Latest / Crypto Coin Show / Kevin Carter says China is now favored to win the AI race outright — and he built two ETFs around that bet
Transcript
- 0:00Speaker 1: I'm Ashton Addison from the Crypto Coin Show. And today
- 0:02Speaker 1: we're joined by Kevin Carter, founder and CIO of EMQQ
- 0:06Speaker 1: Global and EMX ETF. Kevin has been building in the
- 0:10Speaker 1: emerging markets index funds since 2014. And just in the
- 0:14Speaker 1: last month, his team has launched many new funds, the
- 0:17Speaker 1: China AI ETF. And as of this week, the China
- 0:21Speaker 1: AI ETF is under AICH, a broader bet under the
- 0:25Speaker 1: whole AI stack. There's a lot to discuss between Western AI,
- 0:32Speaker 1: Chinese AI, and where we're all at.
- 0:34Speaker 2: It's moving so fast.
- 0:35Speaker 1: And you're the man that has some of the insights
- 0:39Speaker 1: that I would love to dive into today. So thank
- 0:41Speaker 1: you so much for taking the time, Kevin.
- 0:43Speaker 3: Sure. Glad to be here.
- 0:44Speaker 1: Yeah. So most people in AI or investing in the West,
- 0:49Speaker 1: they're looking at AI and they're trying to find where's
- 0:52Speaker 1: the perfect place to be investing. Are we buying the
- 0:56Speaker 1: picks and shovels? Are we investing in the models? Are
- 0:59Speaker 1: we looking at Nvidia for chips, data centers, power companies,
- 1:05Speaker 1: it's all exploding. And some are higher upside, higher risk
- 1:10Speaker 1: than others. And I think as we are in this
- 1:14Speaker 1: exponentially fast growth piece, there's a bit of noise. And
- 1:20Speaker 1: I think that's going to fade out over the next
- 1:23Speaker 1: couple of years and we'll see some clear winners in
- 1:25Speaker 1: where the most asymmetrical upside was. What's your take on
- 1:30Speaker 1: where the money is right now in AI?
- 1:34Speaker 3: Well, right now the money is going into the infrastructure.
- 1:38Speaker 4: And if you, you know, the best way to think
- 1:41Speaker 4: about AI is the five layer stack. And as you know,
- 1:45Speaker 4: the bottom layer is power, the second layer is chips,
- 1:48Speaker 4: the third layer is the infrastructure data centers. And those
- 1:52Speaker 4: first three layers, that's really the platform on which AI runs.
- 1:56Speaker 4: I mean, we had Power before AI, we had chips
- 1:59Speaker 4: before AI, we had data centers before AI. But it's
- 2:02Speaker 4: that fourth layer, the magic layer, as Jensen calls it,
- 2:04Speaker 4: the model layer that requires all of that capital expenditure.
- 2:08Speaker 3: And so it's that picks and shovels layer.
- 2:12Speaker 4: And right now, and not just now, but over the
- 2:15Speaker 4: last couple of years, there's been anxiety about the scale
- 2:19Speaker 4: of the spending. And the hyperscalers have basically deployed all
- 2:23Speaker 4: of their free cash flow towards building data centers. And
- 2:27Speaker 4: there's rightfully a lot of concern about whether that's going
- 2:30Speaker 4: to pay off. But on the flip side of that,
- 2:32Speaker 4: you have companies in Asia and in Taiwan and Korea
- 2:36Speaker 4: in particular, the MAG-3 of AI, TSMC, Hynix, and Samsung.
- 2:42Speaker 3: And they're making a lot of money.
- 2:44Speaker 4: I mean, if you look at the earnings of those
- 2:46Speaker 4: three companies between this year and next year, it's expected
- 2:49Speaker 4: to be about $ 965 billion of profit. So the picks
- 2:55Speaker 4: and shovels are at least in a profitability lens, that
- 3:00Speaker 4: they're the ones making all the money off of that spend.
- 3:04Speaker 4: So that's a clear place the money is being made.
- 3:07Speaker 4: And the models, it's not clear any of the models
- 3:10Speaker 4: anywhere are making money yet, but that's one of the
- 3:13Speaker 4: trillion dollar questions.
- 3:14Speaker 2: Definitely.
- 3:15Speaker 1: And that's an interesting point because it seems like from
- 3:19Speaker 1: an end user perspective right now, people are paying $ 100, $ 200,
- 3:22Speaker 1: and they're getting PhD level inference and coding for a month.
- 3:30Speaker 1: And I feel like this should be worth, you know,
- 3:31Speaker 1: I should be paying tens of thousands for this. How
- 3:34Speaker 1: is the company not, you know, going bankrupt? So how
- 3:40Speaker 1: do you actually see that playing out in the cost
- 3:43Speaker 1: of tokens and the AI companies doing the modeling? You know,
- 3:47Speaker 1: how can they stay profitable in the long run and
- 3:49Speaker 1: how will that work out?
- 3:51Speaker 3: Sure.
