Humanoid Robots Will Replace Millions of Workers — Are You Invested? | Derek Yan

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

Transcript

Welcome And Guest Introduction

SPEAKER_01 0:04

But uh with all that said, my name is Michael Guy, and I'm the publisher of the Lead Lag Report, and of course the founder of Lead Lag Media. Uh, and now uh getting a little bit back into the lead lag live uh podcast side of things, uh Mr. Derek Jan of CraneShare. Before we get into uh the topic everyone wants to talk about, which is war and uh maybe how robotics are playing into warfare, uh maybe a quick intro as far as uh who you are and a quick intro into COID itself.

SPEAKER_02 0:29

Sure. Thank you, Michael. Uh thank you for having me. Like me. Um my name is Derek. I'm a senior investment strategist at Crane Shares. So Crane Shares is um a US-based asset manager. We match over like$10 billion across like around 30 to 40 ETFs. Uh COID is our latest uh launch on emerging technology. Uh we believe that AI is going from digital to physical, uh, and like humanoid, quadruped, all kinds of embodied uh technology is kind of like a nest wave of the AI development. So that's where we launched KOID, COID FUND to really tap into the ecosystem of humanoid and physical AI. Um, so I mean happy to have a discussion on the topic, but like this is like signing timing for this.

Drones Rewrite War Economics

SPEAKER_01 1:17

Yeah, and it's um it's interesting, right? Because I think when people think about uh humanoids or robots, they're thinking about uh what Musk is talking about, right? It's like, you know, it's gonna eliminate poverty and money's not gonna even be a thing, and we'll then be living fat and happy as these robots uh take over our day-to-day activities. But uh we should talk about how maybe uh it's advanced warfare, uh, maybe sort of the the aspects of how robotics and drones and all that are factoring to what we're seeing with Iran uh and the US here. Um first of all, is it fair to say that um military conflict is forever changed? It's not it's not a dynamic of necessarily boots on the ground, it's a function of who has the best technology.

Quadrupeds And Humanoids In Combat

SPEAKER_02 1:55

I think like the technology definitely creates some nuances in the war, right? Like I think starting from the Ukraine war, you can't like see the importance of drones. Uh and then now if come to Iran, uh you can like see Iran is using those very low-cost drones, right, then as like a weapon. Then if US wanna defense the um, say the strait, like you need like millions, like you if you attack that use like a million-dollar missile. That doesn't really make sense. So it's really about like I think going forward, affordability is kind of like uh you think about how the economy, the war is gonna become, right? You don't want to burn another like$100 billion on the war um because that's not sustainable. If you think about like Iran's gonna use very affordable drones to keep attacking. Um so I think that's kind of like the where you see the scene start to change, right? Like going forward, the importance of those um robotics is gonna be um bigger because if you think about like now the drones dominating airspace, but how about like on the ground, right? So I think like now like I don't think like US government or just any government want to send massive troops um to like a modern war because is this just not a fair way, like human fighting all those like technology machines. Um first you have I think Quadrupat, that's a really upcoming. I think Quadrupad, they already have like capability to really go into the buildings, right? Like going to all kinds of terrains. So that's something potentially can be um more on the ground troops. Um so attached like bomb, attached like a machine gun. Um you have seen all kinds of videos or film uh about like using quadruple in in warfare. So so that's kind of like really already happening. Um more like futuristic, I think, like is really the humanoid, right? Like you think about a humanoid um can do a lot of things like together with human troops or um going to like dangerous places um for human. Um just that's kind of like another another uh application, I would think, for humanoid. Um a lot of purchase, definitely research ongoing by the government and the defense companies. Um so I think that's really changing the future for, right? Think about like going forward, uh, you're gonna have increasingly more robots fight for human. Um apparently, like, we don't want to see like robots fighting human, right? So that's kind of like the the least thing we want to see. But like that means like even countries like Iran, right? They're not really advanced economy. They can make very like pretty affordable robotics. And um those robotics is really um driven by very like old technology, even, right? So hardware is very advanced, very affordable. So the shape of war can dramatically change going forward.

SPEAKER_01 5:04

How much um uh influence is there by the US and and maybe other developed economies on the suppliers of these drones, right? So if the future of warfare is around drones and robotics and it's gonna be drones fighting drones and robots fighting robots, um, it seems to me that um uh the edge will ultimately go to the the uh the country that basically prevents those those key supplies from going to Iran or from other you know to other countries that are trying to take advantage of that technology.

