Latest / trading places / TP17: Caplight's Javier Avalos 📊 | Secondary Boom 🔥| Why Banks Need Private Mkts 🏦 | $230B xAI 🚀

Transcript

TP17: Caplight's Javier Avalos 📊 | Secondary Boom 🔥| Why Banks Need Private Mkts 🏦 | $230B xAI 🚀 · trading places

0:00
I think if you are a bank or you are running a massive wealth
0:04
management platform and you do not have a solution for getting
0:09
your clients access to private markets, you are absolutely
0:13
behind in, in terms of like keep keeping up with competition.
0:17
Certainly seems like Oracle decided to hike a good bit of
0:21
their business to the fates of open AI.
0:25
I wonder how that trade will turn out.
0:27
I don't know if I'd want to jump in with my money at this
0:29
valuation on this bet, but I'm at least, you know, I just want
0:32
to think about the other side of it and and decide whether you
0:35
think there's something really special here that it's not
0:38
evident in the current company and the current metrics and what
0:40
they're doing. Welcome again to our no good,
1:08
horrible, terrible, very bad BC podcast, Trading Places.
1:13
But today it's not so horrible 'cause woo Hoo, we're on
1:17
vacation, baby. Well, we're still working here a
1:20
little bit, but we're off the grid at least for the next two
1:23
weeks for practical BC. Aman, you're in Jackson Hole,
1:27
right? You're getting ready for some
1:28
skiing. Yeah, flipped the Jackson Hole
1:31
yesterday with the kids. So we're going to we're it's
1:34
kind of a low snow year so far, but there's snow coming.
1:38
So plan is to wrap up the work this morning and hit the hit the
1:42
mountain this afternoon just in time to get the get the first
1:45
tracks. That's awesome.
1:47
Well, I'm here in Silicon Valley for the next couple of weeks,
1:49
but I'm going to be headed to Hawaii in January for a week, so
1:55
looking forward to that. Already wearing my Hawaii sort
1:58
of shirt here. Yeah, you look and you're ready
2:02
for the beach in the pool already.
2:03
Yeah, you might play now. I am.
2:06
All right. Well, let's jump into some news.
2:07
This week TikTok signed a deal after years long saga.
2:13
What's the deal I'm on? The saga does go back a couple
2:16
years. If you remember how this how we
2:18
got here, there was a congressional ban against TikTok
2:22
at the end of the Biden administration.
2:25
Biden said he wanted to sign it, and then he did.
2:27
And then Trump came in and was supposed to enforce it.
2:29
But in defiance of congressional law, Trump decided.
2:33
To. Pause the deal and it is paused
2:35
that it has been paused for this ban has been paused for nine
2:38
months as they try to find an accommodation which was really
2:41
to sell the US TikTok entity to AUS partner.
2:46
And it seems like in September they tipped it that they have a
2:48
deal in place and the deal that congressional ban was like he
2:51
can no more than 20% to be retained by Bike Dance.
2:54
So as far as I can tell, there may be a little bit of a
2:57
loophole here, but 50% of it is being spun off to a consortium
3:00
for US investors led by Oracle. And then Silver Lake, which
3:04
would definitely Americans, MGX is in there.
3:07
They're definitely not Americans.
3:09
That's an I. Was gonna say Abu Dhabi
3:12
sovereign fund right? Yeah, well, I guess they're not
3:15
Chinese. It, it used to be Mabagla and
3:17
then G42 and, you know, some combination of Abu Dhabi money,
3:21
but I guess it's not Chinese. And maybe that satisfies the
3:23
congressional requirement. And then there's the 30% like
3:26
that's going to be held by, we don't know the, the, the, the
3:30
official deal over the weekend says affiliates of certain
3:33
existing investors and Byte Dance.
3:35
So I guess it's, they are not Chinese that I guess there were,
3:38
they're satisfying the, the requirements of the ban.
3:41
But I suppose what it means is that TikTok will continue to be
3:44
available to US users, which is, which is I guess a good thing.
3:48
The intent of the ban was really to reduce Byte Dance and it's
3:51
direct, the direct involvement of Chinese ownership to 20% or
3:54
less. And I guess the way this is, but
3:56
it's been a months long negotiation.
3:58
You can see the way it's penciling out is looks like Byte
4:00
Dance would retain 19.9% of the company, but another 30%.
4:05
Is this strange? We don't know affiliates of
4:08
existing investors who may or may not be under the control of
4:11
the Chinese or the government. And then, you know, MGX from Abu
4:14
Dhabi is another significant owner.
4:16
But I guess, I guess the bottom line is the governance is going
4:18
to be a new board of directors. It's going to be majority
4:20
American byte dance will technically be below 20%.
4:25
So I guess that satisfies the letter of the law.
4:27
And, and I guess the good news is that TikTok US continues to
4:31
operate for all of its all of its American users, which will,
4:35
which will make my daughter's very happy.
4:36
They are right in that great square demographic and let the
4:41
check out their dance moves and the dance moves with their
4:43
friends online. So they'll be happy with it.
4:44
One of the owners is Oracle, and Oracle's in the news for some
4:48
other reason. Yeah, they've been having a
4:50
little bit of a a rough go of their stock price over the past
4:54
few months. I'm going to share with you this
4:56
infrastructure sell off has essentially began to separate
4:58
some of the AI winners and losers and Oracle has sort of
5:01
been on the on the losing end of it.
5:03
The short way to consider this, this trade is just as the at the
5:06
end of the year, there's better there's been some concern about
5:09
AI valuations, the run up in valuations, the capital
5:12
intensity and the AI trade was coming off.
5:14
I think what's really happening is it's it's separating some of
5:17
the more cash flow positive players with people like Oracle
5:20
who are not. Let me share my screen and I'm
5:22
going to take you through the sell off as we saw last week.
5:25
So the big announcement last week was actually Broadcom,
5:28
which announced a great quarter that beat the revenue estimates.
5:31
They raised their AI revenue outlook.
5:33
They've built a dominant position in in, in custom made
5:37
AI sales on chips, but and the margins are looking pretty good.
5:40
But their stock, their stock price started by 14% on right
5:44
after their they announced on December 16th and they're now
5:47
down about 20% from the peak and Nvidia's down.
5:49
You can see on the on the right hand side of the chart, Nvidia's
5:51
down 15%. Palantir announced speed
5:54
earnings, great revenue growth and yet they were down another,
5:57
you know, 5% from from their peak.
6:00
Oracle is down 40% from where they were and even though they
6:03
had a good quarter in terms of their their overall
6:05
profitability, the concern that they've that's begun to strike
6:09
is really in the in the cash flow area.
6:12
So let me show you the, the free cash flow numbers for Oracle on
6:16
this chart. Back in 2021 when they were just
6:19
a software company, they were generating 14 billion in free
6:22
cash flow. And then and as late as 2024
6:25
they were still at 12 billion in free cash flow.
6:27
And this year they actually went negative.
6:29
And when you see that kind of a number on their on their
6:32
statement of cash flows and you see the CapEx numbers going way
6:36
above within forecast at the start of the year, it now puts
6:39
pressure on their ability to finance the debt that they're
6:41
raising and to finance their ongoing CapEx commitments.
6:45
So now that now there's a lot of concern about whether people at
6:48
the bottom of this chart like Oracle or Broadcom who are very
6:52
involved in the AI trade, but have a lot of CapEx commitments
6:54
and cash flow commitments can keep up.
6:57
And then some of the companies at the top of this list like
6:58
Micron and AMD had big quarters, traded well and had a pop in the
7:02
stock as a result of it, because they don't have the same cash
7:05
flow problems that the that the Oracle do.
7:07
Oracle had their debt downgraded a couple of weeks ago by
7:11
Barclays down to triple B minus, I think.
7:13
So right on the edge of investment grade and dunk bond
7:16
status. And so now all of a sudden, the
7:18
clever markets are beginning to question whether these things
7:20
are, are viable. And that's the AI trade.
7:24
At an LK meeting at the start of the year, you and I took our,
7:27
our LK's role asking what we thought about the MAG 7 and the
7:29
AI companies, right? And we took them through this
7:31
matrix that said, think of the opportunities and the threats
7:35
and let's separate, let's start to separate the MAG 7, which we
7:38
think in 2025 will have a lower correlation in terms of how they
7:41
trade than in the last couple of years.
7:44
And we said Nvidia's a clear winner, fake opportunity in
7:48
terms of threat, like there's no downside than NVIDIA.
7:50
Stock, if not too much. Right.
7:53
It's like they're at the center of everything.
7:54
They're benefiting. There's no real downside.
7:56
Google definitely has a huge opportunity because the AI,
8:00
what's what is AI doing more than anything right now, it's
8:02
transforming LLM in search. Google's at the center of it.
8:05
They can trend, they can transform everything, but they
8:07
have a big position in their current search and that could be
8:09
under threat. And then we said Tesla,
8:11
Microsoft registering Tesla is obviously a car company, but one
8:14
where AI and autonomous driving could differentiate them from
8:18
all the other car companies. Microsoft is kind of a a non
8:21
player in AI two years ago, but got into open AI.
8:24
So that's now a big part of their business.
8:26
And then and then we said net is like right down the middle
8:28
opportunities and risks and Amazon and Apple is sort of, you
8:31
know, it is interesting, not a lot of opportunity, don't seem
8:34
to be at the front and center of that business, but not a lot of
8:37
threat either. We'll look at how the markets
8:39
traded this year. So this is the matrix we walk
8:41
people through at the very top of the comp the evolved of MAC7.
8:45
Google's been the best performer year to date.
8:47
What a surprise. A little bit of a shock.
8:50
They had a bit of a downturn in the middle of the year, but
8:52
they've they came back strong and Googled it.
8:55
NVIDIA, the ones with the biggest AI potential are at the
8:58
top of the list of year to date trades.
9:00
Tesla's right behind again. They had a very soft first half
9:03
of the year, but came they've come back in the last six months
9:05
and traded extremely well. And then Microsoft, you know,
9:08
Next up and then Meta and then Apple and then Amazon.
9:12
So almost perfectly maps the map that we put out.
9:14
The ones with the biggest AI opportunities have done the best
9:16
this year. The ones that have the least
9:18
interesting AI trades have done the worst this year.
9:21
And so the AI companies have been rewarded.
9:24
But to your point, I think what's more important is the
9:26
hyperscalers are safe. The ones that are doing well
9:28
with cash flow have done great and are benefiting.
9:31
The companies are that are AI adjacent like Oracle.