- 3:51Speaker 4: Well, as you know, one of the funds that we,
- 3:53Speaker 4: in fact, the first fund that we launched as part
- 3:56Speaker 4: of EMX ETF was the China AI Tigers, TGRZ, which
- 4:01Speaker 4: tracks the Chinese model. So again, the large language models,
- 4:06Speaker 4: the models, the labs, whatever you want to call them,
- 4:09Speaker 4: the US and China are the.
- 4:11Speaker 3: Only real players. I mean, they are the players.
- 4:13Speaker 4: And in the United States, we, of course, have OpenAI
- 4:17Speaker 4: with ChatGPT. We have Anthropic with Claude. Google with Gemini.
- 4:21Speaker 4: We have other players. Musk has a strong offering as well.
- 4:26Speaker 4: But most or indeed nearly all of our models are
- 4:29Speaker 4: closed models. Now, on the other side, you have the
- 4:32Speaker 4: China models and you have DeepSeat, the one that most
- 4:35Speaker 4: people know because it's the one that sort of rattled
- 4:37Speaker 4: the cage on the US AI story when it launched
- 4:42Speaker 4: 20 months ago in January of 2025. There was already
- 4:45Speaker 4: a lot of anxiety about all the CapEx. And DeepSeq said, well,
- 4:49Speaker 4: we made the same thing for $ 6.
- 4:51Speaker 3: 5 million.
- 4:52Speaker 4: And that really freaked people out. And I think NVIDIA
- 4:56Speaker 4: dropped 35%. And it started to question whether we needed
- 5:01Speaker 4: to deploy that much money. But beyond DeepSeq, which has
- 5:04Speaker 4: continued to push the envelope, you have the so-called AI
- 5:09Speaker 4: tigers of China. These are six China AI companies, all
- 5:14Speaker 4: of which have come out of the Tsinghua University Computer
- 5:17Speaker 4: Science Lab.
- 5:18Speaker 3: This is a leading... science lab in the world, likely.
- 5:22Speaker 4: And those include Z.AI, which has gone public and is
- 5:26Speaker 4: part of the Tigers. That's Minimax, which went public in
- 5:30Speaker 4: January as well. So those are the only two publicly
- 5:32Speaker 4: traded large language models today.
- 5:36Speaker 3: And then you have.
- 5:37Speaker 4: Kimmy and the Moonshot, or Moonshot and the Kimmy K3
- 5:42Speaker 4: model as well, which a lot of people know about.
- 5:45Speaker 3: That was sort of the second deep-seek moment.
- 5:46Speaker 4: And the China model's have a lot of advantages. They
- 5:49Speaker 4: are open weight and can be downloaded and run on-premises,
- 5:54Speaker 4: so you don't have to log in and pay those
- 5:56Speaker 4: expensive inference fees. You can customize them for your own use,
- 6:01Speaker 4: and they're increasingly just as good and getting better, catching
- 6:06Speaker 4: up with our models, and they cost a tenth or
- 6:11Speaker 4: even less the price. And because of that, you're seeing
- 6:15Speaker 4: them get used and not just used in China, they're
- 6:17Speaker 4: getting used in the United States. Airbnb is built on
- 6:22Speaker 4: Alibaba's model. Cursor, which was acquired by SpaceX, was built
- 6:26Speaker 4: using China models. So the China models have actually passed
- 6:29Speaker 4: the US models in terms of token use in.
- 6:31Speaker 3: The last several months. And I think that that's going
- 6:35Speaker 3: to continue.
- 6:36Speaker 1: Yeah, well, I think at least end users in the
- 6:39Speaker 1: West would be happy to be paying less. And the
- 6:43Speaker 1: fact that you can not have to pay the inference fees.
- 6:46Speaker 1: You can download it. I feel like the blockchain industry
- 6:49Speaker 1: and the crypto sector, these guys, they want more open
- 6:54Speaker 1: source models that you can own more or you can download,
- 6:57Speaker 1: but maybe they just, it's always been that matter of convenience.
- 7:02Speaker 1: And I feel like they get caught up and be like,
- 7:04Speaker 1: let's just use ChatCVT and give it all our information
- 7:07Speaker 1: and pay more because that's what other people around me
- 7:10Speaker 1: are doing. I don't know if that's the right way
- 7:13Speaker 1: people should be going about it.
- 7:15Speaker 4: Well, I mean, certainly there is a convenience factor and,
- 7:18Speaker 4: you know, there's going to continue to be use cases
- 7:22Speaker 4: for both models. But again, as we saw earlier in
- 7:25Speaker 4: this year, when the token maxing, you know, became a thing.
- 7:29Speaker 4: And then, you know, within two months when the bills
- 7:31Speaker 4: started to show up, then they started to do token minning.
- 7:35Speaker 3: And, you know, when the bills came in. So you're
- 7:39Speaker 3: going to see a lot more open markets.
