00:00:00
Why do you think I waited to make my
00:00:01
world podcast premiere for Allin? All
00:00:03
the ankle biters called and I said, "No,
00:00:05
I'm just going to wait till the besties
00:00:08
call."
00:00:08
>> CO2 is one of the most successful hedge
00:00:10
funds of the last two decades.
00:00:12
>> $55 billion under management.
00:00:14
>> This is their flagship hedge fund.
00:00:16
>> The reason we decided to kind of get
00:00:17
into this business is to find great
00:00:19
entrepreneurs and find great companies.
00:00:21
>> And they're looking to raise a whole
00:00:22
billion dollars more to invest in AI.
00:00:25
>> We're in an idea business. And when you
00:00:26
have a truly revolutionary idea, it can
00:00:28
get really big.
00:00:31
I hope to do something a little bit
00:00:32
different. Besties, you've been on for a
00:00:34
couple hours, so you can take a break
00:00:35
now for a few minutes. Sit back. We are
00:00:38
going to show you some slides and we're
00:00:42
going to walk you through really an
00:00:43
update on the unicorn economy. So, the
00:00:46
markets are back. We can see that the
00:00:48
unicorn economy on average since
00:00:50
September of 24 is up 70%. I think
00:00:53
that's intuitive to a lot of us. But
00:00:55
what's even more amazing is that the
00:00:57
public market has really made the same
00:01:00
move up. So if we look at the share of
00:01:03
the unicorn economy of the NASDAQ, which
00:01:05
had a significant move up since 2015,
00:01:09
it's really kind of plateaued over the
00:01:11
past few years. And I think it speaks to
00:01:13
the performance of public companies like
00:01:15
Palo Alto and others. So AI is
00:01:18
dominating fundraising.
00:01:20
What's kind of interesting in this slide
00:01:21
is you can see the share continues to
00:01:23
increase. So multiple years in a row now
00:01:26
that AI is increasing its wallet share
00:01:29
of fundraising.
00:01:33
But the composition of that funding has
00:01:35
changed. If you look at the unicorn
00:01:37
factory which really peaked in the Zer
00:01:40
era of 2021,
00:01:42
we've now really normalized at a much
00:01:44
lower level precoid.
00:01:47
So mathematically if you put both
00:01:49
together you can see that the funding
00:01:52
per unicorn has increased 5x since 2021.
00:01:58
So we have fewer unicorns that are each
00:02:01
raising more.
00:02:04
Now I'm going to spend a minute on this
00:02:05
slide because this slide is really about
00:02:07
the health of our ecosystem.
00:02:10
So the way to interpret this is if you
00:02:11
look at the green line which is the
00:02:14
preserp era unicorn cohort of which
00:02:17
there's about 73
00:02:19
you can see that 20 quarters after
00:02:22
becoming a unicorn
00:02:25
80% of them had either raised a new
00:02:27
round or exited which is I would say
00:02:30
pretty healthy.
00:02:33
Now if we look at the 2021 cohort which
00:02:35
is the red line two things stand out.
00:02:37
first that 20 quarters in, you can see
00:02:42
less than 20% less had either exited or
00:02:46
raised. But look at the number 479
00:02:49
versus 73 in the prior cohort. So now
00:02:53
here comes this new cohort, what we'll
00:02:54
call our 2024 cohort of AI companies.
00:02:57
And the key question is what will happen
00:03:00
in the future?
00:03:02
Which of these cohorts will they
00:03:03
resemble the most?
00:03:07
So we talked about how AI is
00:03:08
concentrating the funding base of
00:03:10
unicorns, but what we also see is the
00:03:12
top 10
00:03:14
is capturing a significant share of
00:03:17
funding. So it's not just AI companies,
00:03:21
it's a small number of AI companies,
00:03:24
which probably makes sense since we know
00:03:26
that Enthropic and OpenAI are raising
00:03:29
massive rounds.
00:03:33
And so what I like to think is we kind
00:03:35
of have a new index. If we really
00:03:36
thought about what the index of the
00:03:38
future is,
00:03:40
what for now I'll be able to call the
00:03:41
magnificent 8, but that number is going
00:03:43
to shrink as these companies go public.
00:03:46
The first thing that jumps to my mind
00:03:47
is, wow, what an incredible group of
00:03:49
companies.
00:03:52
And look at the diversity. SpaceX,
00:03:54
Stripe, Anthropic, Data Bricks,
00:03:56
Revolute, Bite Dance, Anderero. We have
00:03:59
internet, we have AI, we have fintech,
00:04:02
we have space tech.
00:04:05
I'd feel pretty comfortable owning this
00:04:07
index if I could for the next decade
00:04:09
plus.