9:34
They have an AI angle, but there's cash flow commitments,
9:37
CapEx commitments, they're building the cloud, the
9:38
infrastructure. Those have really had a struggle
9:41
in the past, in the past couple of months.
9:42
Corey gives Corey is another one, but she had a great IPO,
9:45
came out very strong and then has traded off almost 50% from
9:47
their from their peak. So the those AI fortunes are
9:50
being separated I think based on things like cash flow dynamics.
9:58
Oh, mowed down one cookie too many code brown people.
10:02
When I highlighted an article that was written by a really
10:05
smart BC, Thomas Tungas, he was analysing hyperscaler
10:10
expenditures and, you know, no surprise where he's kind of
10:13
increasing their spend. But then he was highlighting
10:15
specifically the companies that you were talking about before
10:19
were being, you know, not just less cash flow, but also more
10:23
leverage. And I think in particular,
10:26
there's a couple of charts here. 1, you know, Oracle's being
10:29
judged very harshly in the credit default swap market.
10:32
But here specifically is a leverage problem issue that he's
10:36
highlighting. And if you look on the left,
10:38
Oracle and core weave, you know, significant leverage 5 to 1 for
10:42
Oracle, more than one to one for core weave.
10:44
And then the other four major hyperscalers, not quite so much
10:48
more reasonable, you know, single digit or low double digit
10:51
percentages. And the credit ratings for these
10:55
companies kind of fall in line with that.
10:57
So again, whether it's based on cash flow or debt leverage,
11:01
certainly Oracle and Corey were being treated quite differently
11:04
than the other 4 hyperscalers. That's a great chart.
11:07
And just to put some of the some numbers behind this, when you
11:09
see some of like Oracle who's operating on, they only have
11:13
like $20 billion in shareholders equity compared to like 300
11:16
billion plus for a Microsoft or a Google.
11:17
So it's a bit of a denominator problem and you raise a lot of
11:20
debt, you don't have a lot of equity, then it puts pressure on
11:23
the debt equity ratio. And that's something that the
11:26
analysts will look at and say, well, that seems like a, that's
11:28
a, that's a reason for the downgrade.
11:30
The other concern that has just come up with Oracle and there's
11:33
some accounting reasons why that may not be the best ratio.
11:36
It's a good, it's a good ratio to look at.
11:38
But what goes into shareholder earnings, sometimes you miss
11:41
like a, a big run up in a stock with a lot of equity capital
11:44
coming in will sometimes just mean that the, in the, in the
11:49
accounting view of the world is a lot more depth than equity.
11:51
But in terms of the reality of the business, you might have a
11:53
better position than what the balance sheet will tell you.
11:56
But Oracle had guided to about $8 billion in CapEx and Q2 and
12:00
Q3 and they end up spending 12. So once, once you guide to CapEx
12:04
at 8:00 and then spend 12, that's going to set off a wider,
12:06
wider red flags. On the.
12:08
Markets and so they got downgraded and they're still
12:10
holding on to their investment grade status, but they're
12:13
they're now perilously close to junk.
12:14
So their ability to, you know, finance their cash flows and and
12:18
cover those interest charges is going to be something that
12:20
they're going to have to resolve.
12:23
Certainly seems like Oracle decided to hike a good bit of
12:27
their business to the fates of open AI and I I wonder how that
12:32
trade will turn out. Well, our man in space, Elon
12:35
Musk, is suddenly talking a lot about data centers in space.
12:42
Everybody seems to be thinking that data centres in space is a
12:44
really cool idea. I guess access to a lot of solar
12:48
power, access to a lot of cooling.
12:50
I guess maybe challenges in getting stuff up there.
12:53
But probably for SpaceX, that's not as big a challenge as
12:57
everybody else. You know, seems like if you can
13:00
get stuff up there, then data center in space is a great idea.
13:04
And a lot of people talking about that all of a sudden.
13:07
Well, when he asked me who who was working center on it, I had
13:09
two choices. When was Elon Musk?
13:11
When you think about space, you think of Elon.
13:13
The other is Jeff Bezos, though, right?
13:14
He's he also he's the guy he's Dom.
13:16
It's didn't he send his girl corn to space?
13:18
So he's obviously a big fan of wow.
13:22
And so I think those are both. I think the idea is interesting.
13:26
I just think space is going to be difficult and I think it's
13:28
going to be, I guess you're right that now that now that the
13:31
Falcon Heavy capability is within, you know, is, is
13:35
available and whether they can dedicate some of that to, to
13:39
setting getting the equipment out in space.
13:40
I guess that now becomes a real possibility in the next, in the
13:43
next year or two that the computing power that has to go
13:46
into these data centers is it's just a mess like the
13:48
International Space Station. You think that's a big deal?
13:51
The data centers they're talking about star cloud is 100 times
13:54
more compute intensive that the International Space Station.
13:57
So I, I cannot imagine that this is a near, near term success.
14:01
But obviously, you know, it's, it's going to be a, you're right
14:04
about the, the abundant solar power a lot easier to cool.
14:09
So, you know, over the next couple years, I can see that
14:11
being a priority for certainly SpaceX and maybe for Bezos and
14:14
his operations. So to sum up, we've basically Co
14:19
opted Santa Claus and made him synonymous with Coca-Cola.
14:22
Have to say it's a breeze signing endorsement deals with
14:25
fictional characters. Lightspeed Venture Partners, one
14:30
of the largest Silicon Valley VC firms, raised $9 billion, mostly
14:36
to invest in AI. Lightspeed has backed 165 AI
14:39
companies. They were the biggest investors
14:42
in, I guess, a couple of big AI companies, Anthropic being one
14:46
of them, Elon Musk's XII, which we'll talk about a little bit
14:50
later today. I don't know.
14:52
It just seems like they've got plenty of money to invest in
14:55
other AI companies. They're part of a trend where
14:58
you see these mega funds that are raising for what half of all
15:01
venture capital came from the top 10 or 15 firms.
15:04
The last year we've had folks who General Catalyst and A16Z
15:08
are raising these 5 to 8 to $10 billion mega funds.
15:11
I guess the question is what happens to returns when you get
15:14
these mega funds dominating the landscape?
15:16
And certainly the last 15 years we've seen mega funds
15:18
underperform some of the smaller, more strategic funds.
15:23
And the, the lesson I think of that over that period is you can
15:26
have a fund that is too big and then it then ends up making very
15:30
large bets on a, on a handful of companies that get overvalued.
15:33
I don't know, maybe maybe IAI is different, but it feels like
15:36
we're, we're falling back into that, you know, mega fund trap.
15:38
And I don't think that's going to be great for returns going
15:41
forward. I think the VC world is
15:44
bifurcating into two very different styles of investing.
15:47
I think you have these super mega funds raising multi billion
15:50
dollar funds, betting huge amounts on very late stage AI
15:54
companies. And I don't know that returns
15:57
are a critical component of the evaluation of why to do that.
16:02
I mean, I know that sounds a little strange, but people want
16:05
to make sure they have a bet that's going to be on AI
16:09
companies almost regardless of the cost.
16:11
And so I think part of that is there's large amounts of money
16:15
required to finance these companies.
16:17
Part of it is they just want to make sure at whatever stage they
16:19
can they get into those companies.
16:21
The other side of the that is there's a whole bunch of very
16:24
small funds competing for seed round deals all over the place.
16:28
And I would argue that seed rounds have become pretty
16:31
expensive. There's a lot of arguments about
16:33
whether those bets are going to make money or not given, you
16:38
know, we're investing in seed stage companies now at, you
16:41
know, north of 20, thirty, $40 million sometimes as compared to
16:45
maybe, you know, 10 years ago when I think it was more common
16:48
that we invested seed companies at 5 to $10 million valuations.
16:53
Well, regardless, Lightspeed isn't the only VC fund raising
16:57
new capital. Dragon Air is another Silicon
16:59
Valley VC fund that's raising $4 billion for their new VC fund.
17:03
4.3 billion to be exact. Dragon Ear is definitely a more
17:08
later stage investor in a lot of these companies and this is
17:11
their 7th VC fund. I guess they're saying that's a
17:14
slow year. Dragon Ear has over 30 billion
17:17
in assets under management, including a $12 billion
17:19
Evergreen fund that can back both private and public
17:22
companies. Kind of tough when a $4 billion
17:25
fund is a slow year for VC fundraising, huh?
17:28
I. Guess that's a good problem to
17:30
have when you're spender. Smallest fund is only 4 billion
17:32
and and you got to work a little bit harder.
17:35
You got to hustle a little bit harder.
17:36
Dragon here. Get your get your game back.
17:39
Opening eyes in the news this week they're raising $10 billion
17:43
from Amazon. In this case, they're also
17:46
planning to use chips from Amazon.
17:48
Seems like they are now upping their evaluation and discussing
17:51
raising at 750 billion. I don't know if this is
17:54
motivated by Elon raising for SpaceX at 800 billion, which is
17:59
also been in the news recently, but now Sam has upped the price.
18:06
It seems this isn't official news yet, but it sounds like
18:10
sources have disclosed that opening eyes having discussions
18:13
to raise at 750 billion, potentially up to $100 billion
18:17
and ARR recently topping $19 billion.
18:21
So they are certainly growing revenue quickly, but raising up
18:26
to 100 billion at 750 billion, that sounds pretty stratospheric
18:31
there. Aman, What do you think?
18:33
Well, Open AI had just raised at a $500 billion evaluation a
18:37
couple of months ago. So they're already doing a 1.5 X
18:40
step up in just three months, like hot on the heels of our
18:43
evaluation corner where we talked about that deal.
18:46
And looks like you have a new investor in Amazon who was in
18:50
conversations to leave that round or to invest potentially
18:53
up to $10 billion in them, I guess, at that new evaluation.
18:56
And that's also significant because in addition to this
18:59
devaluation being higher, it's broadening the sources of
19:02
funding for Open AI, which they need, and also their
19:05
relationships. Open AI had started with
19:07
Microsoft back in 2019 as their original seed investor.
19:11
Then they moved to NVIDIA who provided money along with VCs
19:15
and EP us the chips they need in order to run their models.
19:19
And now they're going to Amazon and Amazon.
19:21
They just signed a, a cloud deal with Amazon back in, back in
19:26
October, which is a, a cloud partnership where AWS would
19:29
power open a is workloads, put money into the company and also
19:34
provide them Amazon's custom training chips.
19:37
So deepening that strategic relationship as well as just
19:40
being a, a funding source. So it seems like open AI is now,
19:45
you know, running, running to different customers, broadening
19:48
their supply chain for where they get stuff, broadening their
19:50
diversifying their funding sources.