- 7:41Speaker 4: Models and and you're already seeing that the open models
- 7:44Speaker 4: are are getting used more you can use the china
- 7:48Speaker 4: open models at you know aws and and at microsoft
- 7:51Speaker 4: so that the hosts the clouding the host uh companies
- 7:55Speaker 4: are are making them available deep seek even i think
- 7:58Speaker 4: is available at microsoft and and the moonshot people are
- 8:02Speaker 4: working on revenue share deals with all three of our
- 8:04Speaker 4: major cloud companies so it's it's really early in this
- 8:07Speaker 4: in the in the model race and of course there's
- 8:10Speaker 4: been a lot of noise. I mean, you know, you,
- 8:13Speaker 4: you said there's been a lot of noise, but there's
- 8:16Speaker 4: been an incredible amount of noise in the last five
- 8:18Speaker 4: days with the, you know, the, the push by, by
- 8:24Speaker 4: Dario and Sam to, to get regulated.
- 8:26Speaker 3: And then Trump, of course, you know, responding, it's a hoax.
- 8:29Speaker 4: And so it's, it's a pretty heated moment right now.
- 8:32Speaker 4: And with Xi Jinping coming to visit next week, I
- 8:36Speaker 4: think we'll see more fireworks in the next.
- 8:39Speaker 3: Couple of weeks. Definitely.
- 8:40Speaker 1: I think we're going to be seeing a lot of
- 8:41Speaker 1: headlines about China AI, and that's why this conversation is
- 8:46Speaker 1: very timely.
- 8:47Speaker 2: And there was also an.
- 8:52Speaker 1: Announcement of some sort from the leading government bodies in
- 8:56Speaker 1: the US putting out an advisory for the Chinese models,
- 9:02Speaker 1: saying there was distillation, whether they were basing it off
- 9:05Speaker 1: the American models or you know, there's just putting out
- 9:08Speaker 1: a warning, but.
- 9:09Speaker 2: I don't know if the government loves to put out warnings.
- 9:12Speaker 1: What are your thoughts on, should people in the West
- 9:16Speaker 1: be cautious about these China AI models and from what
- 9:19Speaker 1: aspect of that?
- 9:22Speaker 4: I don't think there's any reason to, or need to
- 9:24Speaker 4: be cautious on these China models any more than any
- 9:28Speaker 4: other providers models. In terms of the question of distillation,
- 9:33Speaker 4: I mean, look, The entire AI industry is based on distillation.
- 9:38Speaker 3: I mean, they distilled everything off the Internet.
- 9:40Speaker 4: And by the way, they've been sued by lots of
- 9:44Speaker 4: people whose IP was stolen or distilled. And distillation is
- 9:51Speaker 4: also a common practice amongst our labs. And I can't
- 9:58Speaker 4: speak to the veracity of some of the claims. of distillation,
- 10:01Speaker 4: but I don't think it's, I don't think it's a
- 10:03Speaker 4: long-term issue.
- 10:05Speaker 2: Yeah, no, I agree.
- 10:07Speaker 1: And it, you know, you could put it out, you
- 10:10Speaker 1: can see the warnings and you can then take a
- 10:12Speaker 1: look at the West Western models and yeah, they're based
- 10:15Speaker 1: off of scraping our information off the internet and, And
- 10:18Speaker 1: also the information that we're freely giving them when we
- 10:21Speaker 1: check the terms and conditions box, that they're learning from
- 10:24Speaker 1: whatever we say to make themselves smarter and building their
- 10:29Speaker 1: own models. The frontier is in a recursive build, so
- 10:34Speaker 1: it keeps building from itself. And there's been discussion around
- 10:36Speaker 1: that recently, that that is part of the issue, that
- 10:41Speaker 1: it's moving too fast because of the recursive training.
- 10:45Speaker 2: What are your thoughts on that?
- 10:47Speaker 4: Well, I think that it's good that, again, this is
- 10:54Speaker 4: so much happening in AI so quickly.
- 10:56Speaker 3: It's hard for anyone to really keep their arms around.
- 10:59Speaker 4: But I think that the move that basically started last week,
- 11:03Speaker 4: I mean, you had the guy that quit and said,
- 11:06Speaker 4: everyone's going to die in five years and no one's
- 11:09Speaker 4: taking it seriously.
- 11:10Speaker 3: And people have done that before. But for whatever reason, last.
- 11:13Speaker 4: Week, it did sort of jar up you know, the
- 11:16Speaker 4: main players in this, Dario, Sam and Elon, to agree
- 11:21Speaker 4: that we really should have some procedure to, you know,
- 11:28Speaker 4: review at least the models before they're released. And this
- 11:31Speaker 4: makes a lot of sense. You also have on top
- 11:36Speaker 4: of that, you know, Dario calling for government regulation to
- 11:42Speaker 4: slow down. And they got pushed back pretty hard by
- 11:44Speaker 4: the president and by Sachs. And they basically said, look,
- 11:50Speaker 4: do it yourselves, figure it out. And Elon had been
- 11:53Speaker 4: calling for this previously. And the idea is, look, all
- 11:57Speaker 4: of these guys are in their tunnel. They're doing their thing.
