Lead-Lag Live

The ETF That Owns X.AI and Anthropic — Inside the AI Arms Race with Derek Yan

• Michael A. Gayed, CFA

Share episode

KraneShares’ Derek Yan joins me on Lead-Lag Live to break down the only U.S.-listed ETF that owns equity in private AI giants like X.AI and Anthropic — in partnership with REX Shares.

This isn't just another tech fund. $AGIX is built for the AI age — giving investors rare exposure to the companies powering artificial general intelligence (AGI), including OpenAI rivals that may never hit public markets.

We dive into:
- Why KraneShares is betting on AGI as the next major tech supercycle
- What sets X.AI and Anthropic apart from other AI players
- How KraneShares built a structure to hold private equity inside a 40 Act ETF
- What the AI Score reveals about the next winners in public markets
- Why the next AI boom might happen in applications — not just chips

Subscribe for sharp, high-conviction takes with the people building the future of finance.

 Sign up to The Lead-Lag Report on Substack and get 30% off the annual subscription today by visiting http://theleadlag.report/leadlaglive.


Support the show

Speaker 1:

We're planning to really launch an ETF that captures the company that's going to take the world to artificial general intelligence. We think that's AGI. So that's why we launched AGI X. We truly value our partnership with Ethna Capital because they are early investors to a lot of those AI companies.

Speaker 2:

Hello and welcome back to another episode of LeadLeg Live. I'm your host, melanie Schaefer, head of Media Strategy and Business Development at LeadLag Media, and your host. As some of you know, I recently took over from Michael Guyad, the founder of LeadLag, and I'm excited to keep bringing you the kind of bold, insightful conversations that matter to investors right now, especially in fast-moving spaces like AI. Joining me today is Derek Yan, head of Strategy at CraneShares, where we're going to dive deep into AGI-X, the first US-listed ETF to give direct exposure to major private AI players like XAI and Anthropic. Derek, thanks so much for being here today. Thank you for having me. Gix is getting a lot of attention. It's the first ETF in the US to invest directly in private AI giants. I mean like it's a bold move, but what gave your team the confidence to build something that's this different?

Speaker 1:

Yeah, sure, I mean like looking like back two years ago, when AI is really like taking a storm, right, Like everybody's like shocked on how good like ChatGPT is really changing the way people like communicate or generating content. So we're thinking about like we're planning to really launch an ETF that capture the company that's going to take the world to artificial general intelligence. We think that's AGI. So that's why we launched AGI-X. But a challenge we face is, back then there's so many companies that is critical for AGI are still in the private stage Talking about like OpenAI, Anthropic and later I mean XAI. So those foundational model companies are really driving this round of innovation in the generative AI. So without adding those companies, you end up have kind of like quite a similar exposure compared to, say, NASDAQ 100 or just technology sector ETF, right.

Speaker 1:

So we're thinking about why not create a fund that captures the whole ecosystem, regardless of their liquidity profile? So that's the ideal we had. But like then like it's difficult, right. So then we ask the lawyer what we can do. The lawyer is like yeah, actually, ETF, end of day, is a 40-act fund.

Speaker 1:

So in the 40-act fund actually you can include up to 15% of the private assets. So that's OK. That's interesting. Assets so that's okay, that's interesting. 15% is not a lot, but it actually is decent when it comes to a portfolio of AI company. Think about there's so many AI opportunity in the public space and also in the private space, so 15% actually is a decent allocation that we can allocate to those private AI companies. So fast forward to today. We actually really complete this portfolio by adding XAI and Anthropic into the AGIX ETF. So those two companies now are in the top 10 holdings of the ETF, together with other public AI companies. So this is one interesting solution now just offering the old AI companies, regardless they're public or private, and have a solution to really achieve that goal.

Speaker 2:

Yeah, so I wanted to dig into that just a little bit deeper and follow up on that, Like talking just about sort of maybe XAI and Anthropic in particular, what are these companies that were? And the other private companies within the fund? What are they doing and why do they matter so much in the AI race?