19:53
I guess the question is just will Amazon be allowed to sell
19:55
open AI models here to cloud customers?
19:57
Because as of now, Microsoft, which owns 27% of open AI
20:00
equity, Microsoft has a exclusive right to those
20:05
relationships and those models. And I don't think Amazon will be
20:09
able to sell open a open AI models unless they get some kind
20:12
of a, a deal with Microsoft and open AI.
20:15
Maybe that's a litigation down the road because there are some,
20:18
there's some that has been a. Very difficult relationship
20:20
between Open AI and Microsoft and Open AI has been resisting
20:23
that exclusive and certainly over the past couple of years.
20:25
So I guess we'll see where that goes.
20:29
Well, it seems like the company has been able to grow revenue.
20:32
I guess it's up from 13 billion earlier in the year now to
20:37
almost 20 billion in revenue and they're projecting that should
20:39
rise to 30 billion next year. I guess $500 billion valuation
20:45
when they were only doing, you know, maybe 10-15 billion, now
20:48
$750 billion valuation, one that might be doing 20 to 30 billion
20:52
still pretty high multiple, but I guess, you know getting into
20:55
almost reasonable multiples relative to public markets still
21:01
seems quite optimistic. Yeah.
21:04
Well, if we get to the, if we get to the seven $50 billion
21:07
valuation you said and they're they're $20 billion ARR right.
21:11
So not quite, not quite revenue, but ARR that's still a 30 to 40,
21:16
maybe 45 times ARR. I don't know if you think that's
21:19
reasonable. I guess it's, I guess it's
21:20
reasonable for, you know, for certain AI.
21:22
Companies. AI Company.
21:25
Yeah, exactly. 40 X is still 40X.
21:27
Public company multiples, they're probably more like 7 to
21:30
8X maybe really 11:50 X for a really fast growing company.
21:33
So it's a lot more than that. Well, allow me to repeat a joke
21:39
my lawyers told me the other day.
21:41
It involved a large multinational corporation that's
21:44
been using my likeness without permission or compensation.
21:48
Isn't that hilarious? Another primary round company
21:53
that's been in the news pretty frequently is data bricks.
21:56
We were just talking about data bricks raising at $100 billion
21:58
valuations. Now Data Works is raising funds
22:01
at $134 billion valuation. The company is, you know, now
22:07
move their own revenue up from I guess 3 or 4 billion previously
22:11
to almost 5 billion. Do you think that 134 billion is
22:14
a reasonable valuation for this company?
22:17
You know, we covered them in our very first valuation corner back
22:20
about three months ago and at the time they were raising at
22:24
$100 billion. And we were, they were bullish
22:26
because we're comparing them to Snowflake.
22:29
Snowflake at the time is like a $60 billion market cap company,
22:32
but Data Bricks was similar, 4 billion in ARR, 4 million, four
22:35
billion plus in projected ARR for next year.
22:39
But Data Bricks grown faster, much more, much more quickly
22:42
than Snowflake. So I think they've been since
22:45
made progress on the growth. Snowflake is now at about a $78
22:48
billion market cap itself. So it has actually gone up.
22:52
And I think if you look at the comparable, the comparables
22:54
going up and Data Bricks executing, I don't know 134
22:58
billion, it does feel like a 20 times, you know, 20 times
23:03
multiple, which is on the high side of the range and may be
23:05
consistent with the highest growing SAS companies in the
23:08
public markets. But to date their revenue and
23:11
their performance seems to be at that, you know, top quartile,
23:14
top decile performance. So I think it's, I think it's
23:17
probably rich, but not not surprising given what you said
23:21
about them three months ago. Another primary round with the
23:24
company that we've covered before, Waymo is set to raise
23:28
money, I guess at $100 billion valuation.
23:31
Waymo has certainly been growing very quickly.
23:33
I'm on. We just covered them last week,
23:35
right? We did not, not even not even a
23:39
week ago and we said they were at 45 billion at that at that
23:42
time and but surely print higher the next time.
23:46
Very competitive, very crowded space, but one where they, they
23:49
seem to have, just look at the traction and the, the number of
23:52
rides they're getting now, they seem to be showing a lot of
23:55
success. So you know this, this is a lot
23:57
of excitement. I think that's still much more
23:58
of a feature looking bet. That's a bet on the feature, a
24:00
bet on the technology. It's hard to justify based on
24:03
current financials, but but as we, as we cover, the technology
24:06
is interesting. They're one of the 2-3, maybe 4
24:08
leaders in the space and, and seem to be, you know, seem to be
24:12
in a really good position over the next couple years.
24:14
Another company that's raising currently, which we've
24:16
previously reviewed is Lovable. This is a vibe coding platform.
24:22
They're raising 330 million at a $6.6 billion evaluation.
24:25
That's up significantly from a $1.8 billion evaluation in July.
24:31
I might as one of the fastest growing coding platform
24:34
companies out there, do you feel like 6.6 billion is not too
24:37
crazy? The evaluation definitely seems
24:39
on the high side. They're 200 million in ARR and
24:43
and growing as of now. So it is around 30 times.
24:47
It's a 30X multiple on their ARR, which I guess is on the
24:51
high side of, you know where the space is.
24:52
But we've definitely seen the kind of growth they're at 100%
24:55
year over year growth and seem to be front and center in, you
24:58
know, in a. More than 100% a month.
25:01
It just grew from 100 million in July to 200 million now, so
25:05
doubling revenue in only six months.
25:08
Fair enough. So that that puts them on a on a
25:11
track to being, you know, let's see let's see if it goes next
25:13
year. But you're right, it doubled in,
25:14
in under year and they have an excellent customer base and
25:17
their, their product seems to be catching, you know, catching on.
25:19
So I, I, I don't know, you know, I don't know if you can justify
25:23
just based on those multiples, but it seems like over the next
25:25
two or three years at that kind of growth and then then it's a
25:29
great bet if they can continue to do that.
25:31
Probably our craziest story right now is a company called
25:37
Unconventional AI. And they're unconventional for
25:41
more reasons than just one. They're raising a seed round a
25:46
$475 million at a $4.5 billion valuation.
25:53
And I guess perhaps because their founder CEO was former
25:58
head of AI at Databricks, but $475 million for a seed round at
26:03
a four and a half billion dollar valuation.
26:06
This is for a more energy efficient computer for AI.
26:08
But still, that just seems off the charts, right?
26:14
Half a billion dollars for a seed round.
26:17
Yeah. Is that, is that the biggest
26:18
seed round you've ever you've ever seen?
26:20
That's got to be, that's got to be up there.
26:22
That is an unconventional way to do it.
26:23
You know, it's not the normal way we tell entrepreneurs to
26:26
raise and run a business. But I guess when you're the
26:28
former AI head of data bricks, if you can, you can do things
26:31
unconventionally and you don't have to.
26:33
You don't have to follow the 500 Start-ups playbook.
26:37
Well, I think it seems like they're going to be raising even
26:40
more money, so this is just the beginning of their fundraising.
26:43
But they're already talking about raising a much larger
26:45
financing of a billion dollars. But the two-month old startup
26:50
can raise that 4 1/2 billion dollar valuation.
26:52
I guess as long as you have the former head of AI data breaks on
26:55
your team, money is no object. How did you know that?
27:02
I know a lot of things, Larry. I know that you've been on my
27:05
naughty list since 72. I know that your company has
27:08
been using my image to promote your product since 1931.
27:11
And I know I've never seen a dime.
27:13
And you know what that means, Larry.
27:14
I know you're a thief. Lastly, we have an interesting
27:21
story about Notion. Notion is a company that we
27:24
actually have an indirect investment through one of our
27:26
secondary purchases and they're doing a secondary sale at their
27:31
$11 billion valuation. Now this is a valuation that
27:33
they've had for many, many, many years.
27:37
But what's interesting to note here is that Notion has
27:40
gradually grown into that valuation.
27:42
They're doing about $600 million in revenue, which maybe isn't
27:47
super crazy for an $11 billion valuation.
27:49
When is the flat? Round.
27:50
Not really a flat round. I think the answer is when your
27:53
last round was in the peak of the bubble in 2021 and you were
27:57
able to raise at 100 times revenues.
28:00
Why? Because you could.
28:01
You were able to, and so you did.
28:04
And. Really knock him for raising you
28:05
know back in 2021 and since then, you know then the market
28:09
corrected dramatically and and corrected probably 40 fifty 60%
28:13
but notion good for them they've they've basically what 10 extra
28:16
revenue in the last four or five years and grown into that
28:19
valuation. And so it seems like it's a it's
28:22
a good outcome for the company and they were able to raise
28:24
enough money to have that one way to to build that that
28:26
happened to race again. And so the fact that they got
28:28
here, I think was a real, real prior to them.
28:31
So this charts kind of comparing notions growth in revenue to its
28:36
growth and valuation. Again, we'll we'll observe for
28:39
the record that lots of valuations were crazy in 2021,
28:42
but this was certainly one of the most crazy.
28:45
They went from about a 2 billion to an $11 billion valuation
28:49
between 20 and 21. That's a big jump here in that
28:51
green line. Revenue is taking a lot longer
28:54
to catch up, but finally seems to be justifying that valuation.
29:00
Here we can look at, you know, a couple of milestones for the
29:02
company. When they were raising their
29:04
Series A in 2019, they were valued at 800 million on
29:08
$3,000,000 in revenue. So that's I guess almost a 300X
29:14
both full. Later they raised 50 million at
29:18
2 billion when they were doing 13 million, so 150 times.
29:23
And then gradually things sort of started getting more
29:26
reasonable, but they were basically holding at that ten,
29:28
$11 billion valuation for four or five years.
29:33
Revenue went from 30 million to now 600 million.
29:36
So finally seems to be somewhat justified after only 5-5 years.
29:43
Right, right. The multiple went from like 100X
29:46
revenue to 2020 X revenue because they had to grow their
29:50
grow the revenues. But that was that was the bet
29:53
that Sequoia and CO2 made back in October 21.
29:56
It was the top of the market. They definitely got a premium
29:58
valuation to premium company. And eventually, you know, Notion
30:02
Notion grew into that. So that's probably not the way I
30:04
would want to do it as an entrepreneur.
30:05
But I guess it worked out great for the Notion team.
30:09
I'm sure they're breathing a little bit easier, you know, but
30:11
they've executed and they've done, they've done great.
30:12
It's a great company, so good for them.
30:14
Well, that's all of our news for this week.
30:16
Thanks for spending some time with us a month.