- 12:00Speaker 4: They have their models. You just get locked into something.
- 12:04Speaker 4: And sometimes it's good to have a fresh set of
- 12:07Speaker 4: expert eyes come in and review. And look, it's like proofreading,
- 12:11Speaker 4: you know, if you're writing a book or something, right,
- 12:12Speaker 4: it's good to have other people come and take a
- 12:14Speaker 4: look at it and, you know, give you a reality check.
- 12:16Speaker 3: So I think that.
- 12:18Speaker 4: Those efforts seem to be moving very quickly. And I
- 12:21Speaker 4: think that's a good place to, to at least try
- 12:25Speaker 4: to minimize, uh, you know, risk associated with things like, like, uh,
- 12:31Speaker 4: recursive alerting. Um, And importantly, you also saw Elon Musk,
- 12:37Speaker 4: you know, has been saying, we need China to be
- 12:40Speaker 4: involved with this as well.
- 12:41Speaker 3: Right.
- 12:42Speaker 4: The Chinese models also need to participate in this sort
- 12:45Speaker 4: of industry group. And that's why this summit that's coming
- 12:49Speaker 4: up a week from tomorrow with the president and Xi Jinping,
- 12:55Speaker 4: this could be one of the most important summits in
- 12:58Speaker 4: modern history. I mean, there's a lot going on. A
- 13:02Speaker 4: lot of people are framing this as a race, as
- 13:05Speaker 4: a battle between AI and China. And I think that's
- 13:08Speaker 4: the wrong way to think about it. I mean, this
- 13:10Speaker 4: is a wave, right? Nobody won electricity, right?
- 13:16Speaker 3: We all won.
- 13:17Speaker 4: So I think that if the US and China, because ultimately,
- 13:22Speaker 4: we were the two main superpowers to start with, and
- 13:26Speaker 4: we are clearly the two AI superpowers. And I think
- 13:30Speaker 4: it's very important that we have some alignment and safety
- 13:36Speaker 4: considerations and some sort of apparatus put in place so
- 13:42Speaker 4: as things continue, we can better keep our arms around it.
- 13:47Speaker 2: Yeah, it's a great point, Kevin.
- 13:49Speaker 1: And to that, about limiting the recursive models and people
- 13:55Speaker 1: considering it a battle between the US and China, part
- 13:58Speaker 1: of the thought was that if we do limit it,
- 14:00Speaker 1: then China's models are going to get ahead. And to
- 14:05Speaker 1: your point there about everyone wins when there's electricity.
- 14:08Speaker 2: Do you think that that's an.
- 14:09Speaker 1: Issue with China models getting ahead or does it matter
- 14:14Speaker 1: whose models are better?
- 14:15Speaker 3: Well, here's what I would say.
- 14:18Speaker 4: First of all, I think that the reality is, and
- 14:22Speaker 4: one of my heroes is this man, Nanda Nilakani, who's
- 14:24Speaker 4: one of the founders of Infosys, India's largest tech company.
- 14:28Speaker 4: And as he has said for a long time now,
- 14:32Speaker 4: we don't need the models to get any better for
- 14:34Speaker 4: 10 years. Like for, for the vast majority of things
- 14:38Speaker 4: that AI can do, it's already good enough. And it's
- 14:41Speaker 4: more about figuring out for companies to figure out how
- 14:44Speaker 4: to use it.
- 14:45Speaker 3: Right.
- 14:45Speaker 4: They don't need the latest version of, of, of, you know,
- 14:50Speaker 4: Claude or the latest version of open AI models.
- 14:54Speaker 3: To do almost all of the practical things that need
- 14:56Speaker 3: to be done. So, There will.
- 14:58Speaker 4: Continue to be this race or just drive to improve
- 15:03Speaker 4: the models, improve efficiency. That's one of the other things
- 15:05Speaker 4: that the China models have become is a lot more efficient,
- 15:10Speaker 4: largely because they were forced to, because they didn't have
- 15:12Speaker 4: access to the NVIDIA's leading chips. And so they've been
- 15:17Speaker 4: very crafty in terms of finding ways to tweak the
- 15:24Speaker 4: models and the hardware so that they can compete. And
- 15:29Speaker 4: it very well may be that China models will catch
- 15:31Speaker 4: up with ours. I mean, they're pretty darn close now.
- 15:35Speaker 4: I know last or a couple of months ago, someone
- 15:39Speaker 4: tweeted to Elon about when would the China models catch up?
- 15:44Speaker 4: And his response was first quarter of 2027. And the
- 15:51Speaker 4: founder of Z.AI responded probably sooner. So Look, China's got
- 15:58Speaker 4: more AI patents by a lot. They have more AI scientists.
- 16:02Speaker 4: A lot of the leading scientists in the U.S. labs
- 16:05Speaker 4: are from China. China leads in almost every leading technology.
- 16:10Speaker 4: And even Jensen Wong has stated that he thought China
- 16:14Speaker 4: was going to win AI.
- 16:16Speaker 3: Now, he got.