00:04:11
And obviously the performance of this
00:04:13
index has been incredible. It represents
00:04:15
almost $4 trillion of value and has
00:04:18
really crushed the traditional kind of
00:04:21
mag 7. Almost every single one of these
00:04:24
names has outperformed that index.
00:04:30
We'll go to the next slide.
00:04:34
Now, another positive sign is that if we
00:04:36
look at the exits,
00:04:39
the exits are thawing.
00:04:42
So, one of the things that we've talked
00:04:43
a lot about with the besties over the
00:04:45
years is we know the unicorn economy is
00:04:48
great at consuming cash, but how much
00:04:50
cash is it really returning?
00:04:53
We need to have a balance between the
00:04:56
amount of cash consumed to the amount of
00:04:58
cash returned. That's how an ecosystem
00:05:00
stays in balance. And if we look, 2026
00:05:03
is actually on a pretty good trend.
00:05:06
Not quite where 2021 was, but pretty
00:05:09
good. And we still have half a year to
00:05:11
go.
00:05:15
But
00:05:17
that doesn't include
00:05:19
three companies that we know will be
00:05:21
coming public pretty shortly. SpaceX
00:05:24
obviously in the next few weeks and we
00:05:26
know Anthropic just today the headlines
00:05:29
hit that they've submitted
00:05:30
confidentially for their S1.
00:05:34
And if you add up the totality of just
00:05:36
those three companies, you can see that
00:05:39
it's basically going to be more than the
00:05:40
10 years kind of combined.
00:05:46
Which ultimately means if you remember
00:05:48
and you were there when I presented the
00:05:50
first all-in summit in 2024, we knew our
00:05:53
ecosystem was out of balance, we were
00:05:55
consuming way more cash than we were
00:05:57
returning, which is just a fundamental
00:05:59
imbalance. And you can see that now even
00:06:03
pre
00:06:05
liquidity events that I just mentioned,
00:06:06
our ecosystem is significantly more
00:06:09
balanced
00:06:11
and that will continue to improve.
00:06:16
Part of it is that the growth rates of
00:06:18
open AI and anthropic are unlike
00:06:19
anything that we've ever seen.
00:06:22
So if you look at this chart, just
00:06:24
remember this chart starts in January of
00:06:26
2025.
00:06:28
That was only a year and a half ago.
00:06:32
Just a few months in, these companies
00:06:34
passed Workday, a pretty incredible HR
00:06:36
company.
00:06:38
Then it was Service Now. It was Adobe by
00:06:42
the end of the year.
00:06:44
Salesforce on the way just in January.
00:06:48
Now even bigger than Google Cloud and
00:06:50
Azure.
00:06:53
So what can that look like in the
00:06:55
future? Well,
00:06:57
this is just based on kind of some
00:06:59
assumptions and some forecasts, but you
00:07:01
can see
00:07:03
that we estimate that only not only is
00:07:05
it bigger than Azure, but by the end of
00:07:08
the year could be bigger than AWS. and
00:07:11
potentially bigger than all of Microsoft
00:07:15
by 2028.
00:07:20
Now, these hyperscalers aren't sitting
00:07:22
still. They're seeing the disruption,
00:07:25
but actually they're doing more than
00:07:26
seeing it. They're actually funding it.
00:07:30
Because if you look at the chat GBT
00:07:31
moment that we know happened, look at
00:07:34
how much these companies, the largest in
00:07:36
the world, have invested in enabling and
00:07:40
creating this change. Truly
00:07:42
unprecedented.
00:07:48
So I know SpaceX, a lot of people are
00:07:50
going to talk about SpaceX. So I thought
00:07:52
I would share a little bit of how we as
00:07:53
investors think about SpaceX.
00:07:56
So that as you think about whether it's
00:07:58
a stock that you want to own or you just
00:08:00
want to seem smart at a cocktail party,
00:08:01
you can benefit from our knowledge.
00:08:05
The first thing that pops out when we
00:08:07
look and study SpaceX is that the number
00:08:10
one driver correlated to the valuation
00:08:12
of SpaceX is cadence of launches,
00:08:17
which intuitively makes sense. If your
00:08:19
business is the launch business, the
00:08:21
more you launch, the higher your value
00:08:22
should be. So I think we see that in the
00:08:25
data.
00:08:27
But there's another fundamentally
00:08:29
different ratio that I want to point you
00:08:31
to which is what if we took the
00:08:33
valuation, we divided it by the number
00:08:35
of launches. What would that look like?
00:08:38
Well, you can see it was kind of in a
00:08:39
fixed range for a while and then it
00:08:41
really started to move up.