30:27
Today we have with us Javier Avalos, Co Founder and CEO of
30:31
Cap Light. Javier is not a very bad, but
30:35
very good founder and entrepreneur, has a lot of
30:37
insights about the secondary market, does a lot of private
30:40
company valuation analysis and I guess we'll learn more about
30:44
that. But Javier, welcome to the
30:45
program and thanks for joining us.
30:47
Excited to be here. Thanks for having me, Dave.
30:49
I was. Wondering if you could give us a
30:51
little bit of your background and history.
30:53
Yeah, I'll start at the investment banking piece.
30:55
I started my career working for global investment banks,
30:58
specifically Bank of America Merrill Lynch, Kalin and Company
31:01
and Deutsche Bank. You know, in the four years or
31:03
so that I was at Forge, we grew marketplace volume from less
31:08
than $50 million traded a year to over 2 1/2 billion by the
31:12
time I by the time I left the company.
31:15
I guess before we jump into cap light, I guess Forge was in the
31:18
news recently being acquired by Schwab and kind of on the heels
31:25
of equities that being acquired by Morgan Stanley.
31:28
What do you make of all this? And you know, do you expect
31:31
there to be more acquisitions happening?
31:33
I I don't know if you anybody's been ringing your phone or not.
31:36
1 ringy dinghy. 2 ringy dinghies.
31:41
But I thought you're a pretty popular guy lately.
31:45
Yeah, well, well, I appreciate that.
31:47
It's it's funny, like I look at your your guest list and sort of
31:50
it's a who's who of folks around secondary.
31:52
So I'm glad that I'm able to come in and contribute my my
31:56
cheap two cents where do. So you've spoken to a lot of my
32:01
my peers across the space and I guess reaction to some of these
32:05
recent transactions. I think these are going to end
32:07
up being great deals. Like I think they're going to be
32:10
long term accretive for for for both of the buyers and Morgan
32:14
Stanley and Charles Schwab, I think that in many ways both
32:19
were probably a long time coming.
32:20
It kind of seems like they all they just happened like out of
32:23
nowhere, But I was getting a lot of texts on the forge Charles
32:27
Schwab announcement from friends like, hey, hope this was a great
32:30
exit for you. Like, no, I know you were early.
32:33
It's like, well dude, the thing has been trading since the.
32:37
Company went public four years ago, so.
32:39
Exactly. Yeah, but.
32:41
But thanks for thinking of me, I guess.
32:43
You know, you can you can argue was it the right price?
32:46
Was it not the right price? I think what it does is really
32:49
validate the idea that global financial institutions need some
32:53
solution for getting their wealth management customers
32:56
access to private markets. And I think, you know, as as you
33:01
look at like what forge and equities and are very good at
33:05
doing, it's figuring out how to get small investors.
33:09
You know, the small is, is sort of subjective, but call this
33:12
like $100,000 check writers, repeatable access to the world's
33:18
hottest, most innovative Unicorn companies.
33:21
And I think figuring out how to sort of take that asset and
33:25
workflow and and match it against massive wealth
33:29
management footprints to me seems seems like a no brainer.
33:33
You asked about others in space both, you know, the on both
33:38
sides of the table, those who are still in the secondary space
33:42
as independence and then and then you know, the the large
33:46
banks and global wealth management platforms.
33:48
Is there going to be more activity there?
33:51
I think there probably will be. I mean, we have definitely been
33:54
getting calls. I think if you are a bank or you
33:57
are running a massive wealth management platform and you do
34:03
not have a solution for getting your clients access to private
34:06
markets, you are absolutely behind in, in terms of like keep
34:11
keeping up with competition. And I think, you know great
34:15
thing about these two very splashy Prince is it's put the
34:18
rest of the banking and wealth management world on notice.
34:23
Is that going to result in acquisitions necessarily?
34:25
I don't know, but it's definitely going to result in
34:27
additional partnership opportunities which which we're
34:29
excited about. I guess you could make the
34:31
analogy. It's sort of like, you know, a
34:33
casino not having a blackjack table for their customers or a
34:37
sports book not having, you know, a way to place bets on
34:40
Major League Soccer or, or something like that.
34:43
It seems like there's an appetite from the customer to
34:47
want to, you know, make investments in private companies
34:50
at the at the very least the top 10, you know, most highly
34:53
visible late stage private companies with whether or not
34:56
you can get access might be a question there.
34:58
It seems like, you know, we went through a bit of a doldrums
35:01
after the, you know, late 21, early 22 kind of crash,
35:05
particularly after the Silicon Valley Bank, you know, sort of
35:08
situation in early 22 and there wasn't much activity happening
35:12
maybe in 22 and 23. And then it seems like a whole
35:15
thing, you know, with the AI wave started coming in and
35:18
definitely with SpaceX and a few other defense tech companies as
35:22
well. And it seemed like, you know,
35:24
that secondary volume really took off again, I guess, you
35:27
know, probably last year a little bit and then this year
35:29
for sure. Yeah, I mean, important to keep
35:32
in mind just how sort of subjects to news and headline
35:37
risk this this asset class is. And in asset classes is
35:41
potentially A generous term. I mean, we're, you and I were
35:43
sort of joking right right before this of just how
35:46
concentrated the space is. And you know, we're, we're
35:49
really talking about like 20 companies or even less moving
35:53
the markets or, or potentially, you know, sort of subjecting the
35:59
market to like prolonged periods where there's just less trading
36:03
activity. If nobody's buying the same 20
36:05
companies over and over again because headlines are off,
36:10
suddenly there's just a lot less volume to go around.
36:13
This year has not been a case of that.
36:15
This year has been a fantastic year for the venture capital
36:18
secondary space. A core product that we offer is
36:22
just tracking of like transaction activity from across
36:25
the space. We aggregate trade data from
36:28
several 100 brokers and broker dealers that are active in pre
36:32
IPO secondaries. We aggregate this data up, we
36:35
anonymize it, we clean it up and we make it available to the
36:37
market. That's one of the core like
36:39
value propositions that we have as a company.
36:41
We're on track to collect about 3 1/2 billion dollars worth of
36:45
closed trade data this year. The, the data that we're taking
36:49
in is about 50% higher just in terms of like overall volume
36:54
that's been completed where we are here in the second week of
36:58
December versus where we were same time last year.
37:01
So volumes are objectively up notably we don't track like
37:07
tender activity, we're not tracking VC funding route
37:10
activity. But if you look across those,
37:12
across those metrics, like each of those categories is up as
37:16
well. Why is this?
37:18
Yeah, I mean, there's this little thing called AI.
37:21
People are excited about it like they want, they they, they want
37:24
to invest in it. And you know, it's, it remains
37:27
true that one of the only places you can get like Pureplay AI
37:31
exposure is in the BC secondary's market, right?
37:34
If you're buying an AI asset in the listed market, you're buying
37:38
a bundle of other things, right? If you're buying a hyperscaler,
37:42
you're getting some advertising, some social media exposure in
37:45
there. If you're buying NVIDIA, you're
37:47
buying a derivative on compute and sort of infrastructure layer
37:51
of AI. It's pretty tough to get like
37:53
pure AI, like model exposure unless you're writing a check
37:57
into the venture secondary's market.
37:59
So I think that's going to remain a tailwind.
38:02
We can, we can debate on whether AI is in a bubble or not.
38:06
And I actually think that this debate is going to result in a
38:09
lot more volume happening next year.
38:12
I, I think like disagreement overall is probably one of the
38:15
best things in the near term future of the VC secondaries
38:19
market because when there's conviction on both sides of the
38:22
table, it's going to result in more deal, more deal
38:25
opportunities. You look at the total mix of
38:28
trade volume that we have in 2025, only about 25% to maybe
38:33
1/3 of volume that happened this year was in AI companies.
38:37
But that that sort of doesn't jive with the mix of like buyer
38:41
interest that exists for, for these types of companies.
38:44
Consistently like more than 50% of the buyer interest in the BC
38:48
secondary's market at any given time this year was for AI
38:51
companies. So the fact that you know, at
38:54
most 1/3 of transaction activity was in AI companies indicates if
38:59
more sellers come to market, there's an opportunity for more
39:01
deals to happen, which I think will be a, what will be a, a
39:05
reason that secondary market activity increases next year.
39:08
There's kind of an interesting parallel between the Magnificent
39:10
7 in the public markets in this, you know, top ten, you know, 20
39:14
companies in the private markets.
39:16
The big difference being there's lots of metrics, financial data
39:19
and availability in the public markets.
39:21
There's sometimes that availability in the private
39:23
markets, but not not usually as much fidelity around those
39:26
financials. And certainly if you get outside
39:28
the top ten 20-30 dams, I think you're you're probably flying
39:32
blind on what information is available unless you're on the
39:35
board of those companies or have access to someone who is.
39:40
Santa's so fast he delivers presents to the whole wide world
39:43
in one night. Superman is faster.
39:46
He circled the planet, made us all forget about that earthquake
39:49
that killed Lois Lane. What earthquake exactly?
39:53
But but before we dive into that, maybe if you could just
39:55
give me a rewind on the last five years of you know what,
39:59
what caused you to start Cap light?
40:01
You know, what was the initial ideas and who are your customers
40:05
and why? Why do they find your
40:06
information valuable or interesting?
40:10
Yeah. The reason for starting Cap
40:11
Light was really recognition of just how fragmented the VC
40:16
secondary's market was in in to some extent remains to this day.
40:21
We were having a lot of success at Forge building, building a
40:24
marketplace solution. I think the problem that we
40:27
solved around helping a lot of employee shareholders, you know,
40:31
liquidate their private company holdings, that was a very real
40:35
problem that that we were solving and it was, it was
40:38
awesome to be involved in transactions that were literally
40:41
life changing for for people. I I would say it's still a big
40:44
problem and you know, I was trying to get a handle on this
40:47
the other day. I was doing some research with
40:49
Chachi BT on how many pay for millionaires there might be in
40:53
the Bay Area. I don't, I don't know if you
40:54
have any insight into this number, but I was guessing it's
40:57
not less than 10,000. It might be upwards of 20,000 or
41:00
more that have, you know, at least $1,000,000 in, in equity,
41:05
but largely they can't do anything with that unless they
41:08
sell. They're not able to usually use
41:10
that for financing for a house or borrowing against it unless
41:14
there's, you know, maybe a few specialty lenders can do that,
41:17
but not, not typically most banks, Oregon mortgage lenders.
41:21
Is that kind of a lot of the customer base you guys were
41:23
going after, is that, you know, you still see that as a big
41:26
market? Opportunity it was, yeah.