- 16:19Speaker 4: Booed, basically, by the media after that, and NVIDIA had
- 16:22Speaker 4: to put out.
- 16:23Speaker 3: A press release. But you know, China is very capable.
- 16:26Speaker 4: And I, and again, I don't, I don't think we
- 16:28Speaker 4: should think of it as a race, but I think
- 16:31Speaker 4: that China's very likely to end up leading in AI
- 16:35Speaker 4: as they lead in almost everything technology.
- 16:38Speaker 5: At this point, other than, other than EUV, um, the
- 16:44Speaker 5: extreme ultraviolet lithography, which is what makes, you know, TSMC and,
- 16:50Speaker 5: and Nvidia happy, you know, Cape, um, That's the real
- 16:55Speaker 5: machine that powers the advanced chips, which comes from the
- 17:00Speaker 5: ASML in the Netherlands.
- 17:02Speaker 3: Yeah.
- 17:03Speaker 1: I'd love to jump back into the ETF and sort
- 17:06Speaker 1: of the China AI models that are in there. You know,
- 17:09Speaker 1: we've mentioned and people understand from the end user, it's
- 17:13Speaker 1: much cheaper, much, much cheaper. And the production of the
- 17:18Speaker 1: models was at a fraction of the cost, as we
- 17:20Speaker 1: saw with the DeepSeek, you.
- 17:22Speaker 2: Know, 6 million versus... hundreds of millions.
- 17:24Speaker 1: I don't know what the exact amount it had cost
- 17:26Speaker 1: open AI at the time, but when you compare those numbers,
- 17:30Speaker 1: how much people are paying versus how much it costs
- 17:32Speaker 1: them to make, what does the profitability look like for
- 17:36Speaker 1: investing in these companies versus investing in the U S companies, which,
- 17:42Speaker 1: you know, we we've seen people in San Francisco are
- 17:44Speaker 1: dying to get equity in some of these private companies,
- 17:46Speaker 1: you know, selling their San Francisco Bay house to, to
- 17:50Speaker 1: get anthropic stock. But in terms of the profitability of,
- 17:53Speaker 1: and long-term profitability, importantly, also. How does that look like
- 17:58Speaker 1: for the models in the ETF? Sure.
- 18:01Speaker 4: Well, so the Tigers ETF, TGRZ, which is China's models,
- 18:07Speaker 4: which right now there's three holdings in this, is a
- 18:08Speaker 4: very narrow group of companies. It's a very volatile group.
- 18:13Speaker 4: And they're not profitable. And they have, just like our labs,
- 18:16Speaker 4: they have a lot of revenue growth. They think they
- 18:20Speaker 4: can find a way to profitability. But right now, it
- 18:23Speaker 4: remains a big question mark. So these are amongst the
- 18:27Speaker 4: riskiest investments you can make right now, or especially are
- 18:29Speaker 4: these China models. And again, the volatility reflects that. Now,
- 18:37Speaker 4: we also have AICH, which is the China full-stack ETF.
- 18:40Speaker 4: So that's not just the models. That's where you get power,
- 18:45Speaker 4: so CATL, contemporary amperage, which is a leading a battery
- 18:49Speaker 4: company in the world. On the chip side, everyone knows
- 18:54Speaker 4: NVIDIA and Intel, but they probably don't know Haigon. They
- 18:57Speaker 4: probably don't know Montage Technologies. They probably don't know the
- 19:02Speaker 4: China's leading chip companies. China has its own group of
- 19:06Speaker 4: companies that are competing with NVIDIA now and AMD.
- 19:10Speaker 3: Two of those four companies were.
- 19:12Speaker 4: Founded by NVIDIA's China head and AMD's China head. So
- 19:16Speaker 4: they come with good pedigree. You have optics makers, which
- 19:22Speaker 4: China leads the world in optics. You've got E-OptoLink, for example, InnoLight,
- 19:28Speaker 4: the other leader. And then the models, which we talked about.
- 19:32Speaker 4: And then finally, you get to the application layer. And
- 19:35Speaker 4: that's really the most important layer is how do we
- 19:37Speaker 4: take all this other stuff and create value with it?
- 19:40Speaker 4: And that's where everyone's early, right? You have only a
- 19:44Speaker 4: few sort of AI native apps, Cursor, for example, that
- 19:51Speaker 4: has been acquired, but there's no great example yet of
- 19:55Speaker 4: an application company. And one could argue that it's the applications,
- 20:00Speaker 4: the super apps of China like Tencent and its WeChat
- 20:03Speaker 4: app that will ultimately win at the application there, as
- 20:07Speaker 4: may Apple here in the.
- 20:09Speaker 3: U.S., at some point.
- 20:12Speaker 2: Yeah, we'll see how that plays out with Apple and Google.
- 20:15Speaker 2: I feel like.
- 20:17Speaker 1: They've been a little bit behind compared to some of the,
- 20:21Speaker 1: you know, the two.
- 20:22Speaker 2: Main ones in the US. And I don't know how.
