Want Anthropic exposure before its IPO? Here’s how

Video Transcript

AI investors know the chip story, but what if the next big opportunity isn't making chips, it's moving data between them. Yuri Kojarian is Chief Investment officer at Tema ETFs and joins us now. Yuri, welcome back.

Thanks very much for having me.

So we last talked about space. You're back with an ETF targeting the next piece of the AI buildout, the Tema Photonics and optical ETF, LAZR. Uh, to kick things off, what exactly is photonics and why should investors care?

Um, I think the best way to think about it is to understand that AI is creating a huge amount of computing power within these data centers. And that computing power is just getting bigger and bigger and bigger. And at some point, the other parts of the computing system are going to require, you know, an upgrade in a way. And one of these is how data is moved between parts of the data center, between the individual servers and within even the individual chips. And the reality is that data is moving at such speed and it's also moving with so much bandwidth that the copper connections that we all know, the copper wires that connect the world are just not keeping up. You know, they're hitting their physical limit. And the obvious answer, which has happened in lots of other industries is you move to light, uh which is photonics. The idea that you can use the full spectrum of light to move data, means that you can move data faster and you can move a lot more data, which is what a lot of these AI systems need.

So replacing copper with light, but what am I actually buying with LAZR? Kind of walk me through the portfolio. How do I get exposure to light in my portfolio?

Absolutely. Um, one of the things that we pride ourselves on at Tema is building these curated universes of companies. Um, and we sort of leave no stone unturned. We find the most interesting businesses exposed to this particular theme. In the case of optical and photonics, you're looking essentially at the entire supply chain and value chain of creating the devices that then the basically lead the data sense to be able to connect all of those different parts together. So you're thinking about the machines that make the substrates, that make the materials, that then create those particular lasers or optical components that allow light to carry data inside a data center. So you've got um winners and leaders in terms of companies like Lumentum, you know, which is a transceiver maker and uh one of the world's leading laser producers as well, all the way back to the beginning of the value chain, which is uh companies likeron which make the machines that make the materials to make those lasers. So you can imagine that what we're trying to do is give investors exposure to the entire value chain. and we're particularly focused on Western companies, we're particularly focused on the ones that have the winning capacity and winning technology to deliver optical and photonic solutions to the data center.

So Lumentum is your largest holding, looks like nearly 18% of the fund, but your second largest holding is Anthropic. How does a photonics ETF end up with more than 9% exposure to a private AI company?

So, one of the things that we really focus on at Tema is we're trying to give investors a good package together. So the way we see it is that the way we define the universe and the way we think about it is it's the entire AI infrastructure stack. And if you look at a company like Anthropic, it's obviously the LLM layer, it is the demand layer for a lot of these types of equipment. We had the opportunity to add this as a position. It's something that our portfolio managers look at and consider as part of their entire investment universe. and in the case of LAZR, we thought that there would be a good fit within the portfolio. We also were able to invest in it at what we think is an attractive valuation for what is probably going to be a generational company. Private companies can be in ETFs and we think that that is something that investors want and something that we are prepared to offer after having done all the necessary due diligence. Very similar to our NASA ETF and its pre-IPO exposure to SpaceX.

Mhm. Why Anthropic over Open AI?

That's a great question. Um I think a couple of uh maybe more subtle points. I think one is that Anthropic is a major kind of investor and user of some of the optical technology. So it's a slightly better fit in terms of the infrastructure build out that they're doing. A lot of Anthropic's future compute is going to be based on the TPU ecosystem, which as you know is developed with Broadcom, which is a leader in in photonic and optical technologies and TPUs tend to be a much more or to a certain extent, somewhat much more heavy in terms of its usage of optical technologies than maybe the Nvidia ecosystem. So I think that's one maybe more technical reason why we've gone for it as a fit. But really we think that the most important application of AI is going to be within businesses and Anthropic is a leader in the B2B space for um in terms of its overall revenue exposure. So broadly we think it's a more interesting business, it's also much harder to access uh in terms of buying those shares out into the secondary market.

So if I'm buying LAZR in part because I want access to anthropic before an IPO, what do I need to understand about that exposure and then what happens when anthropic does IPO? What happens to that position within the ETF?

Absolutely. Look, first of all, you have to understand that we're making an investment decision here. The second thing is that we have done the necessary diligence to make sure that this is an exposure um investors can trust. Uh you know, there are multiple layers of due diligence that go into making an investment decision like that that are maybe different to buying a public company. You own or the the the ETF owns on Tropic through an SPV, which is a very standard way of making investments in the private world. Um these SPVs have no fees attached to them. So essentially investors are just paying the management fee of the ETF. So it's a very cost effective way to gain that kind of exposure as well. Finally, we we think it's um once it comes into the ETF, that that's the exposure that you're getting. So you're getting a collection of the leading infrastructure photonic businesses plus the anthropic state.

