All Episodes

October 8, 2025 52 mins

Imagine the internet’s 20-year curve compressed into a few years: costs plummet, capabilities skyrocket, and adoption leaps from novelty to necessity. That’s where AI stands today. In this episode, we go deep into the evolution of reasoning models—how chatbots are transforming into agents that plan, code, and act across real workflows. Then, we explore the next frontier: physical AI, humanoid robots designed to operate in human spaces, where the “final mile” of automation requires hands, balance, and judgment.

We break down the full AI stack so you can invest with intent:

  • Digital AI: From semiconductors and data centers to secure data infrastructure, code generation, and multimodal systems connecting perception to action.
  • Physical AI: Humanoid ecosystems, including integrators, brains, and bodies. We explain why actuators, harmonic drives, tactile sensors, and manufacturing clusters are today’s critical picks-and-shovels plays.
  • Investment insights: U.S. leadership in intelligence and commercialization vs. Asia’s manufacturing depth, hype vs. investable opportunities, and a practical portfolio approach balancing concentrated and globally distributed AI exposure.

 #AI #ArtificialIntelligence #HumanoidRobots #MachineLearning #AgenticAI 

Start your adventure with TableTalk Friday: A D&D Podcast at the link below or wherever you get your podcasts!
Youtube: https://youtube.com/playlist?list=PLgB6B-mAeWlPM9KzGJ2O4cU0-m5lO0lkr&si=W_-jLsiREjyAIgEs
Spotify: https://open.spotify.com/show/75YJ921WGQqUtwxRT71UQB?si=4R6kaAYOTtO2V

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


Foodies unite…with HowUdish!

It’s social media with a secret sauce: FOOD! The world’s first network for food enthusiasts. HowUdish connects foodies across the world!

Share kitchen tips and recipe hacks. Discover hidden gem food joints and street food. Find foodies like you, connect, chat and organize meet-ups!

HowUdish makes it simple to connect through food anywhere in the world.

So, how do YOU dish? Download HowUdish on the Apple App Store today: Support the show

Mark as Played
Transcript

Episode Transcript

Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_00 (00:00):
So you can see like uh this company is already
starting like a mass umproduction.
There's some like a facial umrecognition and facial like
expression.
Their motion control is reallystrong.
Like you if you kick it, if youlike move it, it's very stable.
Uh it's like they're running mymarathon.

(00:20):
And this is another interestinglike company.
You can actually use an app todo like autonomous, like just
like fully autonomous likeretail experiments.

SPEAKER_01 (00:39):
This should be a good conversation with uh Derek
Jan.
We're gonna be talking about uhsome really interesting thematic
ideas, which are getting somereally big traction in today's
environment.
For those advisors that happento be in a physical office
location, do me a favor, tellyour other fellow advisors that
this webinar is taking place.
Uh Crane Chair is one of the uhgreat issuers out there, a
number of unique and innovativefunds, or as Finra likes to say,

(01:00):
distinctive, because unique cannever be said from compliance
perspective for whatever reason,but distinctive is okay.
Uh so they've got a distinctivelineup.
Um, and uh for those that arehere for the CE credit, I will
email you all after thiswebinar.
Uh just stay to the end of thewebinar, I'll get your
information and submit it to theCFP board, and we'll go from
there.
So appreciate the uh support forthose that keep attending these
webinars.
Uh, it's quite a bit of effortand work, and my clients uh like

(01:23):
Crane Shares, obviously, arevery appreciative as I am as
well.
So, with all that said, my nameis Michael Guyad.
Uh, this webinar is sponsored byCrane Shares.
Uh, let's get started with Mr.
Uh Derek Jan and uh talk aboutuh a whole bunch of interesting
ideas.
Go ahead, Derek.

SPEAKER_00 (01:37):
Yeah, thank you, Michael.
I mean, um my name is DerekYoung, for people who don't know
about us.
Um Crane Shares, uh, we areasset manager based in New York,
uh company founded in 2013, uh,and we manage about uh 13
billion right now.
So let me uh share mypresentation today.
Um, just for I think people likebeen talking a lot about uh both

(02:02):
AI and um robotic investment.
So at CraneShares, we have donea lot of research um on this.
Me personally, I've been workinguh on with CraneShares for like
10 years, uh covering globaltechnology companies.
Um so starting from, I thinklike like previously a lot of

(02:22):
investment is really on thesemiconductors, on the SaaS and
cloud.
Uh and now like all of a sudden,like everybody's talking about
the AI robotics.
Um there's a lot of nuanceshappening, and there's a lot of
like debate, like whether it'stoo expensive right now, like
isn't more opportunity outthere?

(02:43):
What's the next big thing?
So with that, like we're like wejust want to have like a little
conversation today, just likeshow you what's our research on
this topic.
Um just before we dive into thepresentation, probably I think
like um a lot of people may haveexperienced this um before.