Speaker 1:

Yeah, if you think about like how this round of AI is really changing the world we have, I mean like think about AI is really. Think about like there's several rounds of technology breakthrough right. There's like the old days, like this PC, then internet was thing, the mobile internet right. Then I think like cloud, become like all those apps and it's now going to cloud and starting from like two or three years ago, we think generative AI is really changing the technology sector, changing the innovation. That's going to create a similar opportunity like the previous generations of technology breakthrough or innovation breakthrough, and this round of really development is really driven by the AI large language model companies. So those large language model companies they train like tons of data and they can generate content, they can have intelligence, that power, the ability for those models to do a lot of work right, like content generation. There's calculation, code generation, so anything like related to code or in the digital world can be automated. That's really come to client services, come to a lot of like I think, white collar jobs in the enterprise world. So that is a game changer for a lot of enterprises and consumers to really automate the whole digital world. So that's a big opportunity and the company is critical. I would think about it in the consumer side, it's OpenAI, which is probably the most famous one. That's what everybody's now using, but actually in the enterprise world, anthropic is a big player that many enterprises is actually using Anthropic's API. Anthropic also has a consumer-facing app called Cloud AI. It's famous for its cloud model, mostly on the, say, legal side or coding side, not engineering side. So that is something I think, like many insiders in the tech world or engineering world or law firm, is usually like using cloud. But I think, most importantly, the coding is a big thing, vertical for the AI adoption and Anthropics really leading in that section by partnering with a lot of application companies. So I mean there's a lot of like autonomous coding right now using the Anthropix API.

Speaker 1:

Xai is another large-slang model company founded by Elon Musk right With its unique access to the data from the social media platform X. So XAI is actually like a very powerful model with, I think, access right Because Elon Musk has bought a lot of the GPU and built a large data center. That's powering the model behind XAI and XAI becoming the most, currently the most intelligent large-dash model so far. That gives them the edge both in the consumer side and the enterprise side. So if you think about like this competition right of AGI, like everybody's now trying to achieve artificial general intelligence in 10 years, 20 years, it's still at an early stage.

Speaker 1:

We don't know if OpenAI, anthropical, xai, which one's going to be probably the AGI companies in the future, we were thinking about just adding more of those, like in the future. So we were thinking about like just adding more of those names in the portfolio. It's really like you don't want to miss any of them. What if, like, anthropi become the solution where XAI become the solution? So adding those companies, those foundational model companies, in the portfolio is pretty critical. If you think about Asia is coming in a decade or two decades.

Speaker 2:

And I guess to follow that up as well. I wanted to talk a little bit more about the future of your allocation strategy. I mean, do you expect that the mix between private and public companies is going to change as the more public companies mature, and I mean especially with so many of them trading at multiples?

Speaker 1:

Yeah, exactly, I mean, when we added those private companies, so we added their primary round, right, so Anthropic, we entered their financing round of Series E. So that's like they're going to do multiple financing going forward. They may IPO in the future. So once they IPO, they become like public companies. So there's, I think, eventually a lot of those private AI companies going to be public. So then we're going to see like how this portfolio going to evolve right in the future. Because, yeah, then I think like currently a lot of the. Then we're going to see like how this portfolio is going to evolve right in the future because, yeah, then I think like currently a lot of the those like high valuation companies are really the foundational model companies or large-net-share model company, like OpenAI, like Anthropic, like XAI, with like a hundred billions or 300 billions valuation. So it's already like a late stage for them to really become IPO in a stage.

Speaker 1:

Then I think at that time when those companies IPO, our focus in the private side is probably going to more on the application side. So we actually, on our phone, we want to capture the three layers of the AI opportunities. One is the hardware right, like those chip company or data center related company to building the hardware for the large selection models. And then there's infrastructure layer. You have cloud, you have data security, data preparation, model preparation and then the model company themselves in the private side. So they are private side now.

Speaker 1:

Then when they become public, I think a lot of opportunities now in the third layer, the application side. We currently we do have a lot of application side company that is public. But I think like when the computing power of the large-next model, when the capability of large-next model becomes so powerful, there will be like tons of new application companies coming up. There will be like tons of new application companies coming up. So we see, currently there's several like, I think, popular one including, say, proplacity, or many in the content creation or like digital media. So there's so many applications right now. Even the physical AI side, right when people are now using the data in the physical world to do automation in the factory, automation of robotics, those are really like future opportunities that we can consider in the private side to maintain a decent exposure into the private application to the CTF.

Speaker 2:

What about the AI score? Can you talk a little bit about that? How does it differ from a traditional tech index sort of approach?

Speaker 1:

Yeah, so I forgot to mention. So for AGIX, we have a partner called Aetna Capital Management. So they're a venture capital firm that they're an investor of the company, like XAI Anthropic, that they are early investor of the company like XAI Anthropic. So with their early connection with the company, we are able to access those deal. So that's the kind of like how we source the deal. So also with them, we give back like AI score together.