Okay. You know, I was just taking a look at some of your other top holdings, not the quintessential AI names that probably appear in uh many investors portfolios. I, you know, do see some familiar names, Tower semiconductors in there. and we had mentioned Lumentum. But how should I be thinking about photonics exposure in my portfolio versus other sort of AI exposure? Is this complimentary? Does this replace something? What if I own names like Nvidia and Broadcom right now? What does this give me that those don't already give me?

Absolutely. Look, I think the way to think about the AI infrastructure build out is to kind of start top down. Uh I think uh it's no secret that there's going to be a large build out, um that it is probably going to be in the trillions of dollars, probably higher than maybe the rest of the market kind of thinks at this moment. And what you need to think about that AI infrastructure build out is to say where are the particular bottlenecks that are being hit as we accelerate that AI demand? If you look at something like GPUs and TPUs, which is what Broadcom and NVI provide, there's a pretty well supplied market. So there is no supply demand mismatch really. Um the only thing that's pacing that demand picture is how quickly can we build the data centers and how quickly can TSMC produce them? But once you start getting into the rest of the AI infrastructure value chain, you start to hit bottlenecks. And there are several that have arisen over this past period and the one that we've focused on with this ETF is the optical bottleneck. Essentially what's happening is the demand is starting to inflect massively upwards because as these data centers are becoming more powerful, the data needs to move a lot faster. So we're starting to replace components with optical fibers. And remember, this is a kind of a networking concept. So the more powerful each particular connection within the network is, the more you need to upgrade the whole network and there are more connections than there are individual parts of that network.

If you look at uh people like Goldman Sachs, they estimate that this market is going to grow something like 10 times over the next couple of wars. So there's an enormous demand impulse. But the key part, which is why this is an interesting investment opportunity, is the supply. There just is a massive constraint of supply delivering the lasers, the transceivers, the optical equipment that goes into these data centers. And this creates a positive investment environment because you have very strong demand, forward orders and companies are built better able to build out their business to take advantage of this opportunity which could be multi year. So that's why we think that it's an interesting area to look at. And so for investors that are looking over the broad investment portfolio in AI, they should be making deliberate decisions. They should be thinking this AI infrastructure build out could be one of the largest things humanity has ever done. Should I be focused on just giving the general exposure through things like in video or should I be thinking about what are the individual bonics? So these are very nice compliments to for example, holding in a video or Broadcom where you're saying actually, I think this part of the build out is going to have the tightest supply and demand mismatch for this period of time as the build out goes on.

One of the things about the optical industry that people maybe don't appreciate is that historically it was an area that was forgotten by the market and that's because most optical technology went into telecoms and telecoms has just not been a growth industry. And so for a lot of investors, you've got this kind of weak starting point because it was an industry that was overlooked and that creates some of the most powerful investment opportunities in the marketplace.

I was going to ask how big the opportunity is, but you already said it's a multi-year opportunity. But I guess what has to happen for this market to grow as much as you expect?

That's a great question. Um and one of the things that there's a bit of a debate is at what point do we start to introduce optical components further and closer to the actual accelerator chips, the Nvidia chips. You can think about a future where even at the chip level, you are using optical components to move data because it's become such a high data burden and we need the data to move so quickly to deliver some of the AI agents that you're seeing out there like Muse or Instinct. So really it's about understanding what the Nvidia and Broadcom roadmaps look like. And that's perhaps one of the debates out there in terms of what will accelerate demand faster. Now, if we move to having co-packaged op as it's called where optical um components are coming much closer to the chip, then the opportunity is much nearer term and much bigger than it is. The way we think about it is that there might be some delay to this, it's going to take some time, but that creates a longer environment where some existing optical components see a bigger pick up in demand and then we have generational technology changes. So that is probably the thing that's pacing how big this opportunity is and when does it happen now or in a couple of years time?

But AI infrastructure has already had this massive run. So investors are investors early to photonics and obstacle at this point or some of these stocks already pricing in the opportunity.

Uh look, I think it's fair to say that the AI infrastructure trade in certain areas is probably well known to people. I think the parts that are less clear are probably how big is the final outcome going to be. Um I think if you had spoken to someone two years ago or a year ago, they would have said, you know, maybe it's a two three trillion investment opportunity. I think there are some estimates now that are out there that are five and and if you look at semi analysis who's a partner research partner for us in this ETF, you know, they're estimating $10 trillion build out, which is bigger than any of the numbers out there in the market. I think it's also going to take longer than people realize and there are certain pacing things that are happening. You know, TSMC can only produce chips at a certain pace. We have a cute need to power these data centers which is going to take time to build out as well. And so those things are pacing the opportunity if you will. Um and so investors can kind of make a bet here that things are going to take longer and probably going to be bigger than the rest of the market thinks. So even if some of these stocks have already started to react, which they have, the kind of AI bet right now is one on length and magnitude rather than whether it's happening or not.