(03:04):
Like when we're like superyoung, like we have this like
doub like connection experience,like when the internet speed was
super low, right?
If you want to download apicture, you're gonna wait for
like five minutes.
Like, as like a young boy, I waslike, yeah, I'm like super
impatient.
Like, I like definitely want tosee like pictures the full

(03:27):
pictures by loading the half.
So I'm like screaming, likesmashing the keyboard.
Um, but like just fast forwardto the day, like the internet
speed is like crazy.
Like we can do anything.
Um, like streaming and we canlike social media, like doing
this webinar, but like like theAI now just give everybody the
capability to like you cangenerate essay, you can generate

(03:51):
like a weekend plan, recommendthe restaurants, you can now I'm
like addicted into also likegenerating pictures like from
AI.
Um, but it just like I'm nowstill waiting for like a few
minutes to to get a picturegenerated.
It's like so funny how my lifeexperience is like a loop.
Um but like is but like showingyou the reality is like the

(04:16):
internet speed is almost went uplike 100 times over the last 20
years.
And in the meanwhile, the costof data dropped 100x.
So that means like if you lookat like 2000s, you have probably
experienced like the Yahoo time.

(04:37):
Um the internet speed and it'smore expensive to go into net to
the internet, and then theapplications is only available
in those formats.
Then fast forward to 2010,you're gonna start to see a lot
of like mobile use cases as likedata become cheaper, the speed
speed becomes like higher thanYouTube, you have like Facebook,

(04:59):
it's booming, then we have now5G, everybody's streaming like
YouTube, um, and like socialmedia, TikTok, like like
probably hate like your kidslike scrolling like TikTok every
day.
But like it just imagine likethat really just happened over
20 years.
Like how fast the internetindustry has really evolved over

(05:24):
the last 20 years and created somuch new applications that's
really um paving the way for thenext generations of applications
and adoptions of the internet.
Um so that just like get likegreat similarity to what's
happening today.

(05:46):
Um the AI, we have seen over thelast few years a very similar
growth trajectory, uh, like whathappened over the last 20 years
on the internet.
Uh, if you look at theintelligence level of the AI
models, it's getting better andbetter um almost every month.

(06:08):
While the cost of the outputfrom the AI is dropping
dramatically.
So starting from the pre-trainedmodels, right, you can see from
2024 to early 2025, the cost ofthose models really dropping to
like 100x, more than 100x.

(06:31):
The intelligence level isincreasing um at like 2 to 3x.
Well, then I think starting fromlike earlier this year, you more
and more reasoning models comingout.
The O Y model, deep C RY modelsshowing you that the
intelligence level is nownip-rocking to a level that's

(06:52):
like 10x better.
And the adoption is gettingcrazy because of the reasoning
capabilities, while the cost isdropping, like what happened to
the pre-trained model.
So we start to see this similarthings happen.
I think like now going back tothe internet era, the AI era is

(07:15):
gonna follow a very similarpattern where the cost could be
probably 100x cheaper, while theintelligence level could be
1,000x better.
So if that's gonna happen,that's gonna drive a lot of the
I think future adoptions of AI.
So already we have seen a lot ofapps like in the AI, like

(07:39):
ChatGPT is probably thefast-growing app out there
getting like 100 million actmonthly active users within two
months.
We have never seen that in theinternet era.
We have the enterprise revenuescoming from like anthropic.
That is like with by like theend of 2024, their revenue is

(08:00):
only like one billion, but justlike on August this year, it's
already five billion.
So five X in seven months.
We have never seen a companygrowing like that.
So decommeralizedcommercialization of AI and the
adoption of AI, the user growth,is actually outpa outpacing the

(08:22):
internet era.
So with that, we think like wehave to really invest uh in a
way that like you have toposition your investment around
the current level or theevolution of intelligence.
If you think of previously, themost AI is really like chatbot,

(08:45):
right?
Like think about like the firstgenerations of ChatGPT, GPT-4,
is really understanding humanlanguage and then doing like one
prompt and give you an answer.
Um then starting from like thisyear, you're gonna see the AI
become like a reasoner.
You can solve problems, you canunderstand context, you can do

(09:07):
some like a chain of thoughts.
That's getting smarter, andthat's unlock a lot of options.
And also with multimodality,like the images, the videos
coming out, AI, that's reallylike creating tons of new
applications out there.

(09:27):
So we're currently at the stagebetween um AI being the reasoner
to AI being uh agent.
I think a lot of people aretalking about like may heard or
may not heard of the agenda AI.
Like AI agent is definitely thehottest keyword this year uh for

(09:47):
most companies.
So if you talk to enterpriseowners, talk to the talk to the
like technology um officers ofeach company, um, they're all
like crazily figuring out how toreally onboard the agent
solution uh to the company.
So, what is the urgentic AI?
It's really autonomouslyplanning and make decisions and

(10:12):
take actions with theenvironment.
So you actually can gather data,you can um understand your
environment, you understand yourworkflow.
So you deploy that, you can makeautonomous decisions.
That is like uh I think like alot of people question about it,
the AI adoption previously onthe enterprise level, because
oh, it's it's the end of theday, it's just like a taboa.