00:08:45
And we believe that markets are
00:08:47
rational. And so we started thinking,
00:08:49
well, why is it that the market is
00:08:51
valuing SpaceX higher on a perlaunch
00:08:54
basis when it's launching more than when
00:08:56
it was just starting out?
00:09:02
And my fundamental view, and we'll kind
00:09:04
of call this our code framework, is that
00:09:06
the reason is
00:09:09
that the quality of SpaceX's business
00:09:11
model increases the more you launch.
00:09:14
So, in phase one, which we call
00:09:16
pre-constellation, you're just trying
00:09:18
your rockets. And we know rockets are
00:09:20
hard
00:09:22
and maybe you have a few government
00:09:24
customers and that's a one-time revenue
00:09:27
business and it's unpredictable.
00:09:30
Then you get into your initial ramp and
00:09:32
now you might have one constellation.
00:09:34
So, why is a constellation important?
00:09:36
Well, it's an end market and it's a
00:09:39
recurring revenue business. The more
00:09:41
satellites you put up, the more
00:09:42
subscribers you have, the more revenue,
00:09:45
etc.
00:09:48
Now you can move from ramp into scale.
00:09:51
Now you don't just have one
00:09:52
constellation, you have multiple
00:09:54
constellations.
00:09:56
And ultimately, we believe that a ride a
00:09:58
wide variety of companies and
00:10:00
governments and militaries will want to
00:10:03
own their own constellation so they can
00:10:05
control their own destiny.
00:10:10
So now you move into being a scale
00:10:12
business which ultimately becomes a
00:10:14
platform and we know how valuable these
00:10:17
platforms are in this technology age and
00:10:20
platform means not only do you have many
00:10:23
more customers in your core business but
00:10:24
you also have new businesses.
00:10:27
It could be space data centers. It could
00:10:29
be the optionality of the moon and Mars
00:10:32
and other space applications.
00:10:38
Now we know one defining feature of this
00:10:41
era has been how quickly these companies
00:10:43
are scaling. And if we just look at
00:10:44
whether it's the PC or the internet or
00:10:46
the mobile, Anthropic in particular is
00:10:49
scaling like no other company that we've
00:10:51
ever seen.
00:10:58
Now, this was kind of an interesting
00:10:59
analysis and this is one I'm I'm curious
00:11:01
to kind of discuss with the besties, but
00:11:04
we looked at essentially three buckets
00:11:06
of companies
00:11:08
and we said, okay, within each bucket,
00:11:10
what is the likelihood that you will
00:11:11
have a 10x,
00:11:13
which I would view as an investor, maybe
00:11:15
not a seed investor like JCAL, but for
00:11:17
us as growth investors, wow, a 10x is
00:11:19
pretty good. They're hard to find.
00:11:23
So the data showed us that if you're a
00:11:25
unicorn,
00:11:27
the odds of you one day becoming a
00:11:29
decacorn are about 8%.
00:11:35
If you're a decacorn, so that means
00:11:37
you're over 10 billion,
00:11:40
the odds of you becoming a hundred
00:11:42
billion dollar company, not much better.
00:11:44
8% to 13%.
00:11:48
But how interesting that if you're a
00:11:50
centacorn 100 billion or more
00:11:53
the odds and by the way we're putting in
00:11:55
public and private companies
00:11:57
you now have a 31% chance of having had
00:12:00
a 10x.
00:12:03
This kind of flies in my opinion
00:12:06
in different than maybe we would have
00:12:09
expected.
00:12:12
And if we look at how quickly these
00:12:14
companies are creating value, this is a
00:12:15
chart that I kind of added at the last
00:12:18
minute because the data is so fresh, but
00:12:19
you can see it typically takes multiple
00:12:22
years to go from 500 billion to a
00:12:25
trillion of market cap. Well, something
00:12:27
happened very recently in the public
00:12:29
market, which is that not only did we
00:12:31
have three companies do it in the same
00:12:32
year, but we had two companies do it in
00:12:35
a matter of weeks.
00:12:38
So we can talk about what conclusions
00:12:40
to take from that.
00:12:44
Now even as these companies were scaling
00:12:46
incredibly quickly from 500 billion to a
00:12:48
trillion,
00:12:50
we had other companies take a long time
00:12:52
to succeed.
00:12:54
And this is a company called Cerebras
00:12:56
that just went IPO. So I thought it'd be
00:12:57
a good candidate. I was very proud to be
00:13:00
a board member for a long time and led
00:13:02
the series B.
00:13:04
But if you look at the company's funding
00:13:06
history, you can see where I put the
00:13:08
little construction icon
00:13:11
that it took a long time and there was
00:13:12
some dark periods, multiple years of no
00:13:16
new capital, of hard grind to develop
00:13:18
their technology.