- 20:27Speaker 1: The China-US race will fit into that, but I'm eager
- 20:32Speaker 1: to see, you know, they really own the devices that
- 20:35Speaker 1: everyone is using. How will they take back control on
- 20:41Speaker 1: the AI front.
- 20:43Speaker 3: Yeah, well, ultimately, you know, you need distribution.
- 20:46Speaker 4: And if you've got a billion people holding a device
- 20:49Speaker 4: that you've made and runs on your software, that's a
- 20:52Speaker 4: pretty good head start. That's another reason why Tencent, I think,
- 20:55Speaker 4: remains very interesting. It's got one and a half billion people.
- 20:59Speaker 4: Tencent's main business is WeChat, which is like the Facebook
- 21:03Speaker 4: of China and other parts of the world.
- 21:05Speaker 3: And so they have a great and they're already deploying it.
- 21:09Speaker 3: I mean, they, you know, we talked.
- 21:09Speaker 4: About this before, but the QCLAW, which is their OpenCLAW
- 21:15Speaker 4: agent is available to all one and a half billion
- 21:18Speaker 4: people on the app now. So, but I think we're
- 21:23Speaker 4: still pretty early in that application layer.
- 21:26Speaker 1: And of the AI models, you know, DeepSeek and Moonshot,
- 21:30Speaker 1: these are also still earlier stage private companies from what
- 21:34Speaker 1: I understand. Does the future of the ETF profitability or
- 21:39Speaker 1: where you see it depend at all on how quickly
- 21:43Speaker 1: these companies grow and if they go public?
- 21:46Speaker 4: Well, so right now you could say the three leading
- 21:50Speaker 4: labs in China are DeepSeek, Z.AI and Moonshot. And Z.AI
- 21:57Speaker 4: is public. It's the largest holding in TGRZ. Moonshot is
- 22:02Speaker 4: an interesting story. Moonshot was actually founded by a quantitative
- 22:04Speaker 4: hedge fund. that happened to have a lot of Nvidia
- 22:09Speaker 4: chips for their money management business. The founder was a
- 22:13Speaker 4: sort of a math prodigy from China who wanted to
- 22:17Speaker 4: become a quantitative hedge fund. And he did that very successfully,
- 22:21Speaker 4: but they never took any outside funding until just a
- 22:24Speaker 4: couple of months ago, they raised $ 7 billion and the
- 22:26Speaker 4: founder put in 3 billion of his own money. So
- 22:29Speaker 4: his hedge funds obviously doing pretty good. So they are
- 22:33Speaker 4: moving towards an IPO, which should happen in the early, 2027.
- 22:37Speaker 4: And then Moonshot and the Kimmy app, they may go
- 22:40Speaker 4: public this year. They've already, I think, started to file
- 22:44Speaker 4: for a Hong Kong IPO.
- 22:46Speaker 3: And so we'll add those as the IPO.
- 22:51Speaker 4: We may also, the fund has the ability to invest
- 22:55Speaker 4: up to 15% in private pre-IPO companies. And so as
- 22:59Speaker 4: the fund scales up, we'll look and see if that's
- 23:02Speaker 4: an opportunity we can find as well.
- 23:04Speaker 3: Hmm.
- 23:05Speaker 2: That's exciting.
- 23:06Speaker 1: And what do you think are the thoughts of U.S. investors,
- 23:12Speaker 1: maybe not ones that are specifically focused on AI, but
- 23:16Speaker 1: branching out into more alternative investments like the China models?
- 23:21Speaker 1: Do you think they, is there a sense of nationalism
- 23:26Speaker 1: or misunderstanding that maybe there's upside, but it's scary territory
- 23:31Speaker 1: to invest in these models that are competing with the U.S.
- 23:36Speaker 4: I think that people in the United States in particular,
- 23:42Speaker 4: a lot of them have biases against China. The one
- 23:45Speaker 4: thing I've learned investing in emerging markets, because we don't
- 23:48Speaker 4: just invest in China, we invest in India and Brazil
- 23:52Speaker 4: and any emerging or frontier market is part of our universe.
- 23:56Speaker 4: It just so happens in the case of AI, there's
- 23:58Speaker 4: really only two players.
- 24:00Speaker 3: It's the U.S. and China.
- 24:02Speaker 4: And then Taiwan and Korea, which are emerging markets, are
- 24:05Speaker 4: making hardware, but they're not really in the real.
- 24:09Speaker 3: Part of the story. But the thing is, most people
- 24:13Speaker 3: have never been to China, right? They've seen headlines. They
- 24:18Speaker 3: hear stories.