(10:35):
Then now with those agendacapabilities, people start to
believe this has like a realeconomic impact uh in a lot of
enterprises.
So with um, I think the currentstage, a lot of applications
really around AI coding, becausecoding is um, I think the best

(10:58):
way to integrate it um with theAI agents, because um you
probably if you try the latestmodels, um uh either from the
GPT model or the NSA big cloudmodels, though the coding
capabilities for those thelatest models is quite amazing.
It's achieving a level that'salmost like a junior to

(11:21):
mid-level software softwareengineer.
So with those capabilities um inplace, you can actually
integrate it into a lot ofworkflow that within your
organization.
Um, and going from here, we'regonna see the intelligence level
increase.
So you can actually createknowledge, it can actually

(11:43):
interact with the physicalworld, it can bring the action
data uh into the training modelor the inference model that can
actually bring the robot to AI,right?
So that's something I think isgonna happen for the next stage.
As probably heard like JensenHuang mentioned on GTC

(12:04):
conference, that robotics biggeropportunity or the next big
thing after AI.
Uh, because we have seen thesame framework that people are
putting intelligence on thedigital world, now are training
the AI in the physical world.
Um then after that, we're gonnasee potentially AGI is gonna

(12:28):
happen, artificial generalintelligence.
And the AI could potentially bean organizer for our society,
for our economy, uh justautomate all the workflow and
organize all the decision makinguh to the maximum at this
effective way.
So that's kind of like theroadmap for the AI revolution.

(12:51):
Um, and we're just just started,I think, in the level two,
between level two and levelthree.
So it makes like a lot of sense.
If you look like uh the workflowfor enterprises, um like we have
this data to showing whatpercentage of each industry now

(13:12):
uh have paid subscriptions or inmodels, platforms, and tools.
Um, so actually, it's not reallysurprised that the technology
sector uh take a far leadcompared to other industries.
The technology sectors uh haslike 72% of uh uh companies

(13:34):
actually now using AI, uhfollowed by like finance, uh a
lot of like I think both banks,insurance, and uh fintech
companies.
Uh and then the rest is reallylagging.
Uh I think manufacturing,retail, healthcare,
construction, those mostlyhappen in the physical world.

(13:54):
It's lagging, just lacking for areason.
As I said, the digital world iseasier to be automated, to be
trained using AI, because thedigital world data is so
available.
There's so many taxes, text andimages, the videos available for
people to train a model, todeploy it, and train live

(14:16):
transaction model.
So most of the AI modelsnowadays is really focusing on
the um digital world.
So I think going forward, thenthese adoptions are gonna
gradually migrate to themanufacturing side, retail,
healthcare, and construction,and our service sector in the

(14:37):
end, maybe our home.
So that um give you an example.
I think like um Anthropic, um,you may not heard of.
Anthropic is funded by theformer OpenAI um like executive.
Um they have been really focusedon the enterprise market um

(14:58):
because they are pursuing moreuh safe, ethical, uh controlled
AI.
And they're, I think one of thekey reasons is their code
generation capabilities verystrong.
Um, so the user experience fromsoftware engineer enterprises to
really get work done in in amore like stable way, that's a

(15:23):
game changer.
So that's why their API is um uhvery popular among like each
enterprise like um agenda AIadoption.
So we as I said, like theirrevenue jumped from like 1
billion last year to now 5billion uh August.

(15:43):
So um their market share ofenterprise solutions is really
increasing, um taking the maxyears from OpenAI.
So we have seen like a companylike that is really a critical
company.
Think about like the next stageof investing in AI.
Um and there's a lot of nuanceshappening, right?

(16:05):
So you think about like the AIopportunity.
Um, so currently there'sthree-layer or four layer if
you're including the modelcompanies.
There's hardware, I think a lotof people already focused a lot
of uh their investment on thehardware still, it's like the
most obvious names, like thesemis and data center companies.

(16:28):
Like um, then you have the AIinfrastructure providing the
cloud, um, providing the datapreparation, the data
monitoring, cybersecurity.
Um, so that's the infrastructurefor the model to be trained, to
be deployed, to be inferenced.
Then you have the model companythemselves, like Anthropix, XCI,

(16:51):
that is being the coreintelligence uh of the whole
ecosystem.
Um, then upon that, then youhave all kinds of applications,
either to the enterprises, um,then you have those in the
middle ground, like contextlayer provider, those just
enterprise like service providerthat's really helping

(17:14):
enterprises using AI.
Or you have um consumer-facingapplications that's using the AI
to transform their businessmodel um uh in the in either in
the healthcare sector, in theeducation sector.
So this ecosystem, we realize,is so dynamic.
Um, just like I thinkeverybody's saying, oh, model is

(17:36):
really commodity, like it'scommoditized, it's like
commodities, like there's nosense to invest in the model.
But we have we do see OpenAI,Anthropic X AI, they've been
growing, they've been expanding,they've been at like the kind of
like the core for all thoseopportunities.
And uh people abandoned some ofthe hardware, I think,

(17:59):
previously, because they thinklike it's overinvesting, like
then we don't need more chips.
Um, so focus on applications,then people realize I think this
year, applications not it'sreally lagging.
Um, so we still need a lot oflike chips, still need a lot of
infrastructure.
So our focus is back toinfrastructure.