00:13:20
All of that time leading up to a massive
00:13:23
open AI contract
00:13:26
which then quintuple the value of the
00:13:27
company.
00:13:32
But we know Cerebrus has been
00:13:34
successful. And frankly, it's not just
00:13:35
Cerebras. Semis are on a generational
00:13:38
run. I was just talking about this with
00:13:40
my friend Brad Gersonner earlier. This
00:13:42
is just since 2024, the all-in summit.
00:13:45
You can see how much the semiconductor
00:13:48
industry has outperformed the index.
00:13:53
What will happen in the future? Well,
00:13:56
one takeaway from having listened to a
00:13:58
lot of speakers this morning is that
00:14:01
there seems to be wide agreement that
00:14:03
the more an AI system knows about your
00:14:06
business or you as a user, the more
00:14:09
useful it is.
00:14:11
You want to know when you go and book a
00:14:13
restaurant that it knows already your
00:14:15
preferences, whether it's what time you
00:14:16
like to eat or what food you like to
00:14:18
have, etc.
00:14:21
So we think ultimately that in this era
00:14:23
the amount of memory per user
00:14:27
could quintuple just based on the demand
00:14:30
that these AI systems are requiring to
00:14:32
provide their services. That helps
00:14:35
explain why we've seen some of these
00:14:37
moves in these memory companies.
00:14:42
And then I want to finish on a point
00:14:44
that I think has a lot of controversy
00:14:45
which is where's the revenue? If we
00:14:49
remember over the past 12 to 24 months
00:14:51
there's been a lot of discussion about
00:14:53
is there revenue is there ROI where is
00:14:55
this associated with
00:14:59
so we tried to look and see okay what is
00:15:01
the size ultimately of the AI ecosystem
00:15:06
we believe that it's about 140 billion
00:15:08
today it'll be about 300 billion this
00:15:11
year and it'll double in 2027 so where
00:15:13
is that revenue coming from
00:15:17
well if If we break it down, we can see
00:15:20
we kind of estimate three key pillars to
00:15:23
this industry. One, we know consumer
00:15:28
number of subs times an arpoo. That
00:15:31
gives you your consumer revenue.
00:15:34
One that I think a lot of people forget,
00:15:36
but it's ads.
00:15:38
We estimate currently that about a
00:15:39
quarter of ads served by Meta and Google
00:15:41
are AI enabled.
00:15:43
We think that penetration will
00:15:45
eventually go to 100%. that's 150
00:15:47
billion
00:15:50
and then obviously we all know about the
00:15:52
breakthroughs in enterprise
00:15:56
and what cloud code and codeex are doing
00:15:58
inside of those businesses. So if you
00:16:00
add all these together you get a good
00:16:02
sense of the size of this ecosystem.
00:16:09
So this will be kind of my second to
00:16:11
last slide. One thing that's different
00:16:14
to me about this era versus the prior
00:16:16
eras in which I was an investor
00:16:19
is that almost every sector of the
00:16:21
economy is being transformed at the
00:16:23
moment. So we know some of the obvious
00:16:25
ones software but look at telco.
00:16:29
I believe that within a few years
00:16:31
Starlink will power a device which will
00:16:33
actually enable you to make a phone call
00:16:35
anywhere in the world. And we think
00:16:37
that's a solved problem. But every time
00:16:40
we get a drop call, we get reminded that
00:16:42
there's a better technology out there.
00:16:44
So back to Nikesha's framework on profit
00:16:46
pools. I think the Starlink profit pool
00:16:49
is the telco global profit pool of
00:16:52
broadband and wireless.
00:16:56
We know compute is driving massive
00:16:58
changes in semis.
00:17:00
We had senators earlier on telling us
00:17:02
how data centers have changed the energy
00:17:04
equation in Pennsylvania.
00:17:07
Just think about the auto business. I'm
00:17:10
sure a lot of us followed what happened
00:17:11
to Ferrari last week, trying to
00:17:13
introduce a new technology of electric
00:17:15
and autonomous,
00:17:17
begging the question of what is the
00:17:19
future of that franchise in an
00:17:20
autonomous and electric world. And I
00:17:22
think the response to that car kind of
00:17:26
fed into this narrative.
00:17:30
And then obviously in consumer we know
00:17:32
GLPs are having a profound impact on
00:17:35
consumption of food, alcohol,
00:17:37
composition of diet and a huge focus
00:17:41
kind of on wellness.
00:17:44
So if we put all that together, what are
00:17:46
our takeaways? Well, my first takeaway
00:17:49
is that the new unicorn economy is
00:17:51
healthier
00:17:53
and we really have kind of AI to thank
00:17:56
for that.