Speaker 1:

Because when we look at this, the traditional index methodology to really capture the AI companies is either based on revenue or is based on certain, say, keyword searching or document searching. We don't think that's a very good way to really capture the AI companies. We don't think that's a very good way to really play this capture the AI companies. Because we really need to look at way that, how large-slash model, how Gen AI is changing the business model across all those technology companies.

Speaker 1:

Think about it right now, every company claim to be an AI company, even like AT&T or like all the other companies now is using AI. So how do you really evaluate the impact of AI? So that's why we partner with Aetna Capital Group. They have a committee and together with their committee, with our index provider. We developed like an index-based methodology to incorporating those AI score to evaluate the readiness and relevance of AI for each company. So we look at current revenue exposure for that company that is exposed to Gen AI and also we look at their business model, how this business model is going to be impacted positively by the Gen AI. So with those two angles we captured the I think, the right company that's really going to be AI beneficiaries.

Speaker 2:

Now, Derek, I want to talk a little bit about the performance of the fund. I mean, as you said, a couple of years old and AGX has already outpaced some major tech benchmarks. What do you think is driving the outperformance?

Speaker 1:

Yeah. So I mean, when we look at the performance of AGX since inception, it has largely outperformed the broader technology or growth indexes, for example NASDAQ 100 or just technology sector index or semiconductor index, technology sector index or semiconductor index. I think the AI score is really creating a comprehensive way to evaluate the long-term AI beneficiaries right. So that framework is actually adding a lot of outperformance since our inception Because when we look at like there's several rotation happening right, like I think at the early stage there's a lot of value just really creating in the AI hardware space. Like NVIDIA is definitely the biggest beneficiary in that there's a lot of data centers being built out, there's a lot of energy cooling and water. Those companies, as you said, it's going to be a benefit. We do have those companies as the AI scores were high at this stage when we launched ETF and gradually we realized the value is migrating to the infrastructure and application companies so along the time the score for the infrastructure and application company become much higher. For example, we are the early investor of company, for example, palantir and Duolingo. That's a perfect example of how the AI is used right in both the enterprise world and the consumer side. For Palantir, they are a leader in the, I would say like creating an operating system for enterprises to use AI, and we realized their AI score is very high and we're including them just before even NASDAQ index. Including the Palantir and Duolingo is another interesting example. Most people know the firm for its language learning, but when we look at how their business model is going to benefit from AI, they just have very good potential and we have validated their technology to incorporate AI in the language learning and the result is very impressive. So with that framework we realized there's application companies is actually adding a lot of performance to the AGI-X ETF as well. Yeah, so that framework is really creating a thing like long-term framework to establish a portfolio that captured the value, gradually migrating from hardware to infrastructure and application, and I would think private companies are going to add in some performance going forward. Because those companies are not public, so they don't have a daily price, so it's not traded on any exchange right now. So how do we really? I think like it's interesting.

Speaker 1:

A lot of people ask us so if you have, like Axia and Anthropic, how are you going to value those? So, because ETF is is a daily offer, daily liquidity for any investors, anyone can buy AJAX ETF with their brokerage account. So that means we have to offer data liquidity and then we actually value those private companies at a daily basis based on all the information, including their primary round valuation, including their secondary market transactions. So we just included the AXA recently and we included Anthropic earlier this year. I think both companies are really growing very strong in their revenue stage. So I'm sure, like with the NAS financing round coming or just like going forward if the IPO, they're going to add a lot of value to agx investors yeah, with the structure of the fund, though, like I mean, the fund actually owns equity and xai and anthropic correct yes, so, uh, actually we own where agx uh uh sit on the cap table of both xai and anthropic uh, so it's a shareholder of the two companies.

Speaker 1:

So that's like direct ownership compared to owning them via another structure. So it's very simple. We like that simplicity. I think that's like most transparent way to invest in those two companies and our investor really loved that way, compared to there's many other, I think, retail trading platform or like a very complicated structure to invest in those companies compared to AGX, which is like simply directly owning those shares.

Speaker 2:

Yeah, it's quite rare and I know fair valuation is something investors worry about quite a bit. I know fair valuation is something investors worry about quite a bit, you know, especially with private equity and with the liquidity in ETFs. Can you explain how CraneShares handles pricing for the private side of the portfolio and what systems are in place to make sure that it's accurate and has the investor money first?