(18:19):
And within each layer, there's alot of dynamic as well.
You think about just among theapplications, the SaaS
companies, there will be a SaaScompany that's ready for AI.
And there will be definitelylike losers.
Um, everyone's gonna, so I thinklike just in the next two to

(18:39):
three years, every company isgonna claim their AI company.
So that's a challenging thingbecause like you have to
understand like how the AI modelis gonna be integrated into each
company's business model.
So that's challenging.
Um, as that's like a challenge,like I think every investor,

(19:01):
including professional portfoliomanagers, is facing.
Um, so we believe the probablythe best way or the only way to
gain the edge um in the AIinvesting is through partners
and knowledge uh sharing withthe AI native researchers.

(19:23):
Um so we launched a fund uhcalled AGIX.
Uh it's supervised by the AInative researchers and venture
capital list, the depthinvesting, like Anthropic SAI
Problesty at an early stage.
So those people they talk to themodel companies, they talk to

(19:43):
the AI researchers uh among eachfirm.
So they understand what's thestatus for the AI deployment,
what's their model, what's whichwhat's kind of like latest
dynamic with within the firm.
So those knowledge are sovaluable.
So as when we talk to them, sowhy not just let's partner um

(20:03):
with those knowledge?
Maybe you can do it quarterly,we can put it into the index
ETF.
So that's why we launched uhAGIX, uh kind of like a first
ETF partnering with like AInative researchers uh last year.
And um then the followingdiscussions like what if we miss

(20:24):
those most critical companies,the foundational model companies
like Anthropic, like XAI.
Um, so the answer is like, whynot just put them all together
um to complete the ecosystem?
So together with them, we wewe've been able to talk to those

(20:45):
uh companies directly.
So we talked to uh Anthropic andXAI, so we being able to really
participate in their run, likefundraising.
Um we sit on the cap table uh asa shareholder.
So AGX on behalf of ClainshipsTrust, sit on CAP table of both
Anthropic and XAI.
We participated the Series Eround of Anthropic on March this

(21:09):
year, and the tender offer roundof XAI on July.
Um so we're happy to offerinvestors a complete solution um
to really navigate this dynamicof the AI development.
Um so that's kind of like a veryinnovative way, one of the first
um kind of like ETF um to haveprivate exposure uh within the

(21:34):
ETF.
Um so that's kind of like wehave more than I think like um
uh 85% into the uh public, thenyou have less than 15% into the
private to really capture thethe full opportunities that uh
within the AI ecosystem.
Um and as I said, um maybe goingto the few years later, we're

(21:59):
gonna see the AI is gonna deployto our physical world.
Um so this is so early, but likeeverybody's like now getting
attention because uh as I said,like um big tech companies,
Tesla, Nvidia, Amazon, they'llbe embedding on this category.

(22:19):
So like Elon definitely thebiggest fan of humanoid, been
talking about like their optimusum um plan like on every every
earning call.
So that's like um kind of likethe the they're gonna like the
Elon thinks a humanoid is gonnatransform the business.

(22:39):
It's not gonna be like a smartcar business, it's gonna be a
humanoid business going forwardfor Tesla.
And um Jensen, as I said, um, isreally developing models and
foundations for um the modeltrainers, um, and to create this
generalized uh robotic modelsthat can really bring the action

(23:03):
data to the model so that peoplecan train that and people can
really um navigate the humanoidwith like human demonstration,
then synthetic data in theomniverse.
So you can train the physicalmodel, like they train it in the
digital model.
Um, so like the Google, Gemini,they're they're doing like a

(23:26):
vision language action model uhwith their very advanced
multimodality capabilities.
Um so they can actually bringthe vision, the images um to the
model so they can reallyunderstand what's happening.
So there's all kinds of effortsaround humanoid uh because
there's there's a real demand,right?