00:17:58
The winners are compounding faster than
00:18:00
ever, which means the costs of not being
00:18:02
in a winner are higher than ever.
00:18:06
Disruption is impacting every part of
00:18:09
the global economy.
00:18:12
And by the way, we don't even have super
00:18:13
intelligence yet.
00:18:16
So if I think that it was about 2 years
00:18:18
since my last all-in summit, I started
00:18:22
thinking, well, gee, what could this
00:18:23
look like in two years? And we know it's
00:18:26
going to be a really interesting time.
00:18:27
And thankfully, we have a great group to
00:18:29
help us navigate what the next two years
00:18:31
will look like. We're going to give this
00:18:33
a title.
00:18:35
The power law rules our lives. The power
00:18:37
law rules our lives. All the great gains
00:18:39
are being consolidated into small
00:18:41
numbers of companies. Uh but we're still
00:18:43
seeing strength in those. How do you see
00:18:46
the private market
00:18:49
uh ecosystem, the game on the field
00:18:52
evolving because of the stay private
00:18:55
longer and these extraordinary outcomes?
00:18:58
Obviously, I operate in the earliest
00:19:00
stages. You have people who are doing
00:19:02
series A's like craft ventures. You have
00:19:06
uh yourselves dipping down into private,
00:19:08
but I was talking to Brad Gersonner, who
00:19:09
you um discussed earlier. He was like, I
00:19:12
I I have to figure out where to put my
00:19:14
time. You know, we have early stage and
00:19:16
and they and they do obviously public
00:19:18
like yourselves. So, and then add to
00:19:21
that, you have people like Andre and
00:19:23
Harowitz maybe going for the average in
00:19:25
a major way and indexing venture. What
00:19:28
What is the playing field going to look
00:19:30
like for people who are LPs, angel
00:19:32
investors, venture firms? What how does
00:19:35
this all sort out into a cohesive
00:19:37
strategy over the next decade or two?
00:19:39
because it's clearly the private markets
00:19:41
are operating much differently than the
00:19:43
playbook 20 years ago.
00:19:45
>> Yeah. So I think the the first breakdown
00:19:47
I would I would submit is on the
00:19:51
positive side of the ledger, the
00:19:53
outcomes are big, right? We're seeing
00:19:55
outcomes that we never thought possible
00:19:56
in private companies and I think that's
00:19:59
good just generally for our ecosystem.
00:20:01
So we have big outcomes.
00:20:03
It's really why I wanted to kind of show
00:20:05
that SpaceX slide. It was somewhat
00:20:07
counterintuitive to me on the launch
00:20:09
business. Why is it that the company
00:20:11
would be valued more as it launched
00:20:12
more? So, I think at least we have a
00:20:15
number of big outcomes and those
00:20:17
outcomes will be public within it seems
00:20:20
like a 12-month period. So, if I think
00:20:22
about, you know, the ZERP era where the
00:20:24
outcomes were smaller and companies were
00:20:27
not going public, I think at least in
00:20:29
this era, we have big outcomes and a
00:20:33
desire of these companies to go public,
00:20:35
right? I think both Anthropic and OpenAI
00:20:38
are both publicly saying that they want
00:20:40
to be public. So I would say that's
00:20:42
good. I'd say the biggest issue is the
00:20:47
it seems like we're talking about Kshape
00:20:49
and power law in every aspect of life.
00:20:52
>> Yeah.
00:20:53
>> And it seems like that's the case in
00:20:54
startups as well. So we've seen if you
00:20:56
looked at my centerorn slide, we've
00:20:59
really kind of been stuck at this number
00:21:00
for a little bit now.
00:21:03
So I think JCL I think the point that
00:21:04
you're asking is if we were to see no
00:21:07
new center coorns right in the next
00:21:09
decade we've basically not really seen
00:21:11
any new one in the past couple of years
00:21:14
I think that's going to be a warning
00:21:16
sign kind of for us what does this mean
00:21:19
for where capital allocators should be
00:21:22
thinking about putting their money
00:21:23
because what you're showing here a
00:21:25
rational person who's an LP we just say
00:21:29
wait for whoever gets to a 100red
00:21:31
billion
00:21:32
and yolo every dollar you can in there
00:21:35
because it's the most sure thing. It's
00:21:37
the least brittle. It's the least amount
00:21:39
of effort and it's the quickest return.