Speaker 1:

Yeah. So first of all, those companies are not really a startup. So for startup there's like a wide range of what is the fair value right for those companies. But the company we invested are mostly the late stage large cap technology companies, right. So if you think about the valuation of those foundational model companies, they're hundreds of billions and now the companies are staying in the private longer. So there's a decent number of investors actually now trading in the secondary market for those companies. So we actually use a lot of third-party data provider to really gather those transaction data happen in the secondary market and there's a lot of information we can use to really have a best practice to have a daily fair value committee and daily fair value process to incorporate all the available information that deliver the daily value for those private companies. So that's why AGX, as ETF can provide daily liquidity to all investors that want to invest in those companies.

Speaker 2:

I'm not sure I think you touched on it, but the fund supports financing rounds for these private companies as well, which is a new dynamic for the ETF world. How does that relationship work in practice and what kind of access does that give you compared to traditional ETF models?

Speaker 1:

Yeah, it's really actually because those companies are. I mean, everybody's like chasing those companies right. So it's really actually because those companies are I'm like everybody's like chasing those companies right. So it's really hard. That's why we truly value our partnership with Aetna Capital. They have because they are early investor to a lot of those AI companies.

Speaker 1:

By the way, I forgot to mention Atmar Capital was founded by AI native engineers and early investors, so themselves have a decent network among the AI researchers and engineers across those Gen AI companies and AI application companies.

Speaker 1:

So that's why I have a close relationship with the company. It's very important because usually now it's over-subscribed for any round of financing, given how popular those companies are at this stage. So getting access through this connection and with the help of theetna team is very important. And we actually, when we launched AGX, we invited both, I think, anthropic and Proclacity team to join our fair reigning on the NASDAQ exchange and we did a panel discussion with them to talk about the future of Gen AI. So we're going to keep doing that, with the close relationship with the company and with their engineering team, research team, to really build this network and ecosystem for the AGI. I think that's kind of like the approach we're going to do. We're going to maintain a good relationship with the company themselves. So in the future, if there's another round of financing or tender offer, we're able to really participate.

Speaker 2:

And so are there more private names coming in the future. Can you talk a bit about that at all?

Speaker 1:

Yeah, I mean like there's a lot of companies we're engaging, that's in the pipeline. There's a lot of like exciting opportunity, I think, both in the financial model round sorry, foundational model round and also a lot of application companies we're talking right now. But because this is still, I think, like when you have this conversation, when you haven't really made the deal, you cannot announce that, unfortunately. But I would say, like investors, please follow Cranesharescom, slash AGIX for more announcement going forward. So just stay in tune.

Speaker 2:

Yeah, derek, and you have other funds as well. Can you talk a little bit about those just more in general, and do you have any new products on the horizon?

Speaker 1:

Yeah, actually we recently launched a humanoid-focused ETF called K-O-I-D Koid. We think, like when we launched AGIX, we think like Gen AI is really exciting and it's phenomenal, so what could be like kind of like nest Gen AI? So when we also, like we were watching, like the CEO of Nvidia, jensen Huang, his keynote at the GTC conference I think he mentioned that the robotics, especially humanoid robotics, can be, could be the next big thing the traditional AI to generative AI to AI agent, where a lot of AI apps can really autonomously get a lot of things down, in the future it could be happening in the physical world. So physical AI is kind of the next big opportunity. So that's why we launched KO-O-I-D recently to really capture the humanoid opportunity. Actually we bring a real humanoid.

Speaker 1:

There's a company called RoboStore based in Long Island, new York. They're a distributor of the humanoid robotics. I was there at their warehouse in Long Island. It's like a dream come true. Like there's robotics, like the robots are everywhere that can move and run like very smooth. So seeing that, I mean since believing the technology breakthrough in the hardware is really at a stage, I think like making humanoid and that's big potential for a lot of applications. And AI is bringing the intelligence to those humanoid robotics, so that's why we created a KOID to really capture that opportunity.

Speaker 2:

And there's a great picture of the humanoid robot ringing the bell at the stock exchange on your website that people can go and look at. Derek, it's been so great having you here.

Speaker 1:

Thank you very much. Thank you for having me.

Speaker 2:

I'm Melanie Schaefer and we'll see you next time on Lead Leg. Live Bye.

People on this episode

Head Shot

Michael A. Gayed, CFA

Host