(23:47):
So um the factories want it.
Um, as I said, like um the last10 years, there's a lot of
automations already happening,but still it's very
labor-intensive for a lot offactories.
Uh, to finish that like lastmile of automation, um, humanoid
becomes a natural solutionbecause those working

(24:09):
environments is designed forhuman.
Uh opening doors to navigatingstairs to kind of like um like
you work, you have to work withhuman actually at the first
stage.
So humanoid in a human shapemakes a lot of sense um to that
way.
Amazon is now, I think, liketesting uh using humanoid as

(24:32):
kind of like the delivery guys.
Um they already have like a lotof robotics on their logistics
and um like sorting on thepackages and the warehouses, but
like you still have to facingvery complicated scenario when
delivering the box.
Um, for example, like the store,like there's um the stock there,

(24:54):
there's like some stairs theyhave to navigate.
Um so we have seen like somelike autonomous food delivery,
but like that's something Ithink like gonna be uh more um
scalable if we have likehumanoid that can really uh take
over a lot of delivery job.
So Morgan Stanley uh actuallythey projected um the humanoid

(25:18):
robotic market can be as largeas five trillion dollars by
2050.
Um and we could see like abillion unit um of humanoid
that's like a mass produced umby then.
And I think Elon has anotherprojection is like they're gonna
have millions of units uhstarted to be produced, uh

(25:41):
starting from next year.
Um I don't know, Elon's timelineis always like very aggressive,
but like he always got adirection right.
So uh give him the credit, um,that the EVs kind of like he
sees this as an early stage, andeven though everybody's
questioning, questioning afteryears, he finally delivered.
So we see like humanoid isprobably a similar development,

(26:06):
like what we've seen in asmartphone, in the smart cars,
now it's the smart robotics.
So at this stage, what we can doto invest in humanoid?
Um, we think there's a threecomponent.
Um, one is the integrity.
Um maybe everybody already knowssome of the companies, uh, but
interpreters is like companiesputting the things together.

(26:29):
Um, so um, like Tesla is one ofthem, like Newbie Tech, like uh
Rainbow Technology.
So those companies areintegrators.
And then there's umintelligence.
The company is providing um thefoundations for model training,
um, the oldest like newertechnologies and the motion,

(26:49):
like the really um from the themodels to motion control,
there's a connection, there's umchips available.
So that's the brain.
Um, but most importantly, Ithink is the body, um, because
um the body uh needs to bereally ready.
The supply chain needs to beready to make it scalable, so

(27:10):
the humanoid can be cheapenough, so everybody can use it,
and everybody's gonna use itlike a phone, like a car, then
people can keep upgrading withnew models.
And that that's the only way tomake it where it's scalable and
sustainable.
Um, so far, um, there's a wholeecosystem out there.
I think nobody reallyunderstands or not like owns a

(27:33):
lot of those positions, becausea lot of those companies are
really listed on theinternational market or emerging
market, even.
So there's accuration systemthat's really allowed each
component to move, uh thecontrol.
Um, there's a mechanical systemthat's really uh functioning uh
with also those verysophisticated move.

(27:55):
Um the hand is very expensive,probably the most expensive
part.
So there's a lot of companiesjust focused on the hand.
Um then there's like a sensingand perception companies
providing sensors, providing allthose like touching
technologies, uh, and there'scritical materials.
Um so that's really um, I thinklike an ecosystem out there.

(28:20):
Um we actually did a trip toChina recently to see, I think
the as people know, they havebeen doing this um um the first
world humanoid robotic game.
Um they also have like a theworld uh humanoid uh conference
there.
So it's a lot of thingshappening.

(28:40):
Like people can see like theracing competition, like boxing,
or just like sortingcompetitions like all over the
place.
So the humanoid uh is at anotherlevel.
Um then this um, as I said, itis very expensive at this stage.
Um the supply chain is notthere, it's not ready.

(29:02):
Uh you have to really keepputting a lot of capital to work
to make it ready, to make itscalable, so that those billion
units of humanoid can be reallyum produced in a very
cost-efficient way, so it'sreally consumer-friendly, work
manufacturing friendly.
Um, so we look at the cost oflike the humanoid manufacturing,

(29:27):
the body takes the most apart.
Um, those like actuators,ammonic reducers, bearings,
encoders, um, those are keycomponents uh as a hand is is
not a very key component.
So we we identified um like asuite of companies that's really
the supply chains um from thefor the for the startups where

(29:52):
the figure of electronicoptimist sanctuary, unit tree in
China, like UbiTech in China.
Um there's some So manycompanies now are providing
those components.
So they have like a vantage ineach of their own category.
So we think this ecosystemapproach is kind of like the

(30:14):
picks and shoulders.
As this is already, and a lot ofkey companies are still in
private.
So that's why a lot ofcompanies, they're gonna need to
invest in the supply chain tomake their humanoid um mass
producing.
So we think at this stage, it'sso already so investor can

(30:37):
position in very diversified wayto capture the ecosystem.
And as we see the ecosystem,it's really defined as the three
bucket.

(30:58):
As we find a methodology toidentify the relevance and their
supply chain strategicpartnership with the humanoid
companies.