41:28
So we were we were definitely going after, you know, sub 20
41:32
employees at some of these global Unicorn companies.
41:36
And the opportunity was really education, education first and
41:40
foremost, like, hey, this is a solution for you if you want to
41:44
do all the things you need cash to do living in the Bay Area or
41:48
any sort of major Geo major, like, you know, urban area.
41:53
That's not cheap. You know kids are expensive.
41:56
You start. At home here, you're right.
41:58
Starter homes here are two, $3,000,000.
42:00
Most you know, private school tuitions are 50 to $100,000.
42:05
But this is like for elementary school.
42:07
It's kind of crazy. That and this is, and this is
42:10
pre COVID where there was a lot less remote work, right?
42:14
And so you could you could there there wasn't yet this arbitrage
42:17
of paying like Midwestern living expenses and and remoting into
42:22
to, you know, San Francisco Bay Area tech jobs.
42:25
So you know, it, it felt good to help with this education.
42:28
But you know, going back to your question of like, OK, well, then
42:31
why did you leave and start cap light?
42:33
One thing that was very clear to me was it was going to be
42:36
incredibly difficult to build a one stop shop venue for trading
42:42
in this in this asset class. And the sort of customer
42:46
acquisition cost, if you will, of going and finding the NTH
42:50
like employee or the NTH investor and getting them to use
42:54
forges platform. It was clear that that was going
42:57
to be really tough. An anecdote I can share is like,
43:00
you know, 2020 was maybe our best year, at least in in my
43:04
tenure there. You know, I remember like doing
43:07
sort of a wrap up in the fall with some of the trading
43:10
partners we had and having a call with each of these brokers
43:14
who they were like independent, they were smaller than Forge's
43:17
platform, but they did a reasonable amount of flow.
43:21
And I just remember being surprised when they were all
43:23
describing it as like the best year they've ever had.
43:26
And it was clear that we weren't like we were growing, but we
43:29
weren't necessarily like eating into anybody else's market
43:33
share. It's like the, the, the market
43:35
was just like rapidly expanding all around us And, and that was,
43:40
that was awesome. But it was sort of validating
43:44
some of my fear of like, OK, it's going to be like pretty
43:46
hard to get every investor who matters onto this platform.
43:51
At the same time, you had like, we started to see some of the
43:53
investment banks really positioning themselves to like
43:56
invest in the space, everything from form, like solidifying
44:01
partnerships, writing investment checks into other platforms,
44:06
hiring people to, to come and build out a, a dedicated
44:09
secondary's desk for them. And, you know, it was just
44:13
clear, OK, the, the banks are going to be a material part of
44:16
this market. We're not eating market share
44:19
from any of the independent brokers.
44:20
At least it didn't seem that way.
44:22
The market is growing, but there's probably some need for
44:27
some system that connects all these dots to each other.
44:30
And that was like the very, very basic idea around, OK, let's
44:34
let's build some infrastructure and tooling that connects all
44:38
the different nodes where transactions are happening to
44:41
each other. Tell me a little bit about who
44:43
your customers are, because it doesn't sound like those are the
44:45
same customers as that set of employees.
44:48
Those were mostly the sellers. It sounds like your customers
44:52
are a little bit more of the buyers.
44:54
And I'm wondering what segment of buyers are the people that
44:57
are trying to, you know, get that data?
44:59
We have two distinct types of customers.
45:02
A big part of our customer group include broker dealers.
45:06
So we think about this as like the brokers themselves or the
45:09
sell side in the traditional sense.
45:12
The other type of customer we have are asset managers.
45:15
Our asset manager customers really spend the gambit of
45:20
everybody from traditional asset managers like large trillion
45:24
dollar plus AUM like mutual fund complexes and the like to VC
45:29
funds. A lot of the top 10 VC funds are
45:33
are cap light subscription customers, hedge funds,
45:36
certainly a large and growing customer concentration for us
45:40
and then LP's. So everybody who could be
45:43
invested in one of these like GP types.
45:47
What these customers all have in common is they're institutional
45:50
asset managers. So we've got the buy side who
45:53
can be buyers and sellers of assets.
45:55
We've got the sell side who are the brokers who are usually
45:59
building relationships with the buy side to try and get them to
46:02
buy and sell stuff. The product set that we sell
46:04
into these groups starts with data.
46:07
So again, this recognition that there are hundreds, if not
46:12
thousands of brokers who are trading pre IPO secondaries, how
46:17
do we go and get all of that data to come into one place so
46:21
that the market can benefit from a little bit more pricing
46:24
transparency, which helps hopefully a price discussion be
46:28
a bit more range bound. Like instead of starting from a
46:32
universe of 0 to Infinity, when you're thinking through trading
46:35
one of these assets, you're now looking at more of a, a, a fixed
46:40
range that circles around like actual completed transactions.
46:45
We think that's the main value proposition that we can offer.
46:48
I guess it's kind of interesting.
46:49
It sounds like you're, you're collecting data from the same
46:52
folks you're almost selling the same data back to.
46:54
Is that, is that accurate or not accurate?
46:57
I, I guess I'm trying to understand, how do you, how do
46:59
you get that bowl of stone soup almost getting started?
47:04
Yeah, you kind of have to offer that you're going to, you know,
47:06
get some data from people, but give them back a larger set of
47:09
data. Is that the idea?
47:10
That's that's exactly it, Yeah. And, and, and I should say like
47:13
it, it's taken several years to get to the point where we are
47:16
today, right. I mean, we started started the
47:19
company in in 2021, we launched our first data product in 2022.
47:24
You know it, it does feel like we're at this moment where cap
47:28
light data is sort of the go to for many of the larger
47:32
institutions who consistently play in the secondary markets.
47:35
And it's incredibly humbling to look at like the logo list of,
47:39
of customers we serve at this point.
47:41
And you know, my, my intern self for my first year investment
47:45
banker self looks, looks at this list and says, man, I couldn't
47:49
get a job at like any of these places out of out of college.
47:52
And now, and now we have like the privilege of selling them a
47:55
software platform and being their like preferred like
47:59
trading venue of sorts. Like it's, it's incredibly
48:01
gratifying. Hey, no fair using reindeer.
48:06
OK, no fair using our yellow sun then.
48:08
You an alien? People seem to be guessing that
48:15
the private market is maybe 3 or $4 trillion worth of illiquid
48:19
assets. Is that is that your guess as to
48:21
what's people are paying attention to?
48:23
Looking at the market as where the last round valued the, you
48:28
know, the valuation of the company versus like where
48:30
they're trading today in the secondary market can give you
48:33
like a pretty, a pretty big difference in terms of total
48:36
valuation. We think about it as around $7
48:38
trillion of, of assets for this like Unicorn herd.
48:44
And yet it feels like the, the entire world is sort of laser
48:48
focused on like 50 of those companies.
48:51
And then the question is like what happens to all of the
48:53
others, right? And I think at, at, at every,
48:58
every layer of like capital markets, we have this like ultra
49:02
concentration risk that is happening.
49:04
You know, it's like if if you look in the public markets and
49:08
it's like mag seven, I think is like a third of the total
49:11
valuation of the of the total market cap of the S&P 500 or
49:15
something crazy like that. It, same thing happens in
49:19
venture, right? So venture fundraising, I, I,
49:23
it's, it's incredible to me how concentrated VC fundraising has
49:27
gotten where you got like 12 funds representing, I think it
49:31
was something like 30 or 40% of the total VC fundraising over
49:36
the last four years went into like 12 funds.
49:40
And then even on the deployment side of things, you know, you've
49:43
got you've got like 10 companies that are that are sucking up
49:47
like a third of all VCVC dollars.
49:51
Trading volume in Q3 according to your stats here, 83% of that
49:54
was in 15 companies and I think probably no, no surprise SpaceX
50:00
is right up there on the left. Stripe is a big one.
50:03
I'm sure Open AI and Anthropic are in there somewhere and rolls
50:07
data bricks cracking, etcetera. Are the, are these companies
50:12
pretty much the same ones from quarter to quarter?
50:14
Does that move around a lot or is it, is it still, you know,
50:18
mostly the same names? So what stays true quarter to
50:22
quarter is this concentration you're highlighting 83% is
50:27
actually quite high. Q3 was it was an even more
50:32
concentrated quarter than is usual.
50:34
But typically, at least over the last year and a half, we've seen
50:37
the concentration in the top 15 companies being north of 60.
50:41
Percent. What that means is so every
50:44
quarter there's more than 60% of the total volume we collect is
50:49
in 15 companies. The names themselves will sort
50:52
of rotate. There's usually a very high
50:55
amount of overlap, but one will fall off and be replaced by
50:58
another. Depending on that on on a
51:01
company having something like a fundraising round or announcing
51:05
a potential IPO or some idiosyncratic events that
51:09
happens at the company might get it to be, you know, a top, a top
51:13
performer in 1/4 and not on the list in the next quarter.
51:17
But there's always the usual suspects.
51:20
You know the your Space XS stripes.
51:23
These days, a lot of these AI companies have found themselves
51:26
on the list repeatedly as well. You know, do you feel like AI is
51:30
going to be eating more and more into that overall market share?
51:33
Is that maybe trending up and down over time?
51:36
Absolutely, yes. On the AI eating up more of the
51:40
volume, I think that the question is as you, as you just
51:43
alluded to, like which part of AI we think about AI as sort of
51:47
breaking out into like, you know, 3 distinct parts of a
51:51
stack with the sort of core infrastructure layer, then the
51:56
model company layer and then the application layer.
51:59
It does feel like model company has a few like determined
52:04
winners. And I think that what's key here
52:08
is just the, the, the amount of like capital requirements needed
52:12
to like operate these businesses, something that we
52:17
think continues to be a thing in 2026 and beyond is just the
52:22
amount of Co investment activity that is going to, is going to be
52:25
needed to keep the AI trade afloat, right.
52:28
So when you have companies going out and raising like $40 billion
52:33
rounds, what does that do to the VC ecosystem there?
52:37
There hardly any VC funds out there that are like a billion
52:40
dollars in a single dedicated fund.
52:43
And so, you know, as these rounds get larger and larger,
52:47
discretionary VC funds don't have enough cash to keep up with
52:51
them, right? Like your blind pool VC fund is
52:53
not going to be able to write a billion and up check to
52:57
participate in one company. And so they they naturally have
53:01
to tap other sources of capital in order to participate in these
53:04
rounds, right? And so you see Co invest and
53:08
single asset SPVS become massive, right?