00:21:41
But as we know, supply demand equals
00:21:43
valuation. These valuations are
00:21:46
disconnecting from any valuation metric
00:21:48
we've ever had. Uh we had it explained
00:21:50
to us today by Bill Aman I think quite
00:21:53
accurately. You're making venture
00:21:55
investments in trillion dollar companies
00:21:57
and giving them 50 times revenue, 100
00:21:59
times revenue evaluations. So talk a
00:22:01
little bit about where people should
00:22:03
rationally as a limited partner, as a
00:22:05
private investor, a high net worth
00:22:07
individual, ultra- high net worth, where
00:22:08
should they be putting their money to
00:22:10
work and do you worry about this
00:22:12
everybody racing to be in three names?
00:22:16
>> Yeah, look, that obviously was the right
00:22:17
strategy for the past 5 years. The
00:22:19
question is about the next five years,
00:22:22
>> correct?
00:22:23
>> Right. So
00:22:25
the one push back I would have just on
00:22:28
the valuation argument is these are not
00:22:30
fake companies.
00:22:31
>> No, absolutely not.
00:22:32
>> So I think we have to I remember the
00:22:34
bubble in 2000. I also remember 2021,
00:22:37
right? These are companies generating
00:22:39
substantial revenue at scale
00:22:42
that are growing faster than anything
00:22:44
we've ever seen. So you know these
00:22:46
businesses are real and they're
00:22:48
performing. And I think it was widely
00:22:51
shown that Anthropic even had a
00:22:52
profitable month, I believe, is what was
00:22:55
reported. So, you know, they're they're
00:22:58
they're also kind of profitable. But
00:23:00
ultimately, and I think Traumath, you
00:23:02
agree with this, the public market is
00:23:06
the great test
00:23:07
>> equalizer.
00:23:08
>> Yes.
00:23:09
>> Yeah.
00:23:10
>> It will be the great antiseptic. it will
00:23:13
not care about my presentation
00:23:16
or you know um and so I love that. I
00:23:19
love that these companies are going to
00:23:21
have to face the scrutiny both SpaceX,
00:23:23
OpenAI and Anthropic of the market,
00:23:26
right? And ultimately, I'm a big
00:23:29
believer in the market. And so, I'm very
00:23:33
excited to see these companies go public
00:23:35
withstand the scrutiny of short sellers,
00:23:39
pontificators, debaters, politicians,
00:23:42
kind of etc.
00:23:43
>> Let me let me ask you two questions on
00:23:44
that. The first is very tactical which
00:23:46
is normally we would say that the
00:23:49
antiseptic or the disinfected happens on
00:23:52
t equals one day right
00:23:56
>> now the rules are changing there's going
00:23:58
to be a lot of passive buying so it's
00:24:00
going to move out that date because
00:24:02
you're going to have to wash through a
00:24:03
lot of supply demand so that's that
00:24:05
could maybe
00:24:06
>> 6 month plus one
00:24:07
>> 6 month plus one is when you'd say we
00:24:09
can really start to get a sense of what
00:24:11
these companies are okay so that's a
00:24:12
tactical question
00:24:14
>> the more strategic question, Thomas, is
00:24:15
do you think that there's something
00:24:17
structurally inefficient or wrong that's
00:24:19
allowing these compounders to accelerate
00:24:22
at scale? Like is that a market
00:24:23
efficiency problem or do you think
00:24:25
that's just a survivor bias and we
00:24:26
shouldn't look too much into that? How
00:24:28
do you look at that?
00:24:30
>> I don't want to read too much into it
00:24:31
because the N of those companies is so
00:24:33
small and look at anthropic, right?
00:24:36
Anthropic pre-cloud code was a
00:24:39
completely different company than
00:24:40
postcloud code, right? So one one event
00:24:42
completely dented the trajectory of
00:24:44
almost that entire industry. So it's
00:24:47
hard for me to know whether that's
00:24:48
truly, you know, whether these companies
00:24:51
were like the mule in the Foundation
00:24:53
series, something that could never be
00:24:55
predicted and just came out of nowhere
00:24:57
and it was just a one-time thing,
00:24:59
>> you know. Um, we'll see. I do think that
00:25:02
the narrative of oh these models are
00:25:04
commodities and these companies are
00:25:06
going to get I think that's been pretty
00:25:08
thoroughly disproven now right um
00:25:11
>> how do you as co you know your asset
00:25:13
base has swelled you've gone into you've
00:25:16
expanded strategy you're now doing you
00:25:18
know data centers you're doing many
00:25:20
things how do you keep it all organized
00:25:23
when maybe a slide like that would say
00:25:26
hold on a second maybe we should have
00:25:28
just plowed 10 billion dollars into
00:25:30
anthropic like how do you balance.