(31:31):
And also like they have a lot ofintegrators that's coming.
And China and uh other emergingmarkets like Japan and Germany,
they're good at manufacturing.
Um, they have a good likeecosystem out there too,
providing the supply chain uhsolutions.
Uh their motion control is likefar beyond the imagination now.

(31:52):
Um so we actually bring one ofthe humanoids to ring the bell.
Um so uh for for our ETF uhcoit.
So uh as I mentioned, uh we haveuh now two ETFs here actually
like um uh offering you likeboth solutions.
Uh on the AI side, there's AGIX.

(32:15):
So it's a public and privatehybrid ETF capturing the
ecosystem for I think from GenAI to AGI.
Um but COI is really focused onthe physical AI, where the
humanoid um and embodiedintelligence is kind of like a
next generation of the investorsfocus.

(32:36):
Um those are really likestructural growth opportunities,
I think.
They're gonna take like a decadeto really play out.
Um and we believe um taking amore like ecosystem and basket
approach um could be a betterway to uh invest in the long run
compared to betting on likesingle companies or single

(32:58):
product.
So that's kind of like the umkind of like presentation today.
Um so AGX uh currently aboutlike um 97 million of we
launched on July last year, um,and is uh listed as listed ETF
on Nasdaq.

(33:19):
So anybody can can see it on ontheir uh on the website um of
crane shares.com slash AGIX.
Um so the for the performance isum it's been very strong, um,
driven both by the public andthe private.
Uh for the COID um is um uhanother ETF we launched on June

(33:39):
this year.
It's about like um uh 67million, I think now.
Um so it's about like 59 equallyweighted.
Uh um and yeah, since inception,we have been doing very
impressive return uh as well.
Um so the we uh I want to showlike a quick clip actually.
Like we did a trip to Chinarecently, and we have been

(34:03):
really filming some like greatfootage.
So just want to share witheverybody.
So you can see like uh thiscompany is already started like
a mass um production.
There's some like facial umrecognition and facial like
expression.
Their motion control is reallystrong.
Like you if you kick it, if youlike move it, it's very stable.

(34:27):
Uh it's like they're runningmarathon.
And this is another interestinglike company.
You can actually, using the appto do like autonomous, like just
like fully autonomous likeretail experience, your app
order, then humanoid is gonnagrab the stains or make coffee.
There's another humanoidactually beside it to making
coffee, uh, then it's gonna bedelivered to you um with uh

(34:51):
online payment.
So that's something like quiteunique.
Um and we within like the handcompany, there's like a
detector's hand, they're like umdoing the sensing technology
that you can feel the pressure,proximity, and um the kind of
like the um it's a touchingtechnology.

(35:11):
So you think like if humanoid isgonna do very sophisticated
tasks going forward, likegrabbing the cell phone or like
a water or something like that'sreally fragile.
You don't want like if you wanta humanoid to do laundry or um
clean the dishes, you don't wantlike it to really break the
dishes.
Um so you need those technology.

(35:32):
Uh we find a lot of those supplychain and um component companies
in China is really reallyinteresting.
Um they've been doing great.
I mean, like growing business.
So it's a it's an interestingecosystem out there.
Um so that's why like I thinklike if you think about like uh

(35:55):
Chat GPT is kind of like themoment where uh investor
realizes like AI is here.
Um we can see like some humanoidum actually like probably gonna
walk in by you and running onTimes Square, like someday then
you're gonna see, oh, this isthis is happening.
Um so um, yeah, so that's kindof like my presentation.

(36:18):
We're gonna see if there's anyquestions from investors.
Um happy to answer any questionsyou've had.

SPEAKER_01 (36:25):
The uh somebody asking about uh thing, great
presentation.
Uh is it possible to get theslides?

SPEAKER_00 (36:32):
Uh yeah, I can if you can uh send me an email um
at uh info at cranchairs.com.
Uh we can definitely send youthe presentations.
Um yeah, and for there's alsolike a lot of articles and uh uh
for the font presentation fontuh deck factories uh it's

(36:55):
available also on the uh craneshares.com uh slash Ajax or
Cranechars.com slash uh K O I D.
So for more information you cancheck there.

SPEAKER_01 (37:08):
So the big talking point now in the media is that
um the bottleneck for all the AIis uh is the actual energy,
right?
That's the the sheer amount ofenergy is needed for these data
centers is enormous.
Um what's the bottleneck forhumanoids?
Like what is there is there anequivalence?

SPEAKER_00 (37:25):
I think we don't have that same problem as the AI
data center right now, becauseif you think about like the AI
models, it's on cloud, right?
So it's really have the you haveto solve the problem that the
now the bigger the model is, thebigger the data center is.

(37:47):
Uh and you need like tons ofenergy.
And for the humanoid models, sofar as I know, most of the
models are running on the edge.
So those edge computing is verydifferent in terms of energy
consumption.
And as you can see, like thosehumanoids really run on battery.