- 24:20Speaker 4: Because it's communist, people assume, oh, it has all these
- 24:23Speaker 4: oppressive elements. And I think people would be stunned if
- 24:27Speaker 4: they went to China. I mean, I tell people, especially
- 24:30Speaker 4: if they're young and don't have a family and a
- 24:33Speaker 4: dog or a cat to kind of hold them down,
- 24:35Speaker 4: it's not hard to get to China. And it's quite
- 24:38Speaker 4: cheap when you get there. And it's amazing. It's really
- 24:42Speaker 4: the most advanced country in many, many ways. I mean,
- 24:46Speaker 4: this is a terrible analogy, but if the aliens decided
- 24:50Speaker 4: they finally wanted to land on the planet and talk
- 24:53Speaker 4: to the leaders and they sort of scoped out which
- 24:56Speaker 4: country just looking at the cities, it's pretty clear they'd
- 25:00Speaker 4: probably go to Chongqing or Shanghai because, I mean, these
- 25:03Speaker 4: are incredible places.
- 25:06Speaker 3: So I recognize that there's.
- 25:08Speaker 4: A fear level that people have with China, but I
- 25:13Speaker 4: think a lot of it's unfounded myself.
- 25:17Speaker 3: Definitely.
- 25:17Speaker 1: Yeah, I have been, I spoke at a blockchain conference
- 25:20Speaker 1: there nearing 10 years ago now in Beijing. So I
- 25:24Speaker 1: did get the chance to experience it, but I know
- 25:27Speaker 1: it's so big. That was just one dot on the map.
- 25:31Speaker 1: So have you personally, as you're talking about China, visited
- 25:35Speaker 1: the companies or the warehouses and, and talk to them? And,
- 25:40Speaker 1: and do you think that's an important factor for building
- 25:43Speaker 1: this fund and presenting it to us investors?
- 25:47Speaker 4: We have admit with all of the companies, but I
- 25:50Speaker 4: have spent lots of time on the ground in China
- 25:53Speaker 4: over the last 20 years since I got involved. Um, Um, and, uh,
- 25:59Speaker 4: you know, for most of the last, um, including still today,
- 26:04Speaker 4: our primary focus has been on the emerging market internet sector, right?
- 26:09Speaker 4: I mean, I, I launched the first China technology ETF
- 26:12Speaker 4: in 2010 with, which has a ticker CQQQ. So that
- 26:16Speaker 4: was my first launch of a emerging markets tech focused product.
- 26:22Speaker 3: And that was in 2010. So 16 years ago.
- 26:23Speaker 4: So, um, But starting around 11 years ago, what I
- 26:33Speaker 4: realized was all the things that were happening in China,
- 26:36Speaker 4: Alibaba and Tencent, those things were starting to spread.
- 26:39Speaker 3: So I launched EMQQ, which is our.
- 26:42Speaker 4: Sort of flagship fund, which does emerging markets internet. I
- 26:47Speaker 4: launched that to capture things like MercadoLibre, the amazon.com of
- 26:52Speaker 4: South America. companies like C Limited in Southeast Asia. So
- 26:57Speaker 4: I've been focused on that, you know, the technology in
- 27:00Speaker 4: emerging markets from the time it basically started, before there
- 27:04Speaker 4: were smartphones, basically. So I, but I haven't made a
- 27:12Speaker 4: trip specifically to go and visit a lot of these
- 27:14Speaker 4: hardware companies, which are part of the EMX portfolios, the
- 27:20Speaker 4: AI-focused portfolios. Mm-hmm.
- 27:24Speaker 1: What are some of your thoughts or predictions on how,
- 27:29Speaker 1: not that there's a race, but the AI race moving
- 27:32Speaker 1: into 2027, and you sort of referenced that one tweet
- 27:36Speaker 1: from the Z.AI and Elon Musk, but are there any
- 27:40Speaker 1: other thoughts and predictions on how the models will evolve
- 27:43Speaker 1: in each of the countries by next summer?
- 27:48Speaker 4: Well, I mean, one thing that is just, about this
- 27:54Speaker 4: AI thing as compared to, say, the internet.com thing. The
- 28:00Speaker 4: speed and breadth of developments, it's staggering.
- 28:07Speaker 3: I mean, consider this.
- 28:08Speaker 4: A year ago, if you asked almost anybody, OpenAI was winning, right?
- 28:16Speaker 4: That was the US model that was going to, you know,
- 28:20Speaker 4: was the leader. And then, Now, less than 12 months later,
- 28:25Speaker 4: people are saying, well, Anthropic's worth twice as much as opening.
- 28:29Speaker 4: Some people have said that. But consider this. One of
- 28:33Speaker 4: the main reasons Anthropic has leapt ahead is the coding product,
- 28:39Speaker 4: which wasn't even a plan. It was a product that
- 28:42Speaker 4: one of the engineers made for his own use in
- 28:46Speaker 4: his spare time. And a couple of his buddies borrowed
- 28:51Speaker 4: it or wanted to use it. And all of a sudden,
- 28:53Speaker 4: it's one of the biggest parts of the company. And
- 28:56Speaker 4: that's been this.
- 28:57Speaker 3: Calendar year that that happened.
- 29:00Speaker 4: So trying to predict where it's going to be a
- 29:03Speaker 4: year from now is quite difficult. I guess I would
- 29:09Speaker 4: probably be pretty safe to say we'll have more U.S.