In terms of valuations, well, it's very easy to build valuation cases if you think it's bigger and longer than the market thinks for most of the AI infrastructure stocks out there. So, if you're prepared to take a longer term view, there is an opportunity there. Of course, short term and especially this year, the AI infrastructure trade has been very strong. So a lot of these stocks have gone up. So short term, it's possible that you're going to have and you see already over the summer, this periods of volatility as the market processes, the run up in share prices against the expected opportunity if you

I want you to take your ETF cap off for just a second and put on your CIO cap and talk to me just about the opportunities that you're seeing in the current market because we have memory stocks selling off so far this month. Obviously still big winner as year to date, but down double digits for a lot of the big names. We have software stocks that have been making a a a comeback or at least some software stocks making a comeback. Then of course, there's the Mag seven. As you think about where the best opportunities are right now based on current valuations, where are they, Yeri?

It's a great question and often when I get asked these things, it sort of it's difficult to say on a short-term horizon where any particular direction the market's going to go in. But I think it's helpful to think about things longer term. So for me, personally and the way we look at the market, we like to look for these opportunities where you've had a period, ideally a very long period of underinvestment. So whether you're looking at the US power situation or you're looking at something like photonics and optical stocks, those were stocks that saw very little growth and there was essentially under investment in the supply chains because there was just no demand growth. They're suddenly now thrust into this growth environment which creates a really interesting valuation opportunity long term. So I see a lot of opportunities there if you're prepared to think longer term and appreciate the demand and supply imbalance that's being built on the market right now.

I think other areas as a CIO that I see is investors need to start to think how is AI going to move uh even if you think the AI infrastructure trade probably has a some way to go. What is the actual impact of AI on different parts of the market? And I think that part is what we're digesting this year. You know, what you've seen in software stocks where they sold off and now they've recovered. I think the interesting trade right now is to understand which vertical is AI going to have the greatest impact in and then starting to think about that verticalization opportunity. So one area that we really like is healthcare AI and we've recently launched an ETF in that area, HTH. It's actually the first and only ETF focused specifically on healthcare AI. So you can see our thinking moving, right? We're in we are big believers in the AI bottlenecks and the build out of AI infrastructure and the second leg is where are the underappreciated opportunities on the application of AI? And we think healthcare is one of those. You've already seen some of these stocks start to react to the market environment as companies like Anthropic and others are investing and building AI for science and for biology. It's a massive opportunity out there. You know, healthcare is 20% of GDP in the United States and growing. I think there is an opportunity here to really build long term value for some of these businesses.

Interesting. And there's an ETF for that of course. All right, I think it's a great time to pivot to our rapid fire round of this or that quick questions, quick answers, no hedging if you can help it. Are you ready, Yuri?

Go for it.

More upside from here, AI chips or AI networking?

AI networking.

Photonics, early innings or getting crowded?

Early innings.

Biggest AI bottleneck, power or data movement?

That's a difficult one. I'm going to say power.

Better AI trade from here, Hyperscalers or their suppliers?

Their suppliers.

Bigger AI risk, valuations or slowing capex?

Oh, absolutely slowing capex.

Hyperscalar capex in 2027, accelerating or peaking?

Oh, I'm accelerating.

AI stocks next year, broader rally or fewer winners?

Probably broader rally.

Anthropic IPO, buy at the open or wait?

Oh, definitely buy at the open.

Or own when it's private. AI IPO boom in 2027, bullish signal or warning sign?

I think probably slight warning sign.

Anthropic or Open AI?

Anthropic.

AI trade from here, chase momentum or wait for a pullback?

Well, I we're sort of in a pullback so chase momentum.

If rates stay higher for longer, AI infrastructure or AI software?

AI infrastructure.

AI stocks or space stocks for the next five years?

Wow, you're making me choose between. I'm going to go for space stocks. Space stocks.

Two babies.

If you could only own one AI stock and it's not an ETF, which is it?

I would go for Lumenta.

Bring it back, bringing it back to LAZR. Finish this sentence. The biggest risk for investors buying LAZR or betting on photonics is

I think it probably is a capex potential capex slowdown in 2028.

All right, we'll leave it there. Yuri Komerian, Chief investment officer, Tama ETFs. Thanks so much for playing along, shedding some light on this ETF and uh we really appreciate your insights as well.

Thanks very much for having me.

If you enjoyed this ETF spotlight, check out our full interview with Dave Ellison. He breaks down what bank earnings tell us about the economy right now and gives his top bank after earnings.