53:12
Because these name brand GPS who are getting allocations in these
53:16
really hot rounds, they want to participate in these rounds.
53:19
They just don't have enough cash to to go and pony up several
53:23
billion into one trade. They put together a vehicle,
53:26
they open up that vehicle to their LP's who can participate
53:29
in that rounds. And that's going to be a big, a
53:32
big way that the AI economy funds itself.
53:36
Now what happens when those LP's stop being able to write checks?
53:40
Well, then those GPS are going to need to look to other sources
53:43
to bring in that that capital in replacement of their LP's.
53:47
And that's where I think there's an interesting meeting of
53:51
primary funding rounds in the traditional secondary market
53:54
because the secondary market will step into those SPV's.
53:57
And we've seen SPV's become a larger and larger part of the
54:01
market as represented by transaction volume.
54:04
SPV's are about half of the total market activity that we're
54:07
tracking now. That's up from just being like
54:10
15% in the beginning of 2021. So we think SP VS as a way of
54:15
like conducting transactions are going to keep going up.
54:19
And it's it's just as much a reason because rounds and and
54:23
supply is getting large as it is as SP VS actually making things
54:28
easier to trade. I wonder if that is a good or
54:31
bad thing and I'd, I'd love to hear your comments.
54:33
I, you know, we hear criticism of SP VS sometimes, you know,
54:37
because people are not aware of who the buyers are, whether the
54:41
assets are actually available in those SP VS maybe if there's
54:44
multiple layers, people are going hunting to find the
54:47
assets. What what's your view on, you
54:50
know, single asset more transparent SP VS versus this
54:54
Wild West of you know potentially multi layered SP VS
54:57
with lack of transparency or visibility there?
55:00
Yeah, I I think long term SP VS are really sophisticated
55:05
solution to solving a hard problem.
55:07
Like I think if used correctly, SP VS checked many boxes.
55:11
One is companies should be able to raise massive rounds if they
55:15
want. And if they do that, they prefer
55:19
to just face the GPS who they already have a relationship
55:22
with. I don't think there are many
55:23
companies who want to go up, go go out and like put together
55:27
financing events that include, you know, a bunch of individual
55:31
check writers. And so SPDS can be a nice way to
55:34
aggregate up demand. GPS want to be able to
55:37
participate in in these rounds, right?
55:40
They they want to participate out of their discretionary fund
55:42
if they love the company, but for some reason or another, they
55:45
maybe can't or they can't to the extent they want to, in which
55:49
case putting together a single asset vehicle and being able to
55:52
charge a little bit of of economics on that makes sense,
55:56
makes sense for the GP. And then of course, investors
55:59
clearly want access to some of these companies.
56:03
Where SP VS breakdown is when you have people going out and
56:07
using them as a way to harvest and extract as much fees as
56:10
possible or when you have people bringing to market some vehicle
56:14
that doesn't even own the underlying.
56:16
I think I think companies need to do a better job of
56:20
acknowledging that SP VS are the way the that the world is moving
56:25
and sort of the price of being able to raise large rounds and
56:29
stay private to begin with. And should have clear protocols
56:32
in place with their GPS of who can come into an SPB.
56:36
What processes need to be followed in order to vet each
56:40
each group coming into the SPB and then figuring out what
56:44
amount of information, if any, is acceptable for the GP to
56:47
share with anybody coming into an SPB.
56:51
And then on on the buy side, if you're being hit up to close in
56:54
a day with something that is like triple layered and has four
56:58
and 40 on it, just say no. Like there, there's hard,
57:03
there's probably no. Opportunity.
57:04
I got to get into this SpaceX deal, I got to get into this
57:08
open AI deal. It's my only chance.
57:11
It's incredible, like how FOMO can just permeate and, and even
57:14
these like really sophisticated investors can get swept up in
57:17
it. And you know, we've seen 3 and
57:19
30 SPDS that are like double layer.
57:22
And it, it always, just, it always makes me wonder, does the
57:27
group know what they bought into?
57:29
Sometimes we see really sophisticated investors going
57:32
into these vehicles and they've got lawyers on staff and they
57:35
know exactly what they're doing. And that's, that's fine.
57:38
Like, OK, you're underwriting that those fees as part of the
57:41
what you're paying for. Like have have fun.
57:44
Like, how I hope it's a good bet.
57:47
And who you have to feel for are those who are new to the space
57:50
and who really want access to the SpaceX or the open AI, and
57:54
they're just getting caught up into the animal spirits of,
57:57
like, trying to get one of these secondaries done.
58:05
Yeah, I think we've, we've been involved in a few cases.
58:07
I mean we, we operate primarily as a fund, but we do some SPVS
58:10
on the side, particularly when we have you know more access to
58:14
something than we would purchase out of our fund would make that
58:16
available to our LP's. Sorry we took so long folks.
58:22
We ran into a few race hazards. Lex Luthor here tried to DQ us
58:27
both permanently. Luckily we stopped him, but only
58:30
by working together. I think they're buying it.
58:34
Bump it reciprocated. I'm bumping away.
58:39
I would say there's really two different audiences of SPVS.
58:41
One, one group is maybe charging carrying fees, but one group
58:45
we've come across doesn't charge carry, they just charge fees.
58:49
And I feel like it's almost antithetical to the idea of
58:52
being, you know, a fund manager because they really don't have
58:56
any skin in the outcome of that investment.
58:59
They're just, you know, basically getting paid almost
59:02
like a broker to facilitate the transaction.
59:05
Not that that's necessarily bad, but they don't have to worry
59:09
about what the outcome of that investment is.
59:11
And I, and I think as, as you know, GPS of funds, we really do
59:15
have a stake in the outcome. It's, you know, sometimes it's
59:18
our money, sometimes other people's money, but our carried
59:20
interest is based on generating a profit.
59:24
What do you what do you think about these kind of fee only
59:26
structures where people are kind of feeders into other larger
59:29
investors? Yeah.
59:31
I think this is an excellent example of something that I'd
59:34
consider a trade. It works right now.
59:36
It'll probably work next quarter in the one after that.
59:39
But long term, we all have to question like does this really
59:43
make sense? And I think, I think the answer
59:45
is probably no. I think over time the way I
59:49
expect fees to work in the venture in the venture market
59:52
generally is you have you, you have investors that are willing
59:56
to pay some really small one time fee for beta and then you
1:00:00
have investors that are willing to pay up in the form of like
1:00:03
full fee loads for alpha. And I think that the way that
1:00:07
those things separate is you should be able to buy.
1:00:12
A a a single name SPV or even like a a basket of very commonly
1:00:18
trafficked late stage companies and not have to pay something
1:00:22
like full venture fee loads for that.
1:00:24
That's that's my view at least. Like I think paying 2 and 20 on
1:00:28
something like an index of like the the 10 most active pre IPO
1:00:33
companies doesn't make a lot of sense when you can go and do the
1:00:36
work and find, you know, 10 individual 00 type of
1:00:39
opportunities where you're maybe paying something more like
1:00:42
market one time fee, which is usually going to be like no more
1:00:46
than 3%. And I would contrast that with
1:00:50
paying for managers to actually be great managers and go find
1:00:55
the non top 20 companies and pick winners consistently.
1:01:00
And I think you should pay something more like full venture
1:01:03
fee load for for that. And the reward should be alpha
1:01:06
against like, you know, a a more passive like market tracking
1:01:11
vehicle. Yeah, I guess you're you're
1:01:14
getting into a subject that's near and dear to my heart.
1:01:16
I know we've had this conversation about, you know,
1:01:18
these secret stallions and maybe are outside the top 2030 dams
1:01:22
where there may be not on everybody's, you know, shopping
1:01:24
list. But there's some really great
1:01:26
companies doing 5000 million in revenue or more and growth.
1:01:29
And most people don't know about them unless, you know, they've
1:01:32
been around in the news, I guess on on the question of sort of
1:01:36
valuation data. How, how do you think about
1:01:39
whether buyers and and sellers, I guess are depending on
1:01:44
valuation data or whether they're just, you know, buying
1:01:46
on, you know, perceived, you know, momentum and the second.
1:01:50
Time to stock. Yeah, right.
1:01:52
Exactly. Like how, how much does, you
1:01:54
know, valuation matter at this point?
1:01:56
I guess that's something my partner and I, we cover every
1:01:59
every week. We cover, you know, something in
1:02:01
our valuation corner and really try to understand whether it's
1:02:04
worth, you know, the premium or, you know, in some cases the
1:02:07
discount. Yeah.
1:02:08
Is that something you guys feel like is fundamental for your
1:02:12
customers or or not? Yeah.
1:02:15
I mean, I, I, I think that valuation is just one input and
1:02:23
in vibes is almost like an equal input in terms of figuring out
1:02:26
where figuring out where people's pricing is going to
1:02:29
shake out. I mean, what like, it does feel
1:02:32
like just given how, how still new this AI technology is, it's
1:02:38
very easy for people to tell a compelling story around it.
1:02:42
And we're just sort of getting to the point now where people
1:02:45
are starting to have to back that story up.
1:02:49
And you know, I I think like it with a long time horizon, which
1:02:54
venture has underwriting something at a trillion dollars
1:02:59
when you're sitting on like what could be like once in a
1:03:03
generation, like once in a century type of technology
1:03:06
development. Like, yeah, you can get a lot of
1:03:08
people comfortable with that story.
1:03:10
And I don't think it's just free IPO secondaries that suffer from
1:03:14
that. You can look at the public
1:03:16
market, at least the maximum and with the with the similar lens.
1:03:20
I think Tesla is definitely that story.
1:03:21
It's not based on current car sales.
1:03:23
It's definitely based on, you know, the future of, you know,
1:03:27
driving robotics. I don't know how what else is in
1:03:30
the Elon magic, but it's data center space.
1:03:33
Yeah, exactly. How do you, how do you value
1:03:36
what is the multiple on data center space?
1:03:39
You know, it's like I, I time will tell, but the other part of
1:03:44
the, the, the question is like, who's the one valuing this?
1:03:46
So like we've seen this massive vacation of individual investors
1:03:51
willing to take like more and more risk and chomping like
1:03:54
deeper and deeper into like the risk curve.
1:03:57
And you see this in other asset classes, like the adoption of of
1:04:01
crypto, the adoption of prediction markets, Yolo trading
1:04:06
and meme stock is is certainly an artifact of this theme.
1:04:09
Pre IPO secondary is is similar, right?