00:25:32
>> The reason I I make a deck like this and
00:25:34
in some ways I should thank you guys
00:25:36
because when we when I do something like
00:25:39
this for you guys and it is tremendous
00:25:42
amount of time from uh myself and our
00:25:44
team and um we really want to present
00:25:46
you with accurate information. So the
00:25:48
past two weeks has pretty much been a
00:25:50
full-time job doing this. But for me it
00:25:53
reanchors my conviction around what to
00:25:55
do. you know, I I can't go and listen to
00:25:58
a thousand people and then I get
00:26:00
distracted and I I don't know what I'm
00:26:01
thinking anymore. So going back to these
00:26:03
ground truths of numbers and valuation
00:26:06
bring me back to a point of okay
00:26:09
conviction, right? So for me whenever I
00:26:11
try and understand the world, I go back
00:26:13
to okay, what do I understand? I
00:26:15
understand models. I understand numbers.
00:26:17
Let me go back and kind of peel this
00:26:18
out. What I think hopefully the deck
00:26:20
will show is look there is substantial
00:26:24
reasons for why right if you look at the
00:26:27
the trillion dollar companies that
00:26:29
became trillion dollar companies in a
00:26:31
matter of weeks these are not fake
00:26:33
companies like these companies have been
00:26:35
around for decades right and they trade
00:26:37
at the lowest multiple of earnings of
00:26:39
the S&P 500 of almost any other company
00:26:43
so there is kind of something kind of
00:26:44
real
00:26:45
>> pentup energy there that just got
00:26:46
released
00:26:47
>> correct and now it's like Someone made a
00:26:49
point to me on on um on uh you'll like
00:26:52
this on memory, right? They said, "Well,
00:26:53
if I want to design a chip like OpenAI,
00:26:55
I can go to TSMC." And I know it's hard,
00:26:58
but at least I have TSMC to help me. If
00:27:00
I want to make memory, well, there is no
00:27:02
TSMC,
00:27:03
>> right?
00:27:04
>> So, what should the memory multiples be
00:27:06
versus ASIC chips as an example? The
00:27:08
wrath of Lena Khan can be se seen
00:27:10
clearly in this and saxs. I want to get
00:27:12
your input into how policy and elections
00:27:16
matter when it comes to outcomes.
00:27:18
>> As I saw, I don't know if it was this
00:27:20
chart, but the the chart where you show
00:27:22
the odds of each category reaching the
00:27:25
next level,
00:27:26
>> the
00:27:27
>> would you have predicted that out of
00:27:28
curiosity?
00:27:29
>> It's very counterintuitive. Where my
00:27:31
mind went was extrapolating one more,
00:27:34
which is what are the odds that the
00:27:35
trillion dollar market cap companies get
00:27:37
to 10?
00:27:38
>> Yeah. Yeah.
00:27:38
>> And the last one was 31%. I mean, it
00:27:41
seems to me what it be like 50%.
00:27:43
>> 100%. I don't know. It seems I'm
00:27:45
thinking is it going to be greater than
00:27:47
or less than 30? And it seems to me it's
00:27:49
greater than 30% are going to hit that.
00:27:51
>> It's it's it's probably the filtering
00:27:54
mechanism of what's the compounding
00:27:56
advantage or the durability of earnings
00:27:58
of that company. And for every step, you
00:28:01
have a filter that says, do you have a
00:28:03
compounding advantage? Do you increase
00:28:05
do you have a stronger durability of
00:28:07
earnings? And if so, you're going to
00:28:08
accelerate to the next phase. It's
00:28:10
almost like fundamental to business um
00:28:14
valuation analysis like Ben Graham style
00:28:16
analysis.
00:28:17
>> To get to that level that's called the
00:28:18
trillion dollar club, you have to have a
00:28:20
dominant business. And then the question
00:28:22
is just at what point do you hit
00:28:23
saturation and it seems like all of
00:28:26
these markets have ended up being so
00:28:27
much bigger than anyone would have
00:28:28
predicted.
00:28:29
>> Yeah. I mean just or or government
00:28:31
intervention because fundamentally if
00:28:33
you think about the break up of the Bell
00:28:36
system. I mean who knows where that
00:28:37
would have gone over time. They could
00:28:38
have had a monopoly on the internet.
00:28:39
They could have had a monopoly on
00:28:40
commerce. They could have had a monopoly
00:28:42
in e-commerce and on and on and on.
00:28:44
>> But as a trading strategy what you'd
00:28:46
like to do is have a bot that just
00:28:48
starts buying up shares of a company
00:28:50
once it hits 1 trillion. And actually if
00:28:53
you had done that I mean a lot of the I
00:28:55
mean I remember who was the first
00:28:56
company to hit trillion was it Apple?
00:28:57
>> I believe so. Yeah. And everyone was
00:28:59
like, "Oh my god, well," and you know,
00:29:01
now there's what, like five or
00:29:02
something.