The Hidden Robotics Supply Chain

SPEAKER_02 5:32

Yeah, exactly. I mean, like currently I don't think this whole supply chain is ready for that because at first, like we still don't have like a massive like war or just like war preparation globally. Um, but even that, like the global defense budget is going up, like in Europe and US, that's kind of like the trend, right? So I think one thing that a lot of decision makers probably they didn't realize that how important those really like low even compared to like missiles or like super hypersonic jet, the drones and robotics, not that expensive, right? So but you need a lot, right? You need to defend um the the other countries having those, right? So to really get ready for the supply chain of the drones for the like quadruped and robotics, that's something quite critical, right? Think about like using um kind of like reducers models and certain batteries for data setup, right? Then you need a lot of actually rivers um for those kind of weapons as well. So that's those really critical. I think that's really, I think uh already got recognized by investors on the railroad side, but and many other components, many other supply chains still under the water. Um so so that's kind of like one area we believe could potentially uh get a lot of value uh going forward, not only because of, well, first, there's many other consumer applications on humanoid and culture pet, um, which is now taking off on on the uh application side, but going forward, potentially the common poachers could be another um big um I think rounding growth for those companies as well.

SPEAKER_01 7:21

Supply chain aspect of it is where it gets interesting. And then I think when um at least when I think about humanoid robots, I think about you know metals, right? Um because you know, and and whatever actually goes into the actual body of it. Um how how flexible are these supply chains? Meaning, you know, if if if we're gonna ultimately keep on going and progressing towards this path of humanoids and robotics, um is there enough stuff available to actually create the these robots?

SPEAKER_02 7:51

Yeah, that's a good question. I mean, like some are, some not, right? Think about like where basic commodities like Lumina, that's probably easy, right? So if you think about like the components part, so what components you need, you need uh models, right? You need actuators. Um the most important part is the actuators because that's the define the f uh the freedom, like kind of like how many, how many like degrees of freedom for robotics, right? Think about like quadruple pad, think about humanoid, you need a lot of flexibility, uh agility. So that's kind of like a very important part. Um then it comes to the reducers, the bearings, uh the wall screw, uh, all those components. Uh, I don't think that's really already being used in some like electric vehicles, um, and other kind of like industries as well, um, uh aerospace. So so those components uh I think is very critical. Um you need to invest a lot of those um supply chain to really make not only like jubino and culture better, but like EVs. And think about like that's that's kind of like where um uh like US and Europe is really um needed, right? So those components um together with the the power, right? Think about the batteries, think about the battery materials, those those come like a very long supply. If you think about like one company, I think the CEO of a humanoid company, they post on social media that there's like thousands of components, and within that thousands of components, there's probably like 20 to 30 countries. So it's it's very dynamic and very global supply chain so far. Um, I mean, like definitely like you every country wanna be very strategic on this, but end of the day, you need to manage this global supply chain.

SPEAKER_01 9:43

I uh I I quickly tried to create an image of an actuator just to kind of visually show this. Uh uh I you uh you would know if this is obviously more accurate uh or not. And yes, I used AI real time as we're doing this. Um is it is this is this accurate? I mean, this this looks pretty um pretty intense.

Actuators And Precision As Moats

SPEAKER_02 9:59

Yeah, this is pretty intense. Um there's uh there's a lineary, there's like a rotary, um uh there's linear uh actuators. So there's many many kinds of actuators um for the robots, for the humanoid. Um so if you think about it, like it's not really just have it, you have to really do the precision control, you have to build a ride, right, like within the your ecosystem of like your components. So I think like there's certain companies now controlling the edge because like it's not really about to have it, but it's what's the quality, right? How precise how precise this can be, because it's very important when think about certain movement, action, motion, those need very precision um control. So so that's kind of like the key for a lot of companies. They have the know-how, they have experience of that technology for I mean 10, 20 years. And because they have been doing that for aerospace, they have been doing that for EVs. So that leaves um a very certain edge, I would say, competitive edge for those companies to um keep dominating in certain component side.

SPEAKER_01 11:14

I gotta imagine it is wildly difficult to even figure out what kind of companies are involved in all these little nuances, right? So um in the context of your ETF COID, let's talk about how you go about uh even identifying the companies that you know are so necessary for all these little components.