1:04:12
And so you see, and that doesn't necessarily mean like individual
1:04:16
investors going and buying straight into something like
1:04:19
SpaceX or open AI, but it does mean like asset managers are
1:04:25
getting the inquiries from their investors say, I want more risk,
1:04:29
I want more concentration. And this shows up in sort of
1:04:33
bidding up all the same, the same assets or, you know, these,
1:04:37
these these rounds and secondaries being hyper
1:04:40
competitive for just like top 20 companies.
1:04:44
To your point, you know, if if you've got some information or
1:04:47
some access angle that no one else has, you can have a field
1:04:51
day looking at companies outside of that top 20 because you're
1:04:54
left to do your real evaluation work and put in a price that's
1:04:58
often going to have no competing bid.
1:05:00
I, I think there's really 2 worlds.
1:05:02
There's 1 based on narrative and momentum and vibes as you call
1:05:06
it. And then there's the other world
1:05:07
that's a little bit more of the Warren Buffett.
1:05:09
You know, what's the revenue, what's the profit, what's the
1:05:12
growth? And you know, whether you're
1:05:14
assigning some crazy multiple or some more rational multiple, at
1:05:17
least you're you're doing that evaluation based on some kind of
1:05:21
equation. I'm noting another slide from
1:05:24
one of your presentations that talks about the gradual
1:05:28
reduction in the multiple for AI companies.
1:05:32
And at least according to this graph, you're suggesting that it
1:05:35
in 24 that was north of 75 times, which which sounds a
1:05:39
little crazy. And then in 25 it's only 50,
1:05:43
only 50 X. Some some bargains on this.
1:05:47
Page I seem to remember all this 2021, you know, 2020, you know,
1:05:52
Zerp era evaluations the people were suggesting 40 and 50 times
1:05:56
revenue multiples were crazy town and you know I guess in the
1:06:00
public markets we saw SAS multiples get to 18 times or 18
1:06:04
to 20 times revenues. Now they're sort of back down to
1:06:08
more like 6 to 8 times revenues. So do you think these, you know,
1:06:12
are artifacts of kind of just the AI world and growth?
1:06:16
Or are we actually starting to see some people, you know,
1:06:19
suggest, OK, as these companies, you know, stop growing at 1000%
1:06:23
per year, maybe only two, 100% per year that we're getting into
1:06:26
more rational, you know, evaluation of metrics?
1:06:29
The story on on this page is really meant to spark a OK.
1:06:32
What is possible in terms of valuations getting cheaper,
1:06:36
becoming more compressed? If you assume that investors are
1:06:41
no longer like two or three Xing the valuation of these companies
1:06:46
because they've hit, let's call it like a critical mass and you
1:06:49
assume that the revenue growth continues at its pace, you can
1:06:53
see valuations coming into more reasonable only.
1:06:56
Two years of forward. Growth with.
1:06:59
Open AI or with not so crazy. Exactly, With only two or three
1:07:03
years of forward growth and of course, nobody coming in and
1:07:07
marking up these these companies by two or three X.
1:07:11
And it's just funny because like that last piece has already been
1:07:14
blown out with open AI saying, OK, we can, we can IPO at a
1:07:18
trillion dollars. Suddenly even just that
1:07:20
sentiment being out there is pushing the price up, right,
1:07:24
right. You know, I, I, I think that the
1:07:27
main thing that these companies are going to have to prove is
1:07:30
that the, the, the revenue in addition to continuing to grow
1:07:34
can stick with them, right? And so the, the, the concept
1:07:38
that that slide does not take into account is how revenue has,
1:07:43
how sticky has revenue been on a year over year basis.
1:07:47
And that's where I think especially as you know,
1:07:50
competition is heated up around the, the quality of the models.
1:07:53
That's where I think a lot of the investors are starting to
1:07:55
sort of like pump the brakes a little bit and seeing, seeing
1:07:59
these companies that were out in front, maybe not having as big
1:08:02
of a lead as as they initially thought.
1:08:04
Santa. Missus Claus, it is a true honor
1:08:10
to meet you. I like.
1:08:12
Thanks for blocking our view, dumbass.
1:08:14
Oh sorry, I didn't know your place was here when I threw the
1:08:16
crystal. Good thing I didn't hook it or
1:08:18
your house would be in my living room.
1:08:21
You know, this is kind of the the article we were talking
1:08:24
about before and something I'm hoping to release here in the
1:08:26
next few weeks is looking at different segments of the market
1:08:30
that those top ten 2030 companies are clearly, you know,
1:08:34
the things that everybody's paying attention to.
1:08:36
These are the biggest companies, in some cases the fastest
1:08:38
growing companies. I think there's a missed segment
1:08:41
that is, you know, the next two or 300 companies and these
1:08:45
probably aren't on everybody's radio, they're probably not in
1:08:47
the news. But these are still companies
1:08:49
doing fifty, 100 million, maybe 200 million plus in revenue,
1:08:52
still doing, you know, solid growth at 30 to 50%, some cases
1:08:56
up to 100%. And I think if you're here in
1:08:59
Silicon Valley, you probably do come across a lot of these
1:09:02
companies. You, you know, talk to people
1:09:03
who are working there, but it seems like they have a challenge
1:09:07
in finding a market for their stock.
1:09:10
Unless there's a tender offer going on or unless there's a
1:09:12
primary around going on, it's probably very hard for at the
1:09:16
very least employees at those companies and investors in those
1:09:19
companies to find a, a liquid market for their shares.
1:09:23
I'm excluding these from, you know, busted unicorns that are
1:09:25
overpriced or even this other group of slow horses that aren't
1:09:29
growing fast enough or aren't big enough.
1:09:31
What what's your thoughts about that middle group?
1:09:33
And, you know, again, I think we're talking about a trillion
1:09:36
dollars, maybe more worth of, you know, assets and
1:09:40
investments, several 100 companies, you know, lots of
1:09:43
different areas. The maybe these are AI verticals
1:09:45
that aren't as expensive as, you know, the LLMS.
1:09:48
Maybe they're not, you know, defense tech or capital
1:09:52
intensive companies, but they don't seem to be the things that
1:09:54
everybody is, you know, selling or buying right now.
1:09:57
They're not. They're not the vibes.
1:09:58
Companies, yeah, I, I mean, I think there's huge opportunity
1:10:02
there as you, as you pointed out and I think about this more from
1:10:05
the technical standpoint of if you are able to underwrite the
1:10:09
stories here and you feel comfortable with pricing that
1:10:12
you're willing to pay, you may be the only buyer in the room
1:10:15
there and. I think that's the case.
1:10:18
And and I think. There's not very many buyers and
1:10:20
and the challenge is just knowing what those, even knowing
1:10:23
the companies exist and then having some sense of what
1:10:26
they're worth is pretty tough. Yeah, I don't think most of
1:10:28
those companies are disclosing their financials.
1:10:31
And again, outside primary rounds or tender offers, it's
1:10:34
probably pretty difficult for employees and early investors to
1:10:37
get a fair price for theirs. The value to you as the investor
1:10:41
in that case compounds over time because yes, you can get in with
1:10:45
maybe a like a, a valuable entry point in the sense of like a
1:10:50
good relative value just because you're not competing or bidding
1:10:53
yourself up. But that also gives you the
1:10:56
ability to build relationships with these companies.
1:10:59
And you know, as the, as the company progresses and your
1:11:03
investment thesis plays out, there's probably additional
1:11:06
shareholders who want liquidity and the company is going to
1:11:09
think about you first. And that should reward your
1:11:11
strategy over time versus chasing the, you know, the top
1:11:16
20 that everybody else is chasing.
1:11:18
It's very difficult to build like a standout relationship
1:11:21
with those companies. And I think the way you have to
1:11:23
differentiate is going to be a a combination of Capital plus
1:11:27
ability to get that company to the next milestone, which is
1:11:31
increasingly hard to do for some of these like DECA corns that
1:11:34
are out there. I guess in terms of looking at
1:11:37
valuation sources and data, who's the canonical source of of
1:11:42
truth for it's this? Or is there more than one?
1:11:45
Obviously I'm biased, but I would say, I would say cap light
1:11:48
is the, the, the reality is there are several places to get
1:11:53
data on private company valuations.
1:11:56
And I think that trading in this asset class requires A mosaic
1:12:00
type approach where you're sort of looking at all the different
1:12:03
data sets that are out there side by side and you're making
1:12:06
an informed decision based on all the information you can get
1:12:09
your hands on. We work pretty closely with the,
1:12:12
the team at Sakura. I think that their approach is
1:12:15
really complementing to ours where they're doing like bottoms
1:12:19
up estimation around things like growth in unit economics to sort
1:12:24
of back into evaluation or evaluation target.
1:12:27
And I think that type of analysis works really hand in
1:12:30
hand with the data that we're providing, which tends to be a
1:12:33
bit more backward looking and based on transactions that have
1:12:36
actually happened that we can verify.
1:12:39
And we're going to, if you're a cap light customer, we're going
1:12:42
to give you that whole list of transactions and say for the
1:12:45
asset you care about, here's the last 10 trades that happened and
1:12:48
here's the type of stock that traded and here's the structures
1:12:51
that were being used. So that's how we think about
1:12:53
evaluation in pre IPO secondary markets and why I think we play
1:12:58
really well with with other data sets.
1:13:01
Well, Javier, thanks for spending some time with us.
1:13:03
This has been a really great conversation.
1:13:06
I, I don't know if you have any crystal ball predictions for 26
1:13:09
or anything you want to suggest. Are there companies you think
1:13:13
are going to go public next year?
1:13:14
Is it going to be a big year for IP?
1:13:16
OS, we think it's going to be a massive year for IP OS.
1:13:19
And we, we have like an IPO tracker that we, we all sort of
1:13:26
huddle over at the end of each quarter, beginning of a new
1:13:29
quarter. And the amount of activity we've
1:13:32
seen with different companies moving through each key
1:13:35
milestone of this tracker. This tracker tracks like 6 main
1:13:38
milestones, including hiring ACFO, you know, preparing for an
1:13:43
actual filing or filing. And then just like even IPO
1:13:47
rumors, there's a few other metrics we've seen the amount of
1:13:51
evaluation as a as a whole double in this tracker in the
1:13:56
last 12 months at the end of the tracker, meaning like closer to
1:13:59
IPO, Like some of the valuations in in these stages are just
1:14:03
massive with names like SpaceX, open AI.