00:29:03
>> Study that showed if you bought, I'm
00:29:05
sorry to interrupt, but that if you
00:29:06
bought the NASDAQ
00:29:08
um over a 10-year period, you get like a
00:29:10
3x multiple or something quite
00:29:12
significant. If you just rebalanced
00:29:13
every year on the top 10 companies in
00:29:15
the NASDAQ, so just buy the top 10
00:29:17
companies by market cap and you
00:29:18
outperform over a decade by like 3x.
00:29:20
>> Yeah, Thomas, why didn't you do that?
00:29:23
What is your maybe maybe just the last
00:29:24
question so we make sure we wrap up
00:29:26
about something that I think you're
00:29:28
uniquely positioned to tell us.
00:29:29
>> What happens when all this money gets
00:29:31
distributed back
00:29:32
>> like what do you think happens to your
00:29:34
competitive dynamics? What do you think
00:29:36
happens to entrepreneurial dynamics?
00:29:38
What happens in Silicon Valley when
00:29:41
three or four trillion dollars gets put
00:29:43
back to GPS then to LPs and then the
00:29:46
recycling happens?
00:29:48
Well, the first thing that comes to mind
00:29:49
is I remember when David was so bearish
00:29:51
California real estate. Um,
00:29:55
>> we'll see whether this influx of capital
00:29:57
>> time to sell.
00:29:59
>> San Francisco homes are selling.
00:30:01
>> Buy the 40%body. Anybody interested in a
00:30:04
40,000 foot mausoleum?
00:30:06
>> Yes.
00:30:07
>> Protesters not included.
00:30:09
>> The one thing I'll say on SpaceX and
00:30:10
look, I don't know whether 1.75 is the
00:30:12
right price for the IPO and you know,
00:30:15
frankly, I have no clue.
00:30:17
What I do know is that the global profit
00:30:20
pool of telco and service providers
00:30:22
across the world is anywhere between 2
00:30:25
to 400 billion depending on who you want
00:30:27
to address. Right? So you do have to
00:30:29
think about a company that just in a
00:30:30
core business which by the way wasn't
00:30:32
even in a couple years ago is addressing
00:30:35
a profit pool of multiple hundreds of
00:30:37
billions of dollars with a substantially
00:30:39
better product. Right? I think all of us
00:30:41
when you think about Starlink works all
00:30:44
the time, no radio towers, you know,
00:30:46
etc. So I go back to it and I think
00:30:52
it's hard to know Chimath because we've
00:30:53
never had anything like this before,
00:30:55
right? Um
00:30:57
>> the ultimate question would be if you
00:30:59
look a bit in the in the ride sharing
00:31:01
wars and in food delivery wars at some
00:31:03
point that excess capital was used to
00:31:05
have a price war,
00:31:07
>> right?
00:31:09
Could we see a price war between Open
00:31:11
Eye and Anthropic? As a question, right?
00:31:14
If these companies have so much capital,
00:31:16
is one of them ever going to pull a
00:31:18
price lever to try and compete with the
00:31:20
other?
00:31:20
>> Rationally, they should.
00:31:21
>> They should. So, we might see things
00:31:24
that we can't predict today, right?
00:31:26
Where companies might say, well, I have
00:31:27
my 200 billion of cash. Now, the issue
00:31:31
is they're spending so much on
00:31:32
infrastructure, right? So, it's it's not
00:31:34
obvious, but I do think we're going to
00:31:36
see some counterintuitive
00:31:39
um changes. You guys all discuss them on
00:31:42
the show every week, and hopefully I'll
00:31:44
come back in 2 years and and analyze
00:31:47
what went right and what went wrong.
00:31:48
>> Honestly, I think what should happen is
00:31:49
you should come back here every year and
00:31:51
we should get the benefit of
00:31:52
>> Yeah, let's lock it in. We'll lock it in
00:31:55
for the two weeks of work that we really
00:31:59
do appreciate the work and the effort.
00:32:01
Yeah. And it's it's really great to have
00:32:03
you bring this to the audience.
00:32:05
>> It just shows also the power of
00:32:06
sometimes slowing down. Yeah.
00:32:08
>> And to meditate on, you know, the the
00:32:11
actual state of reality and it was
00:32:12
incredibly grounded.
00:32:13
>> Incredibly rich coming from you.
00:32:15
>> Incredibly grounded. I think it's a
00:32:17
compliment or an insult. I'm
00:32:19
>> compliment.
00:32:21
Thank you. I'll just say thank you to
00:32:22
you Chimath for that incredible
00:32:24
compliment and for you Thomas for
00:32:26
coming. Thank you.
00:32:28
>> Thanks for doing great job.