SPEAKER_02 11:29

Yeah, exactly. So when we design COID uh index uh together with our index provider, uh we believe um that there's certain components that already have the supply chain agreement with those um humanoid companies, right? Think about like humanoid companies. Um there's so many, right? Like in the in the private space, um you have like figure AI, you have, well, Elon Musk is definitely like now doing a massive production ready. That's then that you have like eltronic sanctuary. Um so then in in in international like South Korean, you have like Rainbow Robotics, um, you have a lot of like traditional EV, like older companies in Japan, like trying to do culture pad and and robotics. Then in China you have like Unisry, you have like AGI, but so like you have so many companies um that's like doing this mass production. Uh we believe that like the companies providing the core components for those companies uh in the private space uh actually can capture a lot of value. Uh the companies uh from the actuator maker, from the the ball uh school, from the bearing companies, um, those mechanical systems, those like actuation systems, sensors, uh models, those companies actually they already have like long-term, some like have a long-term relationship with um like uh Tesla. Um some just like um starting this like tier one, tier two um supply chain test, right? Like check the box, send some sample, so they have like some agreement on sterling years. So those companies uh they already made a lot of announcements, like, oh, uh, we're the tier one, we're the tier two supplier. So I would think like those companies can be the pick and showers basically for the humanoid. If you think about humanoid, it's gonna become the next kind of like um AI, right? So the AI boom created like uh pick and shoulders benefit for like NVIDIA and those SAM conductors. We think humanoid can create like pick and shoulders benefit for those component companies.

Unitree IPO And The S Curve

SPEAKER_01 14:14

And we're obviously still very early on this. Now there is this uh IPO unitry that uh is getting quite a bit of press and headlines. Let's talk about that and how important or maybe uh seminal is that uh is that IPO.

SPEAKER_02 14:28

Yeah, exactly. I mean, like this is kind of like um we have been waiting for IPO, right? So as I said, so many private companies um in the space, in the humanoid side. Um so Unit tree is fighting, it's really setting kind of like a stage, right? Like King Cloud, um in the AI model side, like you have like Minimax, you have like knowledge atlas. Um their IPO is like setting the validation for a lot of private guys. Um so when you look at like fighting with Unitree, you realize this company is already profitable, right? Um the revenue has been growing very fast. Um I think it's 1.7 billion RB uh for the revenue, and the gross margin uh is over like 59%. Um the the revenue contribution perspective, um, I think last year is like over 60% in quadruple pad, uh, but this year the humanoid actually contributed more than 50% of the revenue. So see the rise of humanoid as the gross driver compared to quadrupad. Uh but I believe both have like application um that's gonna uh is what gonna be like widely adopted across many industries. Um but just generally I think that like if you have unitary IPO in Shanghai this year, that valuation is gonna be a reference valuation for a lot of those private unitary uh sorry, private uh humanoid companies. Um so so that's very important because like um when you have multiple humanoid companies um raising a lot of money, um, what's the next step, right? So everyone wants to be the next iPhone or the next Tesla, right, in the EV space. Um the next step for them is to really figure out how do we keep growing, right? So for them to keep growth to happen, you need mass production for their quadruped and humanoid. Um so we think now we are still at the early stage of the S curve. If you think about any new industries, right? Um for EVs, um probably the S curve is very high. The drones, that's like late stage of S curve. Um for humanoid, it's still at the beginning. So when we have those IPOs coming, then a lot of capital deployed, uh, we're gonna see a lot of production line gonna be setting up for humanoids, which can be uh the kind of like S-curve start to happen.

SPEAKER_01 17:05

How um how does intellectual um property protection play into uh some of this? Meaning, you know, uh, as I understand, a lot of obviously is is coming from China. Obviously, we have patents in the US, there's world patents. But um, how how protected are some of these companies from competitors?