1:14:07
I think Anthropic is like a dark horse 2026 IPO candidate hiring
1:14:13
Wilson Sonsini and sort of doing what they need to do on the
1:14:16
legal front to get ready. I also think it'd be awesome for
1:14:20
that company who's kind of viewed as like a little brother
1:14:23
of sorts or like #2 to, to open AI to come out and hit the
1:14:27
public market first. I think that that would be an
1:14:30
awesome event for for the pre IPO secondary market and
1:14:33
overall, yeah, expect a big year in in IPO activity next year.
1:14:37
All right. Well, I hope you're right.
1:14:39
We'd love to see some more liquidity coming next year, but
1:14:42
it certainly seems like the secondary market regardless, is
1:14:44
continuing to grow. So I wish you the best luck in
1:14:47
26. I hope a few folks come calling
1:14:51
on your door, whether you decide to accept their offers or not.
1:14:54
Awesome. I appreciate it, Dave.
1:14:56
Thanks for having me on. And yeah, happy holidays.
1:14:59
Hello, homeowners association. Superman's place is so close to
1:15:03
mine, I can smell it when he takes a crap.
1:15:05
Oh, fine, whatever. All right, see if Superman
1:15:08
brings your kid that new wheelchair this Christmas.
1:15:22
Welcome to another edition of Valuation Corner on Trading
1:15:26
Places. My partner in crime, Professor
1:15:28
Aman Virgie is going to be covering this week.
1:15:32
XAI. The last one of the year, we're
1:15:33
going to wrap up the Augment Top 10.
1:15:36
So this is the last one that we haven't done yet.
1:15:39
This is going to be Xai. This is in the list of all the
1:15:43
prior leaks we've done in the evaluation corners.
1:15:45
I have to admit, I went into each one of them thinking I had
1:15:48
a pretty good idea already, whether it was bullish, bearish
1:15:51
or or somewhere in between. Some of the, some of the big
1:15:54
bull bets we made this year were SpaceX and Canva Bike Dance,
1:15:57
Databricks, Open AI. We, we, we did a couple, couple
1:16:02
of months ago as well. So I think a lot of a lot of
1:16:04
bull bets we've made on the bearer side.
1:16:07
We had grayscale, Rippling, maybe Revolut and Tether.
1:16:10
I thought maybe Xai would be on the bare side of this.
1:16:14
But I want to run you through the valuation, how we how to
1:16:17
think about it and then get your take because I'm not sure I have
1:16:19
a strong even now I have a strong view on it.
1:16:22
So Xai is the developer of Grok. It's a highly conversational
1:16:26
chatbot. They are deeply in Reddit in the
1:16:28
X platform. It's this Elon Musk's company.
1:16:31
They're currently at $113 billion evaluation.
1:16:35
They're now rumored to be in talks at a $230 billion
1:16:38
evaluation, which would put them, you know, right behind
1:16:41
Enthropic and then behind Open AI and SpaceX as the most
1:16:45
valuable in the list of top five, I guess, most valuable
1:16:47
startups in the game. We'll give you a couple of
1:16:50
revenue metrics, but I want to caveat them in a second.
1:16:52
So they're $3.2 billion in in revenue.
1:16:55
They're they're talking about ARR, the end of 2025 S, whether
1:16:59
it's a revenue or ARR, we don't know exactly.
1:17:02
Their revenue has gone up 30X year over year, which sounds
1:17:07
like a lot, but I want to remind you that x.com and XAI
1:17:11
integrated about a year ago. It's a private company.
1:17:14
They're not closing all the financials.
1:17:15
We don't know all the details of this, but x.com is formerly
1:17:19
Twitter. Twitter, when Elon acquired them
1:17:21
a couple years ago, was generating between 4 1/2 and 5
1:17:25
billion in revenue. And so the fact that this new
1:17:27
entity is somewhere in the low threes in terms of revenue, and
1:17:30
presumably a lot of that is media revenue from Twitter, it
1:17:33
means that this is not the same thing as like Open AIARR or
1:17:37
Anthropic ARR. So when open AI says they're at
1:17:40
15 to 20 billion in ARR, it's not quite the same thing as this
1:17:44
3 billion because a lot of open all of open AI stuff presumably
1:17:47
is tech revenue. This is going to be some
1:17:50
combination of tech and media revenue, right?
1:17:52
So right away you're looking at that revenue number and
1:17:54
thinking, I'm not sure I fully buy that this is $3 billion in
1:17:58
ARR growing at the, at the open AI kind of rate, all tech
1:18:01
revenue. But even though I, even if I
1:18:03
give him full credit for 33 billion in revenue, this is like
1:18:07
a 35X40X multiple, right on the revenue, which seems like
1:18:12
really, really high. And at the new valuation of $230
1:18:15
billion, which they seem to be looking at right now, we're
1:18:18
talking about like 80 times revenue, which seems
1:18:21
ridiculously high on the face of it, right?
1:18:24
So I've been a lot. That's what I, that's what I
1:18:26
thought. I went and thinking, this seems
1:18:28
like a lot. This feels like a hard 1 to get
1:18:31
to get behind. Then I considered, well, what
1:18:34
about these other social media platforms and how big is X and
1:18:38
Twitter compared to some of the other players?
1:18:40
Can I get there just based on user accounts and and growth?
1:18:44
Well, X 600 million users around the world, that's, you know,
1:18:46
it's maybe 1/5 of what what Facebook is.
1:18:48
It's maybe 1/3 of where tech doc is.
1:18:51
So it's significant platform, but it's certainly not the the
1:18:54
biggest one out there. If I were to compare this to
1:18:57
some of the other AI players in the game, Open AI says they have
1:19:02
7 to $800 million weekly users. There's certainly north of a
1:19:05
billion monthly users right now. So that's five times what XAI
1:19:09
is. Google is probably $3 billion in
1:19:12
monthly active users right now. Not all of those are on Gemini.
1:19:15
A lot of those are just doing Google, Google searches, but
1:19:17
it's, it's signals the scope of the platform.
1:19:21
So just on the in terms of the metrics and the size, I don't
1:19:23
think XAI is anywhere near where these other two players are in
1:19:27
terms of the products they have and the strategy.
1:19:29
Open AI is just a very broad based general purpose platform.
1:19:32
They're behind, they've got Chachi PT, they've got Dolly,
1:19:34
they've got Whisper, they're able to do multimodal and
1:19:39
graphics based search with Dolly.
1:19:41
Gemini, I would have said six months ago was way behind on the
1:19:44
multimodal game, but they seem to have caught up in the latest,
1:19:48
the latest iterations of the model.
1:19:50
I think they both compare favorably to Grok.
1:19:52
Rock is a little bit more focused on real time more.
1:19:55
They're really limited to the X platform right now and a few
1:19:58
other, few other selected websites, but they're way behind
1:20:01
on their, their API infrastructure compared to those
1:20:03
other two. So there's an angle that they
1:20:05
have to maybe differentiate as a, as a chatbot because they
1:20:09
have the x.com data, they can leverage that for training.
1:20:12
They can they they therefore can be better.
1:20:15
I think real time and maybe new space feeds and maybe Elon has
1:20:18
certainly said he wants the more transparent.
1:20:20
Search bot not not one that has the safety or the caution
1:20:24
restrictions that Google or Gemini might put in and
1:20:26
eliminate some of the biases some of the other players have.
1:20:29
So OK, maybe there's an angle there that they can
1:20:31
differentiate. I think there's also just
1:20:33
knowing you on I think XAXAI will be the AI that powers Tesla
1:20:38
and their optimist robots and you know, potentially SpaceX and
1:20:41
all of that is going to lean on Xai in the future.
1:20:44
So I think there's a lot of lot of opportunity and a lot of
1:20:46
upside, but it's hard to justify the valuation of the company
1:20:49
just based on current revenues, current growth, user count and,
1:20:53
and think of that valuation of $230 billion is, is, is
1:20:58
reasonable. The only reason why I kind of
1:21:00
talked myself back into, hey, maybe this isn't OK bet.
1:21:03
Here's here's here's what I concluded after looking at all
1:21:06
the metrics and, and scratching my head a little bit.
1:21:08
Who are the big investors in this $230 billion round?
1:21:13
Well, the big anchor one is Valor Equity Partners.
1:21:16
It'll be NVIDIA, it'll be double line partners.
1:21:20
What do Valor Equity Partners and double line partners have in
1:21:25
common? Any any guesses?
1:21:28
I'll give you a head. The head of Valor.
1:21:30
I was going to say, yeah, Tony Grassis used to work for Elon.
1:21:34
So there's a very tight relationship there.
1:21:37
Is that similar for DBL? Yeah, Ira Aaron Prize is a is a
1:21:41
wonderful investor by the way, but he is a, he is the head of
1:21:44
double the Double Bottom Line partners.
1:21:47
And so there's a very close relationship between and he's on
1:21:50
the board of Tesla and there's a very close relationship between
1:21:53
Elon Ira from Tesla and Antonio is also on the board of Tesla,
1:21:58
by the way. And Jensen Huang is also a very
1:22:00
strong supporter of, of Elon. And so those are really the, the
1:22:02
fundamental investors who are backing this round in addition
1:22:05
to Andreessen Horwitz and Sequoia and you know, BlackRock
1:22:08
and and Kia. So there's a lot of great
1:22:10
investors in this, in this round, it seems like.
1:22:12
But I think this is really a question for you.
1:22:15
You really believe that this group of investors is aligned,
1:22:18
it is the right one in which you want to jump in at the same
1:22:21
valuation as these guys? I'll give you the, the, I'm not,
1:22:24
I'm not being snarky or difficult.
1:22:25
What I say Elon has a very tight relationship with Antonio and
1:22:28
and Ira. I don't think there's any.
1:22:31
That's not just an opinion, by the way.
1:22:32
That's if you, if you look at the Delaware Court of Chancery,
1:22:34
they actually ruled in that, in that comp package between Elon
1:22:37
and Tesla and the shareholders, they said Antonio and Ira and
1:22:41
Elon are very close friends. That was actually an illegal
1:22:43
ruling and they actually use that as a way to invalidate the
1:22:46
compact to say that Antonio and and the board at Tesla didn't
1:22:51
look out for the fiduciary responsibility to shareholders.
1:22:54
And they gave Elon a cushy package and they invalidated the
1:22:57
package and forced him to go back and do another shoulder
1:22:59
vote, another board vote and everything else.
1:23:01
So there was a little bit of a concern like that's just rounds
1:23:04
being propped up by friends of Elon, right?
1:23:06
And if you think that is, then I don't think I want to be in on
1:23:10
the $230 billion evaluation. But on the flip side, and, and,
1:23:14
and where the way that I talked myself back out of that
1:23:16
conclusion was assuming