(38:07):
So uh we don't, I think like wehave an oversupply of batteries
so far.
Uh who knows if there's a massproduction of humanoid, maybe
there's a battery shortage.
Um, but like so far the energypart is easy to fix.
The hurdle is really on thebrain.
Uh you can see the motioncontrol, the the hardware, like

(38:29):
the action that the humanoid cando, can move, can run, can see,
can flip, can do dances.
It's crazy.
Like, so the hardware is ready.
Um, so you just need AI.
Uh you need physical AI to begood enough um to operate the
humanoid.

(38:50):
So I think that's still lagging,but like that's starting.
Um, just as I said, um the thevision language action model or
the generalized robotic model isgonna change this.
It's good, then you're gonnaunlock the unlimited list of the
possibility for humanoid becausethe hardware is so ready for
this disrupt disruption.

SPEAKER_01 (39:11):
What do you think is more upside potential?
I mean, is that even a fairquestion to ask between uh the
AI side and the and the humanoidside?
I mean, the one needs the other,obviously.

SPEAKER_00 (39:20):
Yeah, so I think a lot of people already have a
position to AI names to theirstandard.
Um, so but like not a lot ofcompanies from the COID, the
humanoid ecosystem, are reallyexposed.
Um so that's kind of like the Ithink opportunity for investors
to really think about uh acompany if you already own AI,

(39:43):
what's like complement, right?
So most of the AI currently isreally US focused, it's like big
tech companies.
Uh, then there's a lot of uh, Ithink innovative companies also
in the US, um, that's not inNasdaq, um, and in the private
side.
But if you think about ahumanoid, uh, as I said, it's
very international.

(40:04):
Uh, it's very hardware-driven atthis stage because a lot of
hardware is so good, um, and themodel is not ready.
And so that makes a wholedifferent set of investment
opportunities in terms ofsectors, in terms of geographic,
in terms of correlation.
Um, so sometimes like you cansee like AI is up, humanoids

(40:28):
down.
Sometimes you can see, well,actually, humanoid outperformed
uh NASDAQ this year.
So that's kind of wheredifferent set opportunities
then, and I think, and in interms of um asset location
perspective, like the AI isprobably not core because those
like trillion dollar companiesare a lot of them, right?

(40:50):
So those are kind of like thethe drier for the economy,
enterprises, that's gonna bemainstream.
But humanoid is so under-owned,and it's not really discovered
by a loting master yet.
So at the early stage, this iskind of like it's like quantum
computing in a way, because Ithink like I think today, just

(41:14):
like HSBC Seth can use IBM'squantum computing.
Like so people don't realize,oh, it's actually come to our
real life.
So maybe like Am's gonna ask,we're gonna replace all the
human delivery guys withhumanoid.
So then people can ask, oh, thisis such a big opportunity.
Um so when that comes, I thinklike those humanoid companies

(41:37):
are gonna definitely get a lotof attention um in a way like AI
did today.

SPEAKER_01 (41:43):
It's not just the US that's in the race, it's also
China.
Uh so the question here, who doyou think is gonna win the
physical AI race, US or China?
You just came back from Asia.

SPEAKER_00 (41:51):
I just came back from Asia.
I was traveling.
Um, so I think is you are a goodquestion.
Because I don't think there's away to each one gonna win
themselves.
Um so China, they have very goodhardware capability,
manufacturing capability.
Um, they have a good supplychain ecosystem.

(42:14):
From each component is verysophisticated.
Um, there's tons of supply chainyou need to figure out from
material to the component to thethe accuration system together.
Um so that's uh advantage so farfor China.
Um they have a lot of companiesthat's together in the same

(42:35):
province that they can just likebeing very scalable.
So if there's um humiliarycompanies like Unitree or uh
anything like Ajiba, there'slike tons of those companies now
ready to they're they're prettyprofitable because they've been
able to source the components ina very cost-effective way.

(42:57):
So that making them uh kind oflike advantage in the the body.
But if you look at intelligence,so far, without a doubt, US is
leading in intelligence, in themodel developing a lot of
talents around um creating thebest models in the world.
China's catching up, but like uhI think the US is still leading

(43:19):
in the space, uh, especiallywith its ecosystem in the
Silicon Valley, with leading bylike media's platform, and all
the AI researchers is here.
Um that gives US a keyadvantage.
Uh, and also on thecommercialization part, I think
US is also leading because thinkabout the labor cost.

(43:40):
Here it's very expensive, right?
So hiring someone um at afactory, warehouses, service
sector, um, it's so expensive.
There's labor shortage, agingpopulation here.
Uh in China, they're starting tosee that in manufacturing, but
the labor cost is still verylow.
So the early timecommercialization makes a lot of

(44:02):
sense here.
If you think about nanny price,in China's so cheap for the
nanny, but here it's crazy.
So the price gap um is is gonnabe different.
So uh like I would think likewhen those humanoid companies,
they starting mass production ofhumanoid, figure out the supply
chain, then that's a problem forcommercialization.

(44:24):
Who's gonna win thecommercialization?
Um, that's probably very far.
Like when we're gonna see AI,we're we're still wondering like
what's the commercialization ofAI at this stage now, enterprise
coming out for the solution.
But for humanoid, I think that'sa long way to go.

SPEAKER_01 (44:43):
You hit on a little bit, sort of the the it's not
core, right?
Although you can argue SP NASDAQis now core AI, uh, because
that's what's driving thereturns.
But um, how should one thinkabout sort of position sizing
for these two mega trends?
Because if these are megatrends, then they probably
shouldn't be small.
But then again, they're also notas diversified.

SPEAKER_00 (45:05):
Yeah, I think like for the AI is kind of like um so
far as the driver for the forthe happy market for Nasdaq.
Uh if you think about like uhoverlap this year, uh SP was up
like, I don't know, like um lowdouble digit, and AI is probably
like 70% of that contribution.
Without AI, uh you're probablygonna be flat on SP 500.

(45:30):
So we have seen that before withsame as PC, internet, cloud,
SaaS.
Um, so on the next five to tenyears, you're gonna see AI gonna
play a similar role um to thoseAI winners.
Um, they're gonna dominate theperformance.
You think about like Dell in theearly times, gonna like thousand

(45:52):
times and Apple, like thoseopportunities kind of happening
in AI, uh, kind of transform theweight.
I think it can even be moreconcentrated uh to AI names
within the even ST 100 or SP500.
So AI concentration can justlead people to overweight AI

(46:13):
naturally.
Um, but you have to play in away that you have to identify
winner out of losers becauselosers are gonna drag the
performance.
Um, so that's one thing.
But humanoid, I think it isreally on the owned.
It's like nobody's has much ofthe position, especially talking
about international emergingmarketing.
Nobody's like US actually hasbeen doing so well over the last

(46:36):
10 years.
Why bother?
Why should I diversify?
Especially on something likehumanoid.
Um there's a US company, butlike, yeah, they're China's
leading, so uh um I don't haveaccess.
So that's why like I I cannottrade for like those single
single stock.
I don't think a lot a lot ofpeople can open an accounting in

(46:57):
Hong Kong and start buying thesecompanies.
Um so that's why like that's wefeel like we have to create a
solution to really bring all thecompanies relevant to um the
investors here um so people canget access.

SPEAKER_01 (47:12):
Folks, for those that are here for the C credits,
I will email you after thiswebinar, get your information
for the CFP board.
Appreciate those that are here.
Um for those who want to learnmore about uh the humanoid side,
right?
Aside from this webinar, and youknow, I'll have this as a
replay.
Any other good sources uh eitheron the clean share side or maybe
even outside the crane shareside that are worth paying

(47:32):
attention to?

SPEAKER_00 (47:33):
Yeah, actually we have seen a lot of research on
humanoid ashi across the majorbanks and uh brokers and
research providers.
Um Moving Stand is definitelythe biggest one.
I don't know if anyone hasaccess to their research, but
like they have done a lot ofgood research on the humanoid
side.
Um their analyst, um Adam Jonas,um at I was talking with him,

(47:57):
he's definitely a genius, uh,who has the vision for Tesla, I
think like 12 years ago orsomething.
It's quite quite like focused onthe like emerging technologies.
And humanoid is their latestfocus.
Actually, he now called himselfhumanoid analyst.
So um that's something Irealized that this is happening.

(48:22):
Um, same as for other uh firm, Ithink Goldman, UBS, they all
have like uh latest research onhumanoid.
Uh at Credentials, we do publishour research on humanoid as
well.
Uh as I said, uh, if you checkcreations.com slash K O I D, uh,
you can find a lot of researchthat is really um we asked what

(48:45):
we what we talked today, uh,what is driving this industry,
uh, which company going tobenefit.
Um, so for uh there's a lot ofarticles about that on our
website, so investor can checkthat out.

SPEAKER_01 (48:57):
That's a uh good place to wrap up this webinar.
Appreciate everybody that joinedhere.
Hopefully you found it veryinteresting.
And uh take a look at the funds.
As you can tell, the rightspace, the right theme, the
right issue.
Uh thank you, Derek.
Appreciate it.

SPEAKER_00 (49:08):
Thank you, Michael, and thank you, everyone.

SPEAKER_01 (49:10):
Cheers up.

SPEAKER_00 (49:11):
Cheers.
Advertise With Us

Popular Podcasts

My Favorite Murder with Karen Kilgariff and Georgia Hardstark

My Favorite Murder with Karen Kilgariff and Georgia Hardstark

My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.

24/7 News: The Latest

24/7 News: The Latest

The latest news in 4 minutes updated every hour, every day.

Dateline NBC

Dateline NBC

Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com

Music, radio and podcasts, all free. Listen online or download the iHeart App.

Connect

© 2025 iHeartMedia, Inc.