Physical AI Software Breakthroughs

How COID Is Built And Used

SPEAKER_02 17:22

Yeah, I mean, like so far, China's really leading in cost efficiency, right? So the cost efficiency uh leads to very affordable hardware, right? Think about like um like Unitry, for that, but like the hardware is not really crazy expensive uh and it's very it's good performance. Um but I think like US still uh so far leading the um the model side, uh leading intelligence side, right? So there's a lot of good progress right now in the physical AI, right? Think about like um uh physical AI is really letting the humanoid, quadrupathic those robotic, understand the environment. They can reason, right? Then can then plan what's the next move for those robotics. So which is super important because the hardware is already ready. You see some performance of those like humanoid now, even can do kung fu, like flip and running. So those capabilities are already there, right? Now it just need to deploy those models into those humanoid and quadruped. Um so several advancements I've been noticing. One is like Faith Feli, the professor like Stanford, um, from the company called uh War Lab. Um so I think like that's kind of like um Faith Lee is famous for kind of like laying foundation for what large extra model is today. Um, training the digital AI, right? I think that like that's two-dimensional, right? In the image side, and there's like text side. Uh but now it's moving to 3D to 4D, um, to really training the robotic to understand uh our real three-dimensional world. So uh I think that's very natural when you have the same, same kind of like large data input to the 2D, the digital world. Now, applying that to 3D to 4D, um, the world model is very, I think, like capable going forward. Um, we're gonna see potentially some breakthrough on those. Um, and another one is the the Gin Fan from Nvidia, right? So I recently noticed like Gin Fan posted a video of those like um robotic hands uh assembling like a like a car. So that's something like very interesting with because previously we saw like humanoid culture pads, they can like they can do very big action, right? So flipping, running those like probably easier for them compared to say fix something very like um sophisticated, right? So um we like seeing that like from the gym fan um of the humanoid hand um assembling like card, that's something I think like major breakthrough. Um so I think the in this space, investors are gonna see more and more uh breakthroughs, like they kind of like seeing the AI in the digital world happening in the early days, where like Chat GPT or GPT 2 come out, then GPT 3, it's kind of like where people call like GPT moment, like then uh we we are waiting for this GPT moment happening in the physical world where those war models, where just like those um the models by startups or like incubators, can create this dramatic change to the intelligence of humanoid and culture pet. So by by that time, we think with the current advancement of the hardware, um the software can be like a major boost to the performance of the humanoid, right? Think about like um doing laundry, doing dishes, those are the early stage, but when you have the model ready, that becomes very possible.

SPEAKER_01 21:02

The last few minutes here, talk about um the the construction of COID, how many holdings, how should one thing about allocating to the ETF how dynamic or maybe not dynamic it is, any kind of context around the funnel would be helpful.

SPEAKER_02 21:14

Yeah, exactly. We think at this stage uh it's quite early. So this index is kind of like dynamic. Uh we have a quarterly rebalance framework. So each quarter we rank the company in the universe um by their relevance to the humanoid. And also look at like the supply chain criticalness. So how important this company is uh as a supplier to the humanoid company. So with those two scores, there's a rank. The 50 top 50 companies can be selected uh and it's equal-weighted. Uh we choose the equal-weighted at this stage because we see um the values really coming from all the places. Uh it's really hard to do market cap with it because that's gonna be concentrating on video or like Tesla, those like big market um uh mega uh large cap like tech names. So it's really to do market cap with it. And it's also really hard to do kind of like thematic weighting because um uh we don't think we can see like clear winners today because there's so many new competitors um coming and there's a lot of um kind of nuance in this industry. But we do see as an industry uh overall, um there's a uh great growth potential. Um so that's why a lot of the holdings actually in the component side. Um Um we I think like today we're well diversified in our geographic. It's about like one third in US, one third in Asia, then one third kind of Europe. So well diversified uh allocation and very uh diversified betting on the uh ecosystem, right? We have some names in integrators where company assemble, commercialize, and humanoid. Um there's some companies in intelligence uh providing the technology, but most of the holdings actually in the supply chain side and the body side, where they're providing the actuators, the models, sensors, critical materials, even. So it's an ecosystem that we believe is gonna benefit from the rise of humanoid. Um so that's kind of like uh the COID basket construction. Uh and we think at this stage, uh, we're gonna see increasingly more clarity uh how this industry is gonna evolve and we're gonna keep updating the methodology as well.

Where To Learn More And Closing

SPEAKER_01 23:32

Yeah, you've had some uh very solid growth on the fund, and the performance obviously has been quite strong. So um, for those who want to learn more about COID, Derek Ridge with it too.

SPEAKER_02 23:41

Yeah, so uh craneshires.com slash KOID, you can see all the COID information about uh fact sheet of presentation. Uh and we have published a lot of white papers and research articles on this topic. Um so that's cranchers.com slash koid. Um yeah, so basically we we are keep getting a lot of information on the humanoid side. This is such a new industry, I'd say, but um so far there's a lot of uh flaw leaders, right, from Silicon Valley, from the um finance world, um, is all betting on this industry. So given that, I I think this is quite an interesting idea because it's not really the technology sector, but it's also a lot of industrial. So it's not really the digital AI, it's the physical AI. So that's a natural, nested phase, right, for AI to happen. Uh, we believe physical AI and the AI happening in the in the real world, uh gonna be uh kind of like a developing phase for the next five to ten years.

SPEAKER_01 24:49

If we learn more about Floyd on cringhears.com, I'll be uh every honor if we do some of these lead leg live podcast episodes live, uh going back to that format. And I appreciate those that uh tuned in here. Thank you, Derek. I appreciate it.

SPEAKER_00 25:00

Thank you. Cheers, everybody. Cheers