time. So this is why optics is one of the the trades we really like. Capital equipment is another area. So as memory as CPU capacity needs to get added, you're going to see the semicap equipment names benefit. So companies like Lamb Research uh KA those supply the equipment. So that's another area that this cycle is just turning because companies are just starting to invest rather than something like the GPUs. there we're already few years into that uh that cycle. >> So okay so when you say they're investing in this so are we already seeing these types of companies whether it's optical or the semicap equipment are they already increasing capacity with the anticipation that they're going to see that I mean they've already seen a a bump in demand but with the anticipation that that's going to last. >> Absolutely. So for optics specifically most companies are planning to double their capacity this year and then double it again next year. But what we're seeing from demand, it could actually be seven to 10 times bigger. And this is because on top of their traditional business, this used to be like telecom suppliers with fiber optic cables. On top of that, you have the data center side, which has already been playing out, but now optical components will find their way into the servers themselves for scale up applications. And it's an entirely new market that didn't really exist before. So this is why the demand is growing 7x while the supply is struggling to keep up with even doubling uh every year and yeah this is a trend we see across the board. So semicap equipment will be the companies that provide the equipment for this doubling of capacity. >> Um obviously you're well aware of the sort of um questions or concerns about this whole cycle. There was an interesting um piece this morning by a macro strategist at Bloomberg named Simon White who said um he pointed specifically to the new Nvidia financing structure which is partnering with a bunch of Wall Street giants to potentially put 500 uh billion dollars out there in the system to try and finance um more of this compute as he said in his piece and he he was drawing parallels between this and the subprime housing crisis and AIG's financing schemes in particular and he says in the housing crisis increasing financial ization artificially boosted demand, muting the signal from the underlying market that there was already too much supply in the aggregate. House prices fell nationwide for the first time since the 1930s, an event not factored into any models. So, do you have any concerns during this cycle that because of the enormous money being poured into the system, if there is a change in demand on the from the end user that it will be masked? >> Well, Julie, absolutely. If there is decline in demand when we have all this supply coming in these industries are cyclical. So this is why tracking the demand side is so important. So we talk to companies either on the private side the public side trying to assess how much demand actually is there and there is also a way to track the spot pricing and right now what we're seeing it's actually the opposite trend. So people are not necessarily chasing like with housing the difference is that people are buying those as an investments or flips here people are not really flipping uh GPU capacity they are buying it for their internal needs. So what we're seeing with pricing, >> well, I think you could argue with most houses you're buying it for to live in it right? >> Well, but when you see the big booms, usually it's exacerbated by investment type activity, right? So it's not necessarily like people didn't buy 10 houses to live in, right? It's really going to be more driven like the bubbles are more driven by investment type uh activities, right? Same thing here. If you start seeing companies that are maybe like contracting GPU capacity and they don't have the buyer, that would be a concern. But as of right now, we see more buyers for that capacity that need these chips to to to use. >> And what about the idea that the GPUs that are already out there at data centers are not at full capacity utilization in many cases. >> Well, that's something that is going to change over time and that's basically where we see a lot of opportunities here today. So companies that provide ways for better GPU and CPU for that matter utilization are really going to benefit here. So any company that so Cloudflare is one example where like they have a a broad network of CPUs and they help companies that want to build AI applications use the CPUs efficiently. So these type of companies are really going to benefit and I think over time we will see utilization increase significantly. So you talked about optical, you talked about semicap equipment. Are there any other areas that you think are I mean it seems like these areas are in general pretty well understood, right? So are there anything are there any other bottlenecks that you see coming down the pike that perhaps are not as anticipated yet? >> Uh well I think the the optical components is relatively newer and semic equipment is relatively newer. CPUs as well I think are not as well understood. uh we ju that cycle just turned with AMD and ARM. This is why we own ARM. I think as you see the agentic AI buildout the AI agents will require a lot more CPUs per GPU versus the data centers themselves. So or model training. So I think what we're going to see is exponential increase in CPU demand. So I would say from the data center trade definitely CPUs, optics and semicap equipment are the places to be. And then broadly speaking, we are very excited about the space theme. I think that's going to open a lot more opportunities and new markets, whether it's defense, whether it's communications or data centers in space themselves. I think that's another interesting idea to put on the radar. >> Vana, good to see you. Thanks so much for coming in. And we also, by the way, just showed you data from Yahoo Finance's Alphaspace platform. That's a new professional-grade financial platform featuring advanced charts, real-time news, customizable investment research, and more. You can access all of those tools by using the QR code on your