- 29:12Speaker 3: Open models. Right? We're already seeing.
- 29:17Speaker 4: I guess that's one prediction I would make, that there
- 29:19Speaker 4: will be more.
- 29:23Speaker 3: Open models.
- 29:24Speaker 4: And I think there's going to be lots of other
- 29:27Speaker 4: smaller models and more use case specific models. But as
- 29:34Speaker 4: you know, as we've touched on, everything's just happening so fast.
- 29:37Speaker 4: In fact, Demis Asavis, who I'm sure you know, and
- 29:42Speaker 4: your listeners, he's the founder of DeepMind, which was, and
- 29:45Speaker 4: he's now the chairman of the DeepMind part of Google.
- 29:51Speaker 4: He was interviewed a couple of months ago, and after
- 29:55Speaker 4: he gave a talk, actually, and then they had a
- 29:57Speaker 4: Q & A afterwards. And the first question was, What
- 30:00Speaker 4: advice would you give to a college graduate, right, that's
- 30:03Speaker 4: coming into this AI world? And he said the same
- 30:06Speaker 4: thing everybody says. He said, I would learn how to
- 30:09Speaker 4: use all the different models. I would become the best
- 30:12Speaker 4: at prompting. I would just make myself really good at
- 30:15Speaker 4: using AI. Now, that was a pretty standard question. But
- 30:18Speaker 4: then he made the following comment.
- 30:20Speaker 3: He said, you know.
- 30:22Speaker 4: I don't even know what's going on because I have
- 30:25Speaker 4: a job. I can't keep up with the models that
- 30:28Speaker 4: have already come out, let alone the five or 10
- 30:31Speaker 4: new things that are going to come out today. So
- 30:33Speaker 4: everybody's sort of in this whirlwind and it's hard to
- 30:37Speaker 4: make predictions because things are going so quickly.
- 30:41Speaker 2: Definitely.
- 30:42Speaker 3: I agree.
- 30:43Speaker 1: And there's going to be a lot. We just don't
- 30:45Speaker 1: know what they're going to stumble upon and that's going
- 30:50Speaker 1: to take off and that will be the thing like
- 30:52Speaker 1: you said, about the coding product. And with the fund,
- 31:02Speaker 1: what's the best way for people to learn more about
- 31:05Speaker 1: the fund and access it for U.S.
- 31:08Speaker 3: Investors?
- 31:09Speaker 1: Sure.
- 31:09Speaker 4: Well, EMX ETF is the website. That's EM as in
- 31:15Speaker 4: Emerging Markets X and then ETF. And then you'll see AICH,
- 31:21Speaker 4: which is the full stack China. So it's a 24 company,
- 31:25Speaker 4: top to bottom portfolio of Chinese AI companies. And then
- 31:29Speaker 4: the Tigers TGRZ you'll find. And again, that's just the
- 31:33Speaker 4: currently three leading companies in the model space.
- 31:39Speaker 1: That sounds great. Well, I'm definitely going to be looking
- 31:41Speaker 1: into it and following it. There's an opportunity that, I
- 31:47Speaker 1: feel like a lot of US investors that are dying
- 31:50Speaker 1: to get into US AI companies that are already a
- 31:53Speaker 1: trillion dollar valuation.
- 31:55Speaker 2: And there's other.
- 31:57Speaker 1: Opportunities outside of this country also that are not at
- 32:01Speaker 1: a trillion dollars yet and they have potential upside. And
- 32:04Speaker 1: I think, you know, whether it's nationalism or misunderstanding of
- 32:09Speaker 1: other cultures and what's going on in other parts of
- 32:12Speaker 1: the world might limit people from having potential opportunities bigger
- 32:17Speaker 1: returns elsewhere, or at least better diversification of their portfolio.
- 32:21Speaker 1: So I really appreciate your insights into the global AI
- 32:27Speaker 1: growth and what we're seeing outside of the US.
- 32:31Speaker 2: And I feel like not enough people are paying attention
- 32:33Speaker 2: to this yet.
- 32:36Speaker 3: Well, it's not going away.
- 32:38Speaker 1: That's for sure. Well, thank you so much, Kevin, for
- 32:42Speaker 1: taking the time. I will leave a link to the
- 32:45Speaker 1: to the, to the platform and all of the other
- 32:47Speaker 1: ETF information as well. And for EMQQ global in the
- 32:51Speaker 1: show notes below, and we'd love to follow up in
- 32:54Speaker 1: the near future. There will be much more to talk
- 32:56Speaker 1: about in the next AI with updates. You know, we
- 33:01Speaker 1: could be back next week and there'll be hours to
- 33:04Speaker 1: talk about, but when there's more on, on the Tigers ETF,
- 33:08Speaker 1: I think that's a, a great part to, to follow
- 33:14Speaker 1: as these other, China AI companies grow and see how
- 33:17Speaker 1: those ETFs evolve. So congrats on the launch of that
- 33:20Speaker 1: and looking forward to following up in the near future.
- 33:23Speaker 3: All right. Thanks, Ashton.