September 18, 2025 • 53 mins
Join me live as I sit down with Derek Yan, Director at KraneShares, to explore how the AGIX ETF is reshaping AI investing by giving everyday investors exposure to private AI unicorns like Anthropic and xAI—companies powering the foundation models behind the next wave of enterprise AI.
In this expert-led discussion, Derek explains how AGIX combines public equities with late-stage private companies, the proprietary AI scoring system that drives stock selection, and why model companies are the epicenter of the AI value chain. We also cover AGIX’s performance to date, its dynamic quarterly rebalancing, and how it navigates valuation concerns in a fast-moving market.
In this webinar:
- How AGIX gives investors rare access to private AI leaders like Anthropic and xAI
- Why foundational model companies could capture the lion’s share of AI value
- How the AI Score framework identifies true innovators vs. “AI pretenders”
- Why enterprise AI could outgrow consumer AI in scale and profitability
- How hybrid ETFs unlock private market access with daily liquidity
Lead-Lag Live Webinars give you exclusive access to deep-dive conversations with industry experts, portfolio managers, and strategists. Subscribe now to gain the insights you need to make smarter, more informed investment decisions.
#AIinvesting #AGIX #AIETF #privateAI #LeadLagLive
Start your adventure with TableTalk Friday: A D&D Podcast at the link below or wherever you get your podcasts! 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!
Youtube: https://youtube.com/playlist?list=PLgB6B-mAeWlPM9KzGJ2O4cU0-m5lO0lkr&si=W_-jLsiREjyAIgEs
Spotify: https://open.spotify.com/show/75YJ921WGQqUtwxRT71UQB?si=4R6kaAYOTtO2V
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 1 (00:11):
Thank you everyone for joining today's webinar.
I mean, it's a quiet summerweek, but there's a lot going on
.
Actually, nvidia just announcedtheir earnings.
There's a lot of earnings callrecently on AI companies,
especially the portfolio companyof our AGIX ETF.
So I mean, like just look attoday's performance of this,
(00:34):
it's doing very well.
So we are that's kind of like inthe mind of everybody like
we're now like in the third yearof AI, right.
So everybody's like either Iown it or the third year of AI,
right.
So everybody is like either Iown it or I have some way to get
access to this.
But we think this is like amulti-year opportunity and if
(00:56):
you can get it right, you'reprobably going to have a good
long-term result benefiting fromthe AI revolution.
I would say so.
With that, I think like a lotof people also like curious,
like what are we talking abouttoday?
It's like private AI unicorn inan ETF.
(01:16):
That is right.
So we're kind of likepioneering in this kind of like
the ETF industry, kind of like apioneering in this kind of like
the ETF industry.
Just to give a little background, craneshares we are 13,.
We launched our firm on 2013.
Now it's over like 12 yearshistory matched $11 billion
(01:40):
across our 30 plus ETFs, mostlylisted on New York Stock
Exchange and NASDAQ, so anyonecan buy it.
So that's kind of like ourbusiness setup.
We are kind of want to be veryinnovative, bring high
conviction ideas, hard to getaccess markets to our investors.
(02:03):
So, like in history, we havebeen focused on a lot of
categories, including a lot ofChina thematics, a lot of
climate related alternatives andalso opportunities globally.
So AI is one of our latestfocus.
We launched the AGIX ETF abouta year ago, on July 17, 2024.
(02:31):
That's kind of like the secondyear when everybody's already
like why did they use ChatGPT,getting used to, oh, bring AI as
part of their life and work.
But we don't think, like mostinvestors still is really
lagging in terms ofunderstanding the true power of
(02:54):
AI and the potential of AI, orlike how to really invest in AI.
So that's why we partner withsomeone I call Aetna Capital
Management.
They are a bunch of AI nativeinvestors, engineers, who are
like venture capitalists, beingin the industry forever
investing.
They're early investors incompanies like Anthropic XAI,
(03:19):
proplacity you probably heardsome of the names.
Some are really leadingfinancial, large-language model
companies, foundational modelcompanies like Anthrop XAI.
Some are like AI nativeapplications, like Proplasty,
who claim to want to buy Chrome.
If you watched news recently,those companies are really the
(03:42):
companies driving this round ofinnovation.
So when we look at the market,there's some AI products mutual
fund or ETS but none of them arebacked by venture capital
insights.
Because if you're not reallyinvesting into those
foundational model companies orAI native applications, you're
(04:05):
not situation aware.
You're not really.
You're not like techie enough,right, you are traditional, like
Wall Street guy or someone likeme.
Sit on New York.
You're not really situationaware.
So we have we need partners totalk to the foundational model
company who know the future ofnext generation of models?
(04:26):
Who knows how Agile AI gonnadeploy?
Which company gonna benefit?
So all those insights, wetranslate that into an index
called Selective AetnaArtificial General Intelligence
Index.
So does AJAX ETF gonna trackthat index Basically based on
(04:46):
that AI score that's going topick the we'll call it winners
out of the whole AI.
Every company is going to claimwe're an AI company, but who's
going to be the true AIcompanies, right?
So that's a process we did forthe public equity.
But when you look at investment,there's something missing.
How can you say you'reinvesting AI, this round of gen
(05:10):
AI, without investing into somecritical companies in the model
space or the AI nativeapplication space, because
they're private?
But even they're private, we'retalking about like
multi-billion or hundred billiondollar valuation companies.
If they IPO today, they couldbe part of the fund.
(05:32):
So the idea is like let's do ahybrid ETF that we can not only
invest in a public listedcompany but also the pre-IPO
company in the ETF directlybecause they know the companies.
So with their introduction weconnected with the company like
Anthropic, xai, Proplacity, andwe actually invested into
(05:54):
Anthropic and XAI currently inthe ETF.
So that makes the fund quiteunique to providing access to
the opportunity that's reallycritical for this round of Gen
AI.
So with that we launched Ajaxand then I mean we see this is
(06:17):
everybody's thinking, oh, thisis thematic, this is not a theme
, it's going to be all hypebubble then over.
Well, when we look at AI, well,there's a lot of themes, right,
like just short term, but AIfundamentally is more like a
structural growth compared to alot of other themes.
(06:40):
Because just look at like thisyear, right, like S&P was up
like 10, 11%.
Actually 60% of that returncontributed by AI names.
So without AI S&P is only like3% 4%.
We have seen that trend likeover the last two years.
(07:02):
But if you're going back wayover you can see each time
there's a technologybreakthrough or new technology
wave you're going to have abunch of companies that's
driving the whole return of theequity market or the broad
benchmarks like S&P 500 orNASDAQ 100.
(07:22):
You, starting from like themainframe, then to the PC, with
the PC you build out a wholeinternet industry.
Like then you have all the dataNow putting the cloud and SaaS
opportunities driving thereturns over the last decade.
Those structural growths arereally sustainable.
There will be ups and downs butif you happen to own those
(07:49):
subsectors of the equity marketyou are likely to outperform the
broad market equity return.
And we believe the AI juststarted.
Right Now we are in the thirdyear.
Think about the early adoption.
Just like currently most peopleknow chat, gpt.
(08:10):
That's only consumer facing,that's a consumer opportunity.
But if you look at like history, like what SaaS did and cloud
did to really change the waythat enterprise across American
and the whole world actuallyoperate, you realize that's a
trillion plus opportunity andyou probably realize that, like
(08:36):
that trillion dollar opportunityis much bigger compared to kind
of like the early internet andPC opportunity.
Kind of like the early internetand PC opportunity and also
internet and PC opportunity isbigger than the mainframe
because each technologybreakthrough is based on the
previous breakthrough, previouswave.
(08:56):
You are building the multiplelayers, right?
You know the first layer.
You have the hardwares and thesoftwares and internet built
upon that.
It's going to be bigger becauseit's leveraging that and we
have seen that for AI.
This is a super cycle that isbased on all the previous
technology breakthrough, so it'sgoing to bring a much larger
(09:18):
opportunity going forward formultiple years and we're now
only three years from the startof AI and that makes us believe
like we should own some likeearly opportunity, right, who's
really disrupting this?
Everybody saw NVIDIA but likewho's really using the
(09:41):
technology to really bring thefoundational model to power?
All the computing of the AIapplications, the cloud
computing and all the adoptionsis really the foundational model
companies.
So that's why we think havingaccess early is important.
(10:02):
So we are currently like ashareholder of XAI and Anthropic
.
Agx, on behalf of GreenShareTrust, sit on the cap table.
We added Anthropic on March 5th2025 through their series E
round of financing and XAI onJuly 18th through their tender
(10:23):
offer.
So those are really ouraddition, our investment in the
private this year that becomelike making those two companies
become part of the ETF holdings.
So looking at like the accessright, a lot of people think
like if I want access to thisprivate companies, I have to
(10:46):
like be a credit investor, Ihave to pay two and 20 to the
private and VC funds and lock upfor 10 years.
So that's very traditional waysto invest into the private
companies and there's someinterval funds right.
So it's like between thetraditional and the mutual fund
you lock up with like worry,uncertainty of like when you're
(11:09):
going to, what's the exposureyou're going to have with the
quarterly liquidity.
Then, like the hybrid ETF, whatwe do is really in between the
interval fund and traditionalETF.
So think about traditional ETFyou have daily liquidity but no
private.
But we have daily liquidity andwe sell in private.
(11:31):
So that I think the number onequestion that people ask how do
you rally those?
Those are private companies.
So just like and actually weare kind of like one of the
first firm doing that in the ETFworld, but there's many
companies that have been doingthat, like Fidelity, barron's
and Tiro.
(11:51):
They have been putting privatecompanies in the mutual fund
space.
So we're just bringing thatbest practice to the ETF
industry.
Etf, at the end of the day, is a48 fund, so that makes like
something.
Well, people should do that.
But like, why is not?
Well, there's, it's notnecessary, right?
(12:13):
What's the rationale to investinto the private vehicle when
you have good returns on thepublic, right, as a fact?
Right, the private investmentis really different by the year
you put your money.
They call it vintage.
So, like early times, like whenyou're putting in some like
(12:36):
private companies, but then theold IPO happening in 2021, once
you have like a lockup expired,you probably like companies like
a very poor performance.
So that's really hurting people.
Or they're now thinking aboutlike private companies.
They have to consider a lot ofthings when they go public.
Right, there's tariff.
(12:57):
There's, like the president'strade and other like policies,
uncertainties, macro.
Why bother?
Right, I just want to dobusiness.
Well, I just want to grow mybusiness Anthropic.
I think their annual recurringrevenue was like about $1
(13:18):
billion last year, july I thinkit's now approaching like $4 to
$5 billion.
So that type of growth is likeamazing, but like, if they're
like a public company, they'regoing to go through the April
Liberation Day and everythinglike that.
So that's kind of like the.
I think that's why a lot ofcompanies, especially even
(13:38):
companies like with variation ofhundreds billions, they're
choosing to stay in the privatelonger.
So that's why I like investorsto if you missed out those
opportunities, you probablymissed out a lot for the private
side, because once they becomeIPOs, it's really hard to get
access.
Actually, looking at like now,just recently, I think like,
(14:01):
look at the Circle, figma, right, corby those IPO markets have
been really coming back.
So, like, if you think about,like the potentially this cycle
of the private market especiallywhat we did is really
late-stage pre-IPO companieslike Antarctic and XAIAI and
(14:25):
potentially in the future, aswell as similar companies is
really capturing like this trend, right, the good companies tend
to be in a private marketlonger, so you need to really
get access.
So, having talked about the fairvalue, right, like, we are
daily liquidity.
So that's why we have a justsimilar to what the practice
(14:49):
that BlackRock sorry Vanguard,fidelity, ethereum and Barron's
doing.
They have a fair valuecommittee.
We do the same.
So we have a daily fair valuecommittee to give a fair value
to those private holdingscommittee.
To give a fair value to thoseprivate holdings, so that's
reflecting in the NAV and theNAV not as a value of the ETF,
(15:10):
so that investor is reallygetting a fair value of those
assets and that's providingdaily liquidity.
And our AP is going tofacilitate that.
The public equity is going tobe in-coin transaction and the
private is going to be cashtransaction creation and
redemption for the ETF.
So that's the mechanism of thisfund.
(15:32):
So you know, that makes thefunnel unique because you have
access to some companies.
Probably you don't get accessanywhere, right?
So, anthropic, for people whoare not familiar with that, they
have a consumer-facing appcalled Cloud.
It's a competitor with theChatGPT backed by OpenAI.
(15:56):
We think OpenAI is greatbecause OpenAI is like kind of
like the most popular consumerapp in the Gen AI space.
Everybody's like, if a consumer, you're going to use ChachiBT
because it's famous, everybodyknows it.
But actually if you'reenterprise, things are a little
bit different For, like, I think, current stage of adoption,
(16:20):
most enterprises are using AIfor, say, coding, right.
So the highest adoption rate ofcompany type is actually
technology companies.
So technology companies areusing AI to read the automatic
code and forming all kinds ofworkflow from the back end to
(16:41):
the front end.
That workflow automation can bereally integrated with the
autonomous decision maker of AIand agents.
So those agentic AI solutionsreally now widely adopt in the
technology space and that's kindof like the fastest growing
adoptions in the technologyspace and that's kind of like
the fastest growing adoptions inthe AI currently.
(17:03):
And who's capturing that?
Actually, we think Anthropichas the potential to really
become a leader in that.
If you happen to be a softwareengineer or you know someone
who's a software engineer whocode every day, they're probably
using uh like cursor, uh or umthe cloud code coding, um
(17:25):
cruiser is also using cloudtechnology.
So like eventually, um, thoselike apis, those traffic, those
revenues, uh, gonna be passedthrough to the anthropics api,
going to be pathed through tothe Anthropix API.
So that's kind of like therevenue making they're doing.
As I said, the revenue has beengrowing very fast.
(17:47):
It was just $1 billion annualrevenue last year.
Now it's almost 4, 5x.
So that's really a competitiveedge, I think, in the enterprise
world.
And so that's really acompetitive edge, I think, in
(18:07):
the enterprise world and webelieve enterprise AI is bigger
than consumer.
At least like next two to threeyears, the enterprise is going
to be the main kind of expanderon AI.
They're going to spend a lot ofmoney.
Think about, like thosetechnology companies, they're
(18:28):
like trillion dollar invaluation.
The tech industry is thehumongous.
So the amount of money going tobe spent to automate things is
going to be people are not goingto see in history, right.
So we call that kind of likethe AI revolution in the
(18:48):
enterprise world.
So I think this trend is waybigger than the SaaS and cloud.
So if that's going to happen,the company is going to provide
that solution to enterprise.
It's going to be critical.
We've seen some companies we own, some companies in the public
space, like ServiceNow, likePatenteer, who really providing
(19:11):
the either contacts layer, orsome like the automation, like
the automation layer in betweenbetween, like Anthropic and
OpenAI and XAI, those financialmodels and their final use cases
and application adoptions onthe enterprise side.
But we do think thosefoundational model companies is
(19:33):
going to take a large part ofthe revenue across the data
chain.
For the XAI we invest aboutlike 1.5 million, I think, in
July.
So XAI is really a company wethink is very special because of
Elon and Elon's ecosystem right.
(19:54):
So Elon on Tesla Robotech,ciatan is driving SpaceX owns
Neuralink.
It's going to do humanoidrobotics.
So all those things are goingto rely on AI and data.
So we feel like Elon'secosystem is quite unique.
They are unique in access.
(20:16):
They have access to the socialmedia data, they have access to
the chips, they have access tothe social media data, they have
access to the chips, they haveaccess to all those kind of like
travel data, driving data,humanoid data going forward.
So all this data becomes sovaluable for XAI going forward.
That's why, even though I thinkrecently, like XAI's GROK4
(20:37):
model become the mostintelligent model because they
have, like, a lot of chips and agood algorithm as well.
So that makes XAI a veryexciting challenge actually to
open an anthropic.
We're going to see Elon's goingto bring XAI to a unique
position within his ecosystem.
So with those two companies, webuilt a portfolio actually
(21:02):
around those foundational modelcompanies.
If you think about the stack ofthe AI ecosystem, you rely on
foundational model to reallytaking input, generate output or
building AI agents.
Then you need an infrastructureright.
The infrastructure is reallythe cloud, the data, all the
(21:25):
data preparation, the contacts,everything that's supporting the
model, the computing.
Then within that infrastructure, you need hardware companies
NVIDIA, of course, then otherenergy-related memories and
manufacturing.
Then you have AI applicationsthat are going to use AI that's
(21:49):
going to either consumer side orenterprise side.
So we have, as I said, we buildthis framework based on the
insights of the AI researchersand winter capitalists, who's
really at the forefront of theAI investment in the private
space.
So their connections and theirknowledge on the ecosystem is
(22:14):
way better than me and manyother like kind like buy-side
analysts.
So I would think like this isframework that's going to really
dynamically capturing likewho's going to be really benefit
from either disruption of theAI models or the disruption in
the business model for a lot ofcompanies, even the companies
(22:37):
that we all know, like Google's,like AMD's.
Those companies somehow aregoing to be disrupted.
So how do you evaluate thoseimpact?
And that's very dynamic everyquarter.
So we have this process toreally identify the company is
going to be in the portfolio,the weight.
The process is really likerule-based right.
(22:58):
So we have a universe of 2,000to 3,000 companies.
It's like the AI industrieslike that's really broad right.
So that then the critical partis the AI score.
You can have the relevance andreadiness of the company.
So, based on the insights of AIengineers and we're going to
weight the companies based onthe score and also the
(23:20):
free-for-market cap, the indexis rebalanced quarterly with
some capping on eithercategories or single companies.
With that process we can reallyidentify companies as really
relevant and ready for AI.
So for relevance is looking atlike things like their revenue
(23:40):
contribution from the AI product.
How's their current productprogress, because I think a lot
of companies really developingand fine tuning their product.
So you have to check theirprogress and their readiness.
So future revenue visibility ifthey can really make money from
AI or they just like spend alot of money and not able to
(24:02):
make money and how's theirbusiness model when AI is going
to disrupt it, when AI is goingto make their business model
weaker or stronger, right.
So we want to really build asystematic framework, holistic
framework actually for investorsto really capture this
(24:23):
long-term AI trend.
But not worry, oh, I'm pickingthe wrong name or I'm like just
betting all my money on onecompany.
So like a basic approach iswhat we think is the best for
long-term, very diversified andrelevant approach to really
capture this long-term trend.
And we are kind of like we'rediversified in terms like
(24:47):
allocation as well.
We diversified across all theAI stack from hardware to
infrastructure models toapplication.
You can see the subsector, subthemes within each category,
each stat, and this weight isalso based on the AI score.
So it's dynamic, right.
(25:08):
So I think this weighting hasbeen changing.
I remember, like back in whenwe launched the company a year
ago, a lot of weight is inhardware Then.
I think now a lot of likeapplication infrastructure
companies now becoming morecrucial.
Then they get a lot of thebenefit and the weight is more
(25:30):
towards that.
So, going forward, it's going tobe this trend is very similar,
right, I think, previoustechnology breakthrough.
You have the value migratingfrom hardware to infrastructure
applications.
You have to see that in AI aswell.
But I think there will be someback and forth based on the
latest dynamics and models andhow the AI research thinks.
(25:50):
So we have this quarterlyprocess really dynamic,
allocating between the threestats and the performance is
showing that the AGS has reallyoutperformed major growth or
technology index since welaunched the fund on July 17,
(26:11):
2024.
You can see that we are likeabout 30%.
I think this data is outdated.
It's as of like July 25th butas of today.
We're like 38%.
So that's something I thinkinvestors should compare
themselves on Bloomberg or Yahooor any, or Cranesharescom slash
(26:38):
Ajax.
You can see the performance ofthe fund compared to NASDAQ 100,
technology Sector Index, s&p500, or Semiconductor Index.
You can see there's some majorperformance out there because
the framework we talked aboutand because we think this is a
long-term trend to reallyidentify the company can truly
(26:59):
benefit from AI to capture thisstructural growth going forward.
And AI is volatile because Iwould think just like any
previous technology breakthrough, any technology if you think
about Nvidia, amazon, google andMicrosoft in the history, you
(27:20):
have some meaningful drawdown.
So that's kind of like I wantto remind everyone who's want to
invest in technology and AIspecifically, you should be
careful if you're not reallyrisk tolerant.
So we think like AI at leasthas like a similar volatility
(27:40):
compared to NASDAQ or technologysector Just seems like
semiconductor is more volatile.
It's more like cyclical, butjust generally technology has
more volatility.
But if you look at like AGXrisk adjusted return,
risk-adjusted return is notreally it's not actually taking
(28:01):
a lot more risk compared to liketechnology sector is less
volatile than semiconductors.
So that makes actually AI.
If you have a long-term horizonand you have a good risk
tolerance, it's a good long-terminvestment to really growth the
capital.
So with that I think I talked alot, but happy to take
(28:29):
questions.
I've seen some like Q&Aquestions.
So before I do that, I justwant to thank Michael and
everybody again to join thisdiscussion.
For AGIX, we are ATF.
We list the fund on NASDAQ andthe latest information is
(28:54):
available on cranesharescomslash AGIX and if anyone want to
know more about coinshares,please check our website.
We have several articles,webinars and, I think, white
papers on the AGIX fund.
(29:15):
It's kind of like our latestinnovation.
So there's a lot of questions,people asking like oh, how do
you do private investment in ETF?
How do you do that?
So we have a white paperactually on our website how to
really unlock private marketswith a hybrid ETF so you can
check that you're moreinterested in into the mechanism
(29:37):
of the fund.
So with that, yeah, I'm goingto go through some like a
disclosure thing, have to takesome questions or we can chat
more about the AI, okay.
So the first question is I'm atOZEAC.
I do not have access to thisETF?
How do I get it?
(29:57):
So if you are like a financialadvisor at a platform, you
definitely can request it and wecan contact the platform really
to onboard the ETF.
So so far, we just launched thephone on year.
So I think a lot of platformsrequire one year track records.
We just we just passed that, sowe are onboarding this fund to
(30:17):
many platforms that currently.
So, yeah, that's, it's great,like you're interesting.
So let's, let's, let's dofollow-ups and if you have more
questions, you can email us atinfo at craneinsurerscom.
We can do a lot more follow-updiscussions and we can talk to
(30:38):
your platform to really makesure, like this platform, this
ETF is available on yourplatform.
We have a question from Hans howwill the private company
positions work?
If the ETF has significantinflows, will the position size
be diluted?
The short answer is yes, sincewe are investing $1 million into
(31:00):
Anthropic and $1.5 million intoXAI.
On Dry, by the time we invested, I think it's about like 8%,
but with more flows coming in,the private is going to be
diluted.
But also, you have considered afair value, because I think we
(31:23):
invest in Anthropic at avaluation of 61.5 billion.
Now the company is talking toNestron, which is a much higher
valuation.
So along the time we have afair value committee that are
considering all the informationthat we're going to have a fair
value of that.
So the total value of theprivate is going to change as
(31:46):
well.
So all those factors are goingto impact the real-time weight
of private exposure within theETF.
But how do we really preventdilution going forward as we
grow the fund?
We have a very strong pipelinethat we work with our partner,
aetna Capital Management.
They're on the cap table ofmany AI companies.
(32:08):
They have a close relationshipwith many AI companies including
.
So first we're on the cap tableof Anthropic and XR already.
So if they do follow up roundswe can actually participate and
invest in more.
If we have more money availableand the pipelines is really
(32:30):
beyond that, we have beentalking to many other companies,
like in the AI space, so thatpipeline is very critical.
So whenever we have a deal flow, a deal opportunity, we can
take it.
We can actually just addingmore private company going
forward once the fund becomesbigger.
So generally we are talkinginvesting about like 10% by the
(32:57):
time we invest and we're goingto keep adding new names to the
portfolio as the fund growsbigger.
Okay, so, hans, you have anotherquestion.
When new ETF shares are created, assume new XAS shares cannot
be purchased.
Yes and no right.
So that's what I answer.
Right, with the inflow you'regoing to have new shares created
(33:21):
, so the public equity is goingto be in-coin creation.
So our AP, our broker, is goingto deliver the shares of those
public equities and the privateportion is going to be cash
creation.
So we're going to receive cash.
So we cannot invest in XCIdirectly at that moment.
(33:41):
But, as I said, we have astrong pipeline that's available
.
We're going to keep investingmore money to target a decent
weight of the private within thefund.
So that's kind of like the planfor the ETF going forward to
maintain the right exposure tothe private AI companies, as we
(34:03):
talked about today.
Speaker 2 (34:05):
Derek, I just wanted to ask a question as well.
I'm talking about where you'reallocating to private companies,
public companies, etc.
More like in terms of thesegment of the AI ecosystem,
with hardware, infrastructure,applications, etc.
Which of those three areas doyou think offers the most
untapped potential in terms ofAI?
Speaker 1 (34:25):
So currently I think a lot of people just don't have
any exposure, any directexposure to the model company.
When you think about modelcompanies OpenAI, anthropic, xai
I think they should be part ofthe infrastructure or they
should be a standalone.
Actually, they are at the coreof the ecosystem, right?
(34:47):
Actually they're at the core ofthe ecosystem, right?
Because think about, like, whatis driving this random
innovation?
Chatgpt is everybody knows that, so people don't have access.
So now, what's driving thisenterprise?
Ai is really Anthropix, api,all those like models really
(35:10):
powering the AI agents toautomate workflow for
enterprises.
People don't have access tothose models as well.
So those model companies Ithink is very unique is at the
epicenter of this round of AIinvestment, but that's only in
the private side.
You do have a little exposurelike Microsoft a little bit,
(35:33):
some like OpenAI.
You have the other like Google,have this like Gemini.
You have Facebook meta apps.
You have some models throughlike public holding, maxapp, but
it's not direct, it's notreally pure exposure to the
model companies.
Then I would think, applicationoffer more opportunities going
(35:58):
forward, just like manytechnology breakthroughs.
Previously you do buy NVIDIA.
Now then you're going to buildout all the database data
centers.
All those CapEx are going to bedeployed and they're going to
build out all the database datacenters.
All those CapEx are going to bedeployed and they're going to
generate returns.
Cash flows are high goingforward, but all that is based
(36:19):
on the application Right.
So previously we have seen thattrend similar in the Internet
the consumer apps, enterpriseapps, all those then those apps,
those software companies, nowbecoming like now, multi-billion
dollars from a few hundredmillions so then now trillion
(36:40):
dollar now.
So that trend is similar.
I would think going forward,where companies can really build
out applications is going tocapture a big part of the value.
Where companies can reallybuild out AI applications is
going to capture a big part ofthe value going forward.
So that's potentially, I think,a future investment we're going
to focus on is really AI-nativeapplications, maybe on the
(37:03):
private side and public side.
So that's kind of like theeight to four planning.
Yeah.
Speaker 2 (37:10):
And then that relates as well to the AI score, right,
when a company starts out beinghundreds of millions and then
goes to billions.
So when you're adding privatecompanies, you're looking at
that.
They have a clear path towardsmonetization, right.
And then how do you take thatand work it into the AI score?
Speaker 1 (37:27):
Yeah, so that's part of the score we're calling
revenue visibility.
So basically means they canreally make money going forward.
So if you think about it,there's a lot of AI companies
there's.
If you think about many otherAI ETFs.
There are like 70, 80 companiesthat any company putting AI in
(37:51):
their description or prospectusearnings call it's going to be
screened as an AI company by theindex provider.
Then it's included in the ETF.
So we don't like that approachbecause I think, going forward,
every company AT&T is going tosay we're going to use AI.
Che is going to say we're goingto use AI.
Che is going to say we're goingto use AI, but that doesn't
(38:12):
mean that's a good investment.
Well, potentially, but that'sreally relevant for this
structural growth.
You really have to look at howmuch impact from AI to this
business model and how muchmonetization they can make.
And even some pure AI companieslike Kuser, like I talked about
(38:34):
is the private company, but Ithink a lot of the revenue
eventually going to be passedthrough to Anthropic, because
they pay Anthropic for API toreally making that software
service available, to reallymaking that software service
available.
So themselves, if themselves isnot making a lot of money, but
(38:56):
actually they rely on anothercompany.
We don't think that's likehealthy.
So, similar to many otherpublic AI application companies,
I think like if the companyjust like wrapping AI with a
very thin margin, they're notreally creating a business model
.
That's going to be.
You have this flying wheel ofmonetization.
(39:19):
It's really hard, I think,especially at this stage when
everybody's questioningdemonetization so investors
become more picky on companies.
I think it was going to be.
There'll be a lot of questions,a lot of cautious going forward
If just company claim to befrom the earning call.
(39:40):
You probably noticed that a lotof companies that we don't own
but like people thought oh, it'sAI play actually tanked after
earning call because it's notmaking money right.
So we have seen that trend fora year now.
So that monetization, thebusiness model, is very
important and really you needsomeone to assess that.
(40:00):
You really need someone to besituation where you have to talk
to the model companies,application companies who are
they providing service to andwho's actually generating a?
The model companies,application companies who are
they providing service to andwho's actually generating a
healthy mode that can reallymake money?
Going forward is critical.
It's not just like do keywordsearching and finding AI in
(40:23):
their description.
Speaker 2 (40:26):
Right.
So then, sort of a follow-upquestion from that would be
since you've been so involved inthis space and you're
critiquing and going over everycompany that's not only in your
fund but that you've decided tonot have included in the fund,
what surprises have you seen?
Speaker 1 (40:43):
We have seen like I think it's a lot of companies
that people ask why are you notowning this?
So we have to explain why.
So it's a little bit riskybecause what if the company is
doing well, but after one year,I think compared to our holdings
(41:03):
, to the companies we don't hold, actually for the picture, have
we doing?
Okay, just like.
Even just like when theearnings call come out like a
lot of business models reallynot really having mode right.
So they're just like.
On cybersecurity, a lot ofpeople know one company I'm not
(41:25):
going to say the name, but likeit's like oh, that's the AI play
.
It was like no, it's not, it'snot, it's not.
It's like oh, that's the AIplay.
No, it's not, it's not.
It's like end-to-end.
It's like traditionalcybersecurity is not really
integrated into the AI workflow.
We're not including that.
The investor is like pissedbecause he owns the name.
So it's like but like now,after a year, he's like okay,
(41:45):
you're right.
So we have severalconversations like that because
now everyone is finding AI names.
Yeah, some Western may be doingbetter than Fun, better than
AGX, but I think it's reallyhard to pick several names and
claims like oh, I'm AI readybecause this industry is so
(42:09):
dynamic, the business industryis so dynamic, the business
model is always changing.
The capability of AI, of themodels of agenda AIs, is
evolving every day.
Even like we make mistakes,right, everybody makes mistakes
because nobody can see thefuture 100% right we all.
I think we live in a time thattechnology breaks through so
(42:31):
fast.
You have to keep running everyday to really keep up.
Oh, what is the genetic AI?
What is that?
So when we're talking about, Ithink, like especially in the
financial industry, when we aredoing a lot of discussion on AI,
some terminology is confusing,If not really to the industry,
(42:52):
you have no idea what I'mtalking about.
That's why you have to partnerwith venture investors who come
like they're one step ahead oftraditional public equity
investors, because they know theupcoming stuff, they know the
trend.
I think you have to acknowledgethat there's a limitation
public equity investors becausethey they know the upcoming
stuff, they know the trend.
So I think, like you have toacknowledge that like there's
(43:13):
limitation of information,knowledge and insights we have.
So keep learning, um, and behumble and just like um, keep,
keep.
Like you can make mistakes, butlike it's this long trend,
long-term structural goals forcommunity.
So we have time, uh, to reallyadjust and adapt so we can make
(43:34):
right decision going forward.
Speaker 2 (43:37):
Yeah, there's another question that's just come in on
the Q&A and that is how arevaluations impacting the
strategy today and the returnsgoing forward?
Speaker 1 (43:47):
So that's a really good question because it is
really like a mindset, right?
So if you think about like AI,if you think about like AI
investment or technologyinvestment generally in the past
, you have like a period ofusually like look at like a SMB
(44:07):
founder, right, simple, theone-on-one like buy low, sell
high.
When the valuation is high, yousell.
When the valuation is low, youbuy, and in the next five years,
10 years, you're going to bedoing better than others.
That's still valid.
I'm thinking in terms ofgeneral equity investment,
(44:27):
especially like S&P 500.
But you have some period thatthe valuation is really not that
relevant when you have I thinkthis is especially true when the
internet was really taking offand you have a period of high
(44:47):
forward PE, but the subsequentfive-year return and 10-year
returns is amazing.
So why is that?
Because the forward PE is forone-year PEs when you look at
next year and without takingconsideration of the long-term
growth rate, right?
(45:08):
So when you have a growth ratelike this at the early stage,
you have to look at the end game.
So in five years, in 10 years,well, how much company, how much
money the company gonna make?
What's the revenue, what's theearnings?
Um, you cannot just linearlycalculating the growth return
(45:29):
and forecasting the four PEratio and determine oh, that's
the valuation.
So I would think liketraditional valuations becoming
tricky for this type ofinvestment, especially when AI
is like internet, I believe.
So I would think likeconsidering valuation is
(45:50):
important on the short term,right.
So it could be when sentimentchange, when tariffs come,
there's some like short termdisruption corrections that
those high forward PE companiestends to be sold by short term
investors.
So it depends on the strategy.
(46:11):
You can do that, of course,right.
So you can do some trading riskmanagement.
You can do hedging if you wantto manage risk.
But it's challenging, it's hardto time the market and it's
short-term Usually.
Each correction, just like whathappened this April right, we
(46:31):
have a huge drawdown actuallyacross the market, especially
tech and AI, but the rebound isway bigger than the drawdown and
actually looking at the yeartoday, the Ajax hugely
outperformed S&P 500 and NASDAQ100.
So how do you really managethat risk?
I think like, just as I said,you either be patient, long-term
(46:55):
horizon and then just embracingit, or you have even more risk
where you maybe do some hedgingor like trading strategy to
re-manage down the risk.
But it's tricky, it's hard.
I think a lot of people justmissed it.
A lot of friends like oh, Isold like 50% of my portfolio
(47:18):
doing the April like tariff,like other things.
I never bought it back.
Then I missed all the rally.
So there's a risk upside aswell.
So just be careful.
I think there's like juststaying long term is a good
strategy, but also if you wantto manage risk, you got to be
really smart, I guess.
(47:38):
So, okay, so I think those arethe next questions.
Some of these companies aretrading at 10X price to sales.
Is that not a concern?
How do you decipher who's goingto win relative to extreme
variations?
Yeah, so if someone is readinglike 10xps, so that's something
(47:59):
I think like really like I'm notgoing to talk about each
company.
As I said, there's like forwardlooking is only like one or two
years, right, so you have tolook at like longer term.
Then a company is like you haveto really think about the how
the analyst usually, when youlook at like I think like
(48:21):
internet or like just likeeither industry analyst or macro
analyst or economist tend to bevery cautious.
I think like when sometechnology breakthrough happened
, just happened during theinternet era right.
So you have a lot of cautious,even like today.
I think, like people talkingabout, oh, inflation could be a
(48:44):
very like like a triggerrecession or stagflation,
trigger recession or stagflation.
Well, I think in the short-termperiod, those inflation can be
like the macro topic, can reallydrive the sentiment around, but
just like long-term, you see,technology is actually adding a
lot of productivity and usuallythe real productivity add is
(49:09):
more than the economists oranalysts forecasted and that, if
you have a longer horizon, thatreally has changed a lot of the
GDP or assumptions or therevenue growth assumptions for a
lot of companies, revenuegrowth assumptions for a lot of
(49:30):
companies.
So dynamics, I think, like inshorthand, those macro and
variations and all those thingsmatters.
But in the long term you have along horizon which I think I
recommend it for type ofinvestment like this.
You should have a very longhorizon to really navigating
through all the noises and allthose kind of like corrections
(49:52):
and disruptions along the way.
Thank you everyone for joining today's webinar.
I mean, it's a quiet summerweek, but there's a lot going on
.
Actually, nvidia just announcedtheir earnings.
There's a lot of earnings callrecently on AI companies,
especially the portfolio companyof our AGIX ETF.
So I mean, like just look attoday's performance of this,
(00:34):
it's doing very well.
So we are that's kind of like inthe mind of everybody like
we're now like in the third yearof AI, right.
So everybody's like either Iown it or the third year of AI,
right.
So everybody is like either Iown it or I have some way to get
access to this.
But we think this is like amulti-year opportunity and if
(00:56):
you can get it right, you'reprobably going to have a good
long-term result benefiting fromthe AI revolution.
I would say so.
With that, I think like a lotof people also like curious,
like what are we talking abouttoday?
It's like private AI unicorn inan ETF.
(01:16):
That is right.
So we're kind of likepioneering in this kind of like
the ETF industry, kind of like apioneering in this kind of like
the ETF industry.
Just to give a little background, craneshares we are 13,.
We launched our firm on 2013.
Now it's over like 12 yearshistory matched $11 billion
(01:40):
across our 30 plus ETFs, mostlylisted on New York Stock
Exchange and NASDAQ, so anyonecan buy it.
So that's kind of like ourbusiness setup.
We are kind of want to be veryinnovative, bring high
conviction ideas, hard to getaccess markets to our investors.
(02:03):
So, like in history, we havebeen focused on a lot of
categories, including a lot ofChina thematics, a lot of
climate related alternatives andalso opportunities globally.
So AI is one of our latestfocus.
We launched the AGIX ETF abouta year ago, on July 17, 2024.
(02:31):
That's kind of like the secondyear when everybody's already
like why did they use ChatGPT,getting used to, oh, bring AI as
part of their life and work.
But we don't think, like mostinvestors still is really
lagging in terms ofunderstanding the true power of
(02:54):
AI and the potential of AI, orlike how to really invest in AI.
So that's why we partner withsomeone I call Aetna Capital
Management.
They are a bunch of AI nativeinvestors, engineers, who are
like venture capitalists, beingin the industry forever
investing.
They're early investors incompanies like Anthropic XAI,
(03:19):
proplacity you probably heardsome of the names.
Some are really leadingfinancial, large-language model
companies, foundational modelcompanies like Anthrop XAI.
Some are like AI nativeapplications, like Proplasty,
who claim to want to buy Chrome.
If you watched news recently,those companies are really the
(03:42):
companies driving this round ofinnovation.
So when we look at the market,there's some AI products mutual
fund or ETS but none of them arebacked by venture capital
insights.
Because if you're not reallyinvesting into those
foundational model companies orAI native applications, you're
(04:05):
not situation aware.
You're not really.
You're not like techie enough,right, you are traditional, like
Wall Street guy or someone likeme.
Sit on New York.
You're not really situationaware.
So we have we need partners totalk to the foundational model
company who know the future ofnext generation of models?
(04:26):
Who knows how Agile AI gonnadeploy?
Which company gonna benefit?
So all those insights, wetranslate that into an index
called Selective AetnaArtificial General Intelligence
Index.
So does AJAX ETF gonna trackthat index Basically based on
(04:46):
that AI score that's going topick the we'll call it winners
out of the whole AI.
Every company is going to claimwe're an AI company, but who's
going to be the true AIcompanies, right?
So that's a process we did forthe public equity.
But when you look at investment,there's something missing.
How can you say you'reinvesting AI, this round of gen
(05:10):
AI, without investing into somecritical companies in the model
space or the AI nativeapplication space, because
they're private?
But even they're private, we'retalking about like
multi-billion or hundred billiondollar valuation companies.
If they IPO today, they couldbe part of the fund.
(05:32):
So the idea is like let's do ahybrid ETF that we can not only
invest in a public listedcompany but also the pre-IPO
company in the ETF directlybecause they know the companies.
So with their introduction weconnected with the company like
Anthropic, xai, Proplacity, andwe actually invested into
(05:54):
Anthropic and XAI currently inthe ETF.
So that makes the fund quiteunique to providing access to
the opportunity that's reallycritical for this round of Gen
AI.
So with that we launched Ajaxand then I mean we see this is
(06:17):
everybody's thinking, oh, thisis thematic, this is not a theme
, it's going to be all hypebubble then over.
Well, when we look at AI, well,there's a lot of themes, right,
like just short term, but AIfundamentally is more like a
structural growth compared to alot of other themes.
(06:40):
Because just look at like thisyear, right, like S&P was up
like 10, 11%.
Actually 60% of that returncontributed by AI names.
So without AI S&P is only like3% 4%.
We have seen that trend likeover the last two years.
(07:02):
But if you're going back wayover you can see each time
there's a technologybreakthrough or new technology
wave you're going to have abunch of companies that's
driving the whole return of theequity market or the broad
benchmarks like S&P 500 orNASDAQ 100.
(07:22):
You, starting from like themainframe, then to the PC, with
the PC you build out a wholeinternet industry.
Like then you have all the dataNow putting the cloud and SaaS
opportunities driving thereturns over the last decade.
Those structural growths arereally sustainable.
There will be ups and downs butif you happen to own those
(07:49):
subsectors of the equity marketyou are likely to outperform the
broad market equity return.
And we believe the AI juststarted.
Right Now we are in the thirdyear.
Think about the early adoption.
Just like currently most peopleknow chat, gpt.
(08:10):
That's only consumer facing,that's a consumer opportunity.
But if you look at like history, like what SaaS did and cloud
did to really change the waythat enterprise across American
and the whole world actuallyoperate, you realize that's a
trillion plus opportunity andyou probably realize that, like
(08:36):
that trillion dollar opportunityis much bigger compared to kind
of like the early internet andPC opportunity.
Kind of like the early internetand PC opportunity and also
internet and PC opportunity isbigger than the mainframe
because each technologybreakthrough is based on the
previous breakthrough, previouswave.
(08:56):
You are building the multiplelayers, right?
You know the first layer.
You have the hardwares and thesoftwares and internet built
upon that.
It's going to be bigger becauseit's leveraging that and we
have seen that for AI.
This is a super cycle that isbased on all the previous
technology breakthrough, so it'sgoing to bring a much larger
(09:18):
opportunity going forward formultiple years and we're now
only three years from the startof AI and that makes us believe
like we should own some likeearly opportunity, right, who's
really disrupting this?
Everybody saw NVIDIA but likewho's really using the
(09:41):
technology to really bring thefoundational model to power?
All the computing of the AIapplications, the cloud
computing and all the adoptionsis really the foundational model
companies.
So that's why we think havingaccess early is important.
(10:02):
So we are currently like ashareholder of XAI and Anthropic
.
Agx, on behalf of GreenShareTrust, sit on the cap table.
We added Anthropic on March 5th2025 through their series E
round of financing and XAI onJuly 18th through their tender
(10:23):
offer.
So those are really ouraddition, our investment in the
private this year that becomelike making those two companies
become part of the ETF holdings.
So looking at like the accessright, a lot of people think
like if I want access to thisprivate companies, I have to
(10:46):
like be a credit investor, Ihave to pay two and 20 to the
private and VC funds and lock upfor 10 years.
So that's very traditional waysto invest into the private
companies and there's someinterval funds right.
So it's like between thetraditional and the mutual fund
you lock up with like worry,uncertainty of like when you're
(11:09):
going to, what's the exposureyou're going to have with the
quarterly liquidity.
Then, like the hybrid ETF, whatwe do is really in between the
interval fund and traditionalETF.
So think about traditional ETFyou have daily liquidity but no
private.
But we have daily liquidity andwe sell in private.
(11:31):
So that I think the number onequestion that people ask how do
you rally those?
Those are private companies.
So just like and actually weare kind of like one of the
first firm doing that in the ETFworld, but there's many
companies that have been doingthat, like Fidelity, barron's
and Tiro.
(11:51):
They have been putting privatecompanies in the mutual fund
space.
So we're just bringing thatbest practice to the ETF
industry.
Etf, at the end of the day, is a48 fund, so that makes like
something.
Well, people should do that.
But like, why is not?
Well, there's, it's notnecessary, right?
(12:13):
What's the rationale to investinto the private vehicle when
you have good returns on thepublic, right, as a fact?
Right, the private investmentis really different by the year
you put your money.
They call it vintage.
So, like early times, like whenyou're putting in some like
(12:36):
private companies, but then theold IPO happening in 2021, once
you have like a lockup expired,you probably like companies like
a very poor performance.
So that's really hurting people.
Or they're now thinking aboutlike private companies.
They have to consider a lot ofthings when they go public.
Right, there's tariff.
(12:57):
There's, like the president'strade and other like policies,
uncertainties, macro.
Why bother?
Right, I just want to dobusiness.
Well, I just want to grow mybusiness Anthropic.
I think their annual recurringrevenue was like about $1
(13:18):
billion last year, july I thinkit's now approaching like $4 to
$5 billion.
So that type of growth is likeamazing, but like, if they're
like a public company, they'regoing to go through the April
Liberation Day and everythinglike that.
So that's kind of like the.
I think that's why a lot ofcompanies, especially even
(13:38):
companies like with variation ofhundreds billions, they're
choosing to stay in the privatelonger.
So that's why I like investorsto if you missed out those
opportunities, you probablymissed out a lot for the private
side, because once they becomeIPOs, it's really hard to get
access.
Actually, looking at like now,just recently, I think like,
(14:01):
look at the Circle, figma, right, corby those IPO markets have
been really coming back.
So, like, if you think about,like the potentially this cycle
of the private market especiallywhat we did is really
late-stage pre-IPO companieslike Antarctic and XAIAI and
(14:25):
potentially in the future, aswell as similar companies is
really capturing like this trend, right, the good companies tend
to be in a private marketlonger, so you need to really
get access.
So, having talked about the fairvalue, right, like, we are
daily liquidity.
So that's why we have a justsimilar to what the practice
(14:49):
that BlackRock sorry Vanguard,fidelity, ethereum and Barron's
doing.
They have a fair valuecommittee.
We do the same.
So we have a daily fair valuecommittee to give a fair value
to those private holdingscommittee.
To give a fair value to thoseprivate holdings, so that's
reflecting in the NAV and theNAV not as a value of the ETF,
(15:10):
so that investor is reallygetting a fair value of those
assets and that's providingdaily liquidity.
And our AP is going tofacilitate that.
The public equity is going tobe in-coin transaction and the
private is going to be cashtransaction creation and
redemption for the ETF.
So that's the mechanism of thisfund.
(15:32):
So you know, that makes thefunnel unique because you have
access to some companies.
Probably you don't get accessanywhere, right?
So, anthropic, for people whoare not familiar with that, they
have a consumer-facing appcalled Cloud.
It's a competitor with theChatGPT backed by OpenAI.
(15:56):
We think OpenAI is greatbecause OpenAI is like kind of
like the most popular consumerapp in the Gen AI space.
Everybody's like, if a consumer, you're going to use ChachiBT
because it's famous, everybodyknows it.
But actually if you'reenterprise, things are a little
bit different For, like, I think, current stage of adoption,
(16:20):
most enterprises are using AIfor, say, coding, right.
So the highest adoption rate ofcompany type is actually
technology companies.
So technology companies areusing AI to read the automatic
code and forming all kinds ofworkflow from the back end to
(16:41):
the front end.
That workflow automation can bereally integrated with the
autonomous decision maker of AIand agents.
So those agentic AI solutionsreally now widely adopt in the
technology space and that's kindof like the fastest growing
adoptions in the technologyspace and that's kind of like
the fastest growing adoptions inthe AI currently.
(17:03):
And who's capturing that?
Actually, we think Anthropichas the potential to really
become a leader in that.
If you happen to be a softwareengineer or you know someone
who's a software engineer whocode every day, they're probably
using uh like cursor, uh or umthe cloud code coding, um
(17:25):
cruiser is also using cloudtechnology.
So like eventually, um, thoselike apis, those traffic, those
revenues, uh, gonna be passedthrough to the anthropics api,
going to be pathed through tothe Anthropix API.
So that's kind of like therevenue making they're doing.
As I said, the revenue has beengrowing very fast.
(17:47):
It was just $1 billion annualrevenue last year.
Now it's almost 4, 5x.
So that's really a competitiveedge, I think, in the enterprise
world.
And so that's really acompetitive edge, I think, in
(18:07):
the enterprise world and webelieve enterprise AI is bigger
than consumer.
At least like next two to threeyears, the enterprise is going
to be the main kind of expanderon AI.
They're going to spend a lot ofmoney.
Think about, like thosetechnology companies, they're
(18:28):
like trillion dollar invaluation.
The tech industry is thehumongous.
So the amount of money going tobe spent to automate things is
going to be people are not goingto see in history, right.
So we call that kind of likethe AI revolution in the
(18:48):
enterprise world.
So I think this trend is waybigger than the SaaS and cloud.
So if that's going to happen,the company is going to provide
that solution to enterprise.
It's going to be critical.
We've seen some companies we own, some companies in the public
space, like ServiceNow, likePatenteer, who really providing
(19:11):
the either contacts layer, orsome like the automation, like
the automation layer in betweenbetween, like Anthropic and
OpenAI and XAI, those financialmodels and their final use cases
and application adoptions onthe enterprise side.
But we do think thosefoundational model companies is
(19:33):
going to take a large part ofthe revenue across the data
chain.
For the XAI we invest aboutlike 1.5 million, I think, in
July.
So XAI is really a company wethink is very special because of
Elon and Elon's ecosystem right.
(19:54):
So Elon on Tesla Robotech,ciatan is driving SpaceX owns
Neuralink.
It's going to do humanoidrobotics.
So all those things are goingto rely on AI and data.
So we feel like Elon'secosystem is quite unique.
They are unique in access.
(20:16):
They have access to the socialmedia data, they have access to
the chips, they have access tothe social media data, they have
access to the chips, they haveaccess to all those kind of like
travel data, driving data,humanoid data going forward.
So all this data becomes sovaluable for XAI going forward.
That's why, even though I thinkrecently, like XAI's GROK4
(20:37):
model become the mostintelligent model because they
have, like, a lot of chips and agood algorithm as well.
So that makes XAI a veryexciting challenge actually to
open an anthropic.
We're going to see Elon's goingto bring XAI to a unique
position within his ecosystem.
So with those two companies, webuilt a portfolio actually
(21:02):
around those foundational modelcompanies.
If you think about the stack ofthe AI ecosystem, you rely on
foundational model to reallytaking input, generate output or
building AI agents.
Then you need an infrastructureright.
The infrastructure is reallythe cloud, the data, all the
(21:25):
data preparation, the contacts,everything that's supporting the
model, the computing.
Then within that infrastructure, you need hardware companies
NVIDIA, of course, then otherenergy-related memories and
manufacturing.
Then you have AI applicationsthat are going to use AI that's
(21:49):
going to either consumer side orenterprise side.
So we have, as I said, we buildthis framework based on the
insights of the AI researchersand winter capitalists, who's
really at the forefront of theAI investment in the private
space.
So their connections and theirknowledge on the ecosystem is
(22:14):
way better than me and manyother like kind like buy-side
analysts.
So I would think like this isframework that's going to really
dynamically capturing likewho's going to be really benefit
from either disruption of theAI models or the disruption in
the business model for a lot ofcompanies, even the companies
(22:37):
that we all know, like Google's,like AMD's.
Those companies somehow aregoing to be disrupted.
So how do you evaluate thoseimpact?
And that's very dynamic everyquarter.
So we have this process toreally identify the company is
going to be in the portfolio,the weight.
The process is really likerule-based right.
(22:58):
So we have a universe of 2,000to 3,000 companies.
It's like the AI industrieslike that's really broad right.
So that then the critical partis the AI score.
You can have the relevance andreadiness of the company.
So, based on the insights of AIengineers and we're going to
weight the companies based onthe score and also the
(23:20):
free-for-market cap, the indexis rebalanced quarterly with
some capping on eithercategories or single companies.
With that process we can reallyidentify companies as really
relevant and ready for AI.
So for relevance is looking atlike things like their revenue
(23:40):
contribution from the AI product.
How's their current productprogress, because I think a lot
of companies really developingand fine tuning their product.
So you have to check theirprogress and their readiness.
So future revenue visibility ifthey can really make money from
AI or they just like spend alot of money and not able to
(24:02):
make money and how's theirbusiness model when AI is going
to disrupt it, when AI is goingto make their business model
weaker or stronger, right.
So we want to really build asystematic framework, holistic
framework actually for investorsto really capture this
(24:23):
long-term AI trend.
But not worry, oh, I'm pickingthe wrong name or I'm like just
betting all my money on onecompany.
So like a basic approach iswhat we think is the best for
long-term, very diversified andrelevant approach to really
capture this long-term trend.
And we are kind of like we'rediversified in terms like
(24:47):
allocation as well.
We diversified across all theAI stack from hardware to
infrastructure models toapplication.
You can see the subsector, subthemes within each category,
each stat, and this weight isalso based on the AI score.
So it's dynamic, right.
(25:08):
So I think this weighting hasbeen changing.
I remember, like back in whenwe launched the company a year
ago, a lot of weight is inhardware Then.
I think now a lot of likeapplication infrastructure
companies now becoming morecrucial.
Then they get a lot of thebenefit and the weight is more
(25:30):
towards that.
So, going forward, it's going tobe this trend is very similar,
right, I think, previoustechnology breakthrough.
You have the value migratingfrom hardware to infrastructure
applications.
You have to see that in AI aswell.
But I think there will be someback and forth based on the
latest dynamics and models andhow the AI research thinks.
(25:50):
So we have this quarterlyprocess really dynamic,
allocating between the threestats and the performance is
showing that the AGS has reallyoutperformed major growth or
technology index since welaunched the fund on July 17,
(26:11):
2024.
You can see that we are likeabout 30%.
I think this data is outdated.
It's as of like July 25th butas of today.
We're like 38%.
So that's something I thinkinvestors should compare
themselves on Bloomberg or Yahooor any, or Cranesharescom slash
(26:38):
Ajax.
You can see the performance ofthe fund compared to NASDAQ 100,
technology Sector Index, s&p500, or Semiconductor Index.
You can see there's some majorperformance out there because
the framework we talked aboutand because we think this is a
long-term trend to reallyidentify the company can truly
(26:59):
benefit from AI to capture thisstructural growth going forward.
And AI is volatile because Iwould think just like any
previous technology breakthrough, any technology if you think
about Nvidia, amazon, google andMicrosoft in the history, you
(27:20):
have some meaningful drawdown.
So that's kind of like I wantto remind everyone who's want to
invest in technology and AIspecifically, you should be
careful if you're not reallyrisk tolerant.
So we think like AI at leasthas like a similar volatility
(27:40):
compared to NASDAQ or technologysector Just seems like
semiconductor is more volatile.
It's more like cyclical, butjust generally technology has
more volatility.
But if you look at like AGXrisk adjusted return,
risk-adjusted return is notreally it's not actually taking
(28:01):
a lot more risk compared to liketechnology sector is less
volatile than semiconductors.
So that makes actually AI.
If you have a long-term horizonand you have a good risk
tolerance, it's a good long-terminvestment to really growth the
capital.
So with that I think I talked alot, but happy to take
(28:29):
questions.
I've seen some like Q&Aquestions.
So before I do that, I justwant to thank Michael and
everybody again to join thisdiscussion.
For AGIX, we are ATF.
We list the fund on NASDAQ andthe latest information is
(28:54):
available on cranesharescomslash AGIX and if anyone want to
know more about coinshares,please check our website.
We have several articles,webinars and, I think, white
papers on the AGIX fund.
(29:15):
It's kind of like our latestinnovation.
So there's a lot of questions,people asking like oh, how do
you do private investment in ETF?
How do you do that?
So we have a white paperactually on our website how to
really unlock private marketswith a hybrid ETF so you can
check that you're moreinterested in into the mechanism
(29:37):
of the fund.
So with that, yeah, I'm goingto go through some like a
disclosure thing, have to takesome questions or we can chat
more about the AI, okay.
So the first question is I'm atOZEAC.
I do not have access to thisETF?
How do I get it?
(29:57):
So if you are like a financialadvisor at a platform, you
definitely can request it and wecan contact the platform really
to onboard the ETF.
So so far, we just launched thephone on year.
So I think a lot of platformsrequire one year track records.
We just we just passed that, sowe are onboarding this fund to
(30:17):
many platforms that currently.
So, yeah, that's, it's great,like you're interesting.
So let's, let's, let's dofollow-ups and if you have more
questions, you can email us atinfo at craneinsurerscom.
We can do a lot more follow-updiscussions and we can talk to
(30:38):
your platform to really makesure, like this platform, this
ETF is available on yourplatform.
We have a question from Hans howwill the private company
positions work?
If the ETF has significantinflows, will the position size
be diluted?
The short answer is yes, sincewe are investing $1 million into
(31:00):
Anthropic and $1.5 million intoXAI.
On Dry, by the time we invested, I think it's about like 8%,
but with more flows coming in,the private is going to be
diluted.
But also, you have considered afair value, because I think we
(31:23):
invest in Anthropic at avaluation of 61.5 billion.
Now the company is talking toNestron, which is a much higher
valuation.
So along the time we have afair value committee that are
considering all the informationthat we're going to have a fair
value of that.
So the total value of theprivate is going to change as
(31:46):
well.
So all those factors are goingto impact the real-time weight
of private exposure within theETF.
But how do we really preventdilution going forward as we
grow the fund?
We have a very strong pipelinethat we work with our partner,
aetna Capital Management.
They're on the cap table ofmany AI companies.
(32:08):
They have a close relationshipwith many AI companies including
.
So first we're on the cap tableof Anthropic and XR already.
So if they do follow up roundswe can actually participate and
invest in more.
If we have more money availableand the pipelines is really
(32:30):
beyond that, we have beentalking to many other companies,
like in the AI space, so thatpipeline is very critical.
So whenever we have a deal flow, a deal opportunity, we can
take it.
We can actually just addingmore private company going
forward once the fund becomesbigger.
So generally we are talkinginvesting about like 10% by the
(32:57):
time we invest and we're goingto keep adding new names to the
portfolio as the fund growsbigger.
Okay, so, hans, you have anotherquestion.
When new ETF shares are created, assume new XAS shares cannot
be purchased.
Yes and no right.
So that's what I answer.
Right, with the inflow you'regoing to have new shares created
(33:21):
, so the public equity is goingto be in-coin creation.
So our AP, our broker, is goingto deliver the shares of those
public equities and the privateportion is going to be cash
creation.
So we're going to receive cash.
So we cannot invest in XCIdirectly at that moment.
(33:41):
But, as I said, we have astrong pipeline that's available
.
We're going to keep investingmore money to target a decent
weight of the private within thefund.
So that's kind of like the planfor the ETF going forward to
maintain the right exposure tothe private AI companies, as we
(34:03):
talked about today.
Speaker 2 (34:05):
Derek, I just wanted to ask a question as well.
I'm talking about where you'reallocating to private companies,
public companies, etc.
More like in terms of thesegment of the AI ecosystem,
with hardware, infrastructure,applications, etc.
Which of those three areas doyou think offers the most
untapped potential in terms ofAI?
Speaker 1 (34:25):
So currently I think a lot of people just don't have
any exposure, any directexposure to the model company.
When you think about modelcompanies OpenAI, anthropic, xai
I think they should be part ofthe infrastructure or they
should be a standalone.
Actually, they are at the coreof the ecosystem, right?
(34:47):
Actually they're at the core ofthe ecosystem, right?
Because think about, like, whatis driving this random
innovation?
Chatgpt is everybody knows that, so people don't have access.
So now, what's driving thisenterprise?
Ai is really Anthropix, api,all those like models really
(35:10):
powering the AI agents toautomate workflow for
enterprises.
People don't have access tothose models as well.
So those model companies Ithink is very unique is at the
epicenter of this round of AIinvestment, but that's only in
the private side.
You do have a little exposurelike Microsoft a little bit,
(35:33):
some like OpenAI.
You have the other like Google,have this like Gemini.
You have Facebook meta apps.
You have some models throughlike public holding, maxapp, but
it's not direct, it's notreally pure exposure to the
model companies.
Then I would think, applicationoffer more opportunities going
(35:58):
forward, just like manytechnology breakthroughs.
Previously you do buy NVIDIA.
Now then you're going to buildout all the database data
centers.
All those CapEx are going to bedeployed and they're going to
build out all the database datacenters.
All those CapEx are going to bedeployed and they're going to
generate returns.
Cash flows are high goingforward, but all that is based
(36:19):
on the application Right.
So previously we have seen thattrend similar in the Internet
the consumer apps, enterpriseapps, all those then those apps,
those software companies, nowbecoming like now, multi-billion
dollars from a few hundredmillions so then now trillion
(36:40):
dollar now.
So that trend is similar.
I would think going forward,where companies can really build
out applications is going tocapture a big part of the value.
Where companies can reallybuild out AI applications is
going to capture a big part ofthe value going forward.
So that's potentially, I think,a future investment we're going
to focus on is really AI-nativeapplications, maybe on the
(37:03):
private side and public side.
So that's kind of like theeight to four planning.
Yeah.
Speaker 2 (37:10):
And then that relates as well to the AI score, right,
when a company starts out beinghundreds of millions and then
goes to billions.
So when you're adding privatecompanies, you're looking at
that.
They have a clear path towardsmonetization, right.
And then how do you take thatand work it into the AI score?
Speaker 1 (37:27):
Yeah, so that's part of the score we're calling
revenue visibility.
So basically means they canreally make money going forward.
So if you think about it,there's a lot of AI companies
there's.
If you think about many otherAI ETFs.
There are like 70, 80 companiesthat any company putting AI in
(37:51):
their description or prospectusearnings call it's going to be
screened as an AI company by theindex provider.
Then it's included in the ETF.
So we don't like that approachbecause I think, going forward,
every company AT&T is going tosay we're going to use AI.
Che is going to say we're goingto use AI.
Che is going to say we're goingto use AI, but that doesn't
(38:12):
mean that's a good investment.
Well, potentially, but that'sreally relevant for this
structural growth.
You really have to look at howmuch impact from AI to this
business model and how muchmonetization they can make.
And even some pure AI companieslike Kuser, like I talked about
(38:34):
is the private company, but Ithink a lot of the revenue
eventually going to be passedthrough to Anthropic, because
they pay Anthropic for API toreally making that software
service available, to reallymaking that software service
available.
So themselves, if themselves isnot making a lot of money, but
(38:56):
actually they rely on anothercompany.
We don't think that's likehealthy.
So, similar to many otherpublic AI application companies,
I think like if the companyjust like wrapping AI with a
very thin margin, they're notreally creating a business model
.
That's going to be.
You have this flying wheel ofmonetization.
(39:19):
It's really hard, I think,especially at this stage when
everybody's questioningdemonetization so investors
become more picky on companies.
I think it was going to be.
There'll be a lot of questions,a lot of cautious going forward
If just company claim to befrom the earning call.
(39:40):
You probably noticed that a lotof companies that we don't own
but like people thought oh, it'sAI play actually tanked after
earning call because it's notmaking money right.
So we have seen that trend fora year now.
So that monetization, thebusiness model, is very
important and really you needsomeone to assess that.
(40:00):
You really need someone to besituation where you have to talk
to the model companies,application companies who are
they providing service to andwho's actually generating a?
The model companies,application companies who are
they providing service to andwho's actually generating a
healthy mode that can reallymake money?
Going forward is critical.
It's not just like do keywordsearching and finding AI in
(40:23):
their description.
Speaker 2 (40:26):
Right.
So then, sort of a follow-upquestion from that would be
since you've been so involved inthis space and you're
critiquing and going over everycompany that's not only in your
fund but that you've decided tonot have included in the fund,
what surprises have you seen?
Speaker 1 (40:43):
We have seen like I think it's a lot of companies
that people ask why are you notowning this?
So we have to explain why.
So it's a little bit riskybecause what if the company is
doing well, but after one year,I think compared to our holdings
(41:03):
, to the companies we don't hold, actually for the picture, have
we doing?
Okay, just like.
Even just like when theearnings call come out like a
lot of business models reallynot really having mode right.
So they're just like.
On cybersecurity, a lot ofpeople know one company I'm not
(41:25):
going to say the name, but likeit's like oh, that's the AI play
.
It was like no, it's not, it'snot, it's not.
It's like oh, that's the AIplay.
No, it's not, it's not.
It's like end-to-end.
It's like traditionalcybersecurity is not really
integrated into the AI workflow.
We're not including that.
The investor is like pissedbecause he owns the name.
So it's like but like now,after a year, he's like okay,
(41:45):
you're right.
So we have severalconversations like that because
now everyone is finding AI names.
Yeah, some Western may be doingbetter than Fun, better than
AGX, but I think it's reallyhard to pick several names and
claims like oh, I'm AI readybecause this industry is so
(42:09):
dynamic, the business industryis so dynamic, the business
model is always changing.
The capability of AI, of themodels of agenda AIs, is
evolving every day.
Even like we make mistakes,right, everybody makes mistakes
because nobody can see thefuture 100% right we all.
I think we live in a time thattechnology breaks through so
(42:31):
fast.
You have to keep running everyday to really keep up.
Oh, what is the genetic AI?
What is that?
So when we're talking about, Ithink, like especially in the
financial industry, when we aredoing a lot of discussion on AI,
some terminology is confusing,If not really to the industry,
(42:52):
you have no idea what I'mtalking about.
That's why you have to partnerwith venture investors who come
like they're one step ahead oftraditional public equity
investors, because they know theupcoming stuff, they know the
trend.
I think you have to acknowledgethat there's a limitation
public equity investors becausethey they know the upcoming
stuff, they know the trend.
So I think, like you have toacknowledge that like there's
(43:13):
limitation of information,knowledge and insights we have.
So keep learning, um, and behumble and just like um, keep,
keep.
Like you can make mistakes, butlike it's this long trend,
long-term structural goals forcommunity.
So we have time, uh, to reallyadjust and adapt so we can make
(43:34):
right decision going forward.
Speaker 2 (43:37):
Yeah, there's another question that's just come in on
the Q&A and that is how arevaluations impacting the
strategy today and the returnsgoing forward?
Speaker 1 (43:47):
So that's a really good question because it is
really like a mindset, right?
So if you think about like AI,if you think about like AI
investment or technologyinvestment generally in the past
, you have like a period ofusually like look at like a SMB
(44:07):
founder, right, simple, theone-on-one like buy low, sell
high.
When the valuation is high, yousell.
When the valuation is low, youbuy, and in the next five years,
10 years, you're going to bedoing better than others.
That's still valid.
I'm thinking in terms ofgeneral equity investment,
(44:27):
especially like S&P 500.
But you have some period thatthe valuation is really not that
relevant when you have I thinkthis is especially true when the
internet was really taking offand you have a period of high
(44:47):
forward PE, but the subsequentfive-year return and 10-year
returns is amazing.
So why is that?
Because the forward PE is forone-year PEs when you look at
next year and without takingconsideration of the long-term
growth rate, right?
(45:08):
So when you have a growth ratelike this at the early stage,
you have to look at the end game.
So in five years, in 10 years,well, how much company, how much
money the company gonna make?
What's the revenue, what's theearnings?
Um, you cannot just linearlycalculating the growth return
(45:29):
and forecasting the four PEratio and determine oh, that's
the valuation.
So I would think liketraditional valuations becoming
tricky for this type ofinvestment, especially when AI
is like internet, I believe.
So I would think likeconsidering valuation is
(45:50):
important on the short term,right.
So it could be when sentimentchange, when tariffs come,
there's some like short termdisruption corrections that
those high forward PE companiestends to be sold by short term
investors.
So it depends on the strategy.
(46:11):
You can do that, of course,right.
So you can do some trading riskmanagement.
You can do hedging if you wantto manage risk.
But it's challenging, it's hardto time the market and it's
short-term Usually.
Each correction, just like whathappened this April right, we
(46:31):
have a huge drawdown actuallyacross the market, especially
tech and AI, but the rebound isway bigger than the drawdown and
actually looking at the yeartoday, the Ajax hugely
outperformed S&P 500 and NASDAQ100.
So how do you really managethat risk?
I think like, just as I said,you either be patient, long-term
(46:55):
horizon and then just embracingit, or you have even more risk
where you maybe do some hedgingor like trading strategy to
re-manage down the risk.
But it's tricky, it's hard.
I think a lot of people justmissed it.
A lot of friends like oh, Isold like 50% of my portfolio
(47:18):
doing the April like tariff,like other things.
I never bought it back.
Then I missed all the rally.
So there's a risk upside aswell.
So just be careful.
I think there's like juststaying long term is a good
strategy, but also if you wantto manage risk, you got to be
really smart, I guess.
(47:38):
So, okay, so I think those arethe next questions.
Some of these companies aretrading at 10X price to sales.
Is that not a concern?
How do you decipher who's goingto win relative to extreme
variations?
Yeah, so if someone is readinglike 10xps, so that's something
(47:59):
I think like really like I'm notgoing to talk about each
company.
As I said, there's like forwardlooking is only like one or two
years, right, so you have tolook at like longer term.
Then a company is like you haveto really think about the how
the analyst usually, when youlook at like I think like
(48:21):
internet or like just likeeither industry analyst or macro
analyst or economist tend to bevery cautious.
I think like when sometechnology breakthrough happened
, just happened during theinternet era right.
So you have a lot of cautious,even like today.
I think, like people talkingabout, oh, inflation could be a
(48:44):
very like like a triggerrecession or stagflation,
trigger recession or stagflation.
Well, I think in the short-termperiod, those inflation can be
like the macro topic, can reallydrive the sentiment around, but
just like long-term, you see,technology is actually adding a
lot of productivity and usuallythe real productivity add is
(49:09):
more than the economists oranalysts forecasted and that, if
you have a longer horizon, thatreally has changed a lot of the
GDP or assumptions or therevenue growth assumptions for a
lot of companies, revenuegrowth assumptions for a lot of
(49:30):
companies.
So dynamics, I think, like inshorthand, those macro and
variations and all those thingsmatters.
But in the long term you have along horizon which I think I
recommend it for type ofinvestment like this.
You should have a very longhorizon to really navigating
through all the noises and allthose kind of like corrections
(49:52):
and disruptions along the way.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
CrimeLess: Hillbilly Heist
It’s 1996 in rural North Carolina, and an oddball crew makes history when they pull off America’s third largest cash heist. But it’s all downhill from there. Join host Johnny Knoxville as he unspools a wild and woolly tale about a group of regular ‘ol folks who risked it all for a chance at a better life. CrimeLess: Hillbilly Heist answers the question: what would you do with 17.3 million dollars? The answer includes diamond rings, mansions, velvet Elvis paintings, plus a run for the border, murder-for-hire-plots, and FBI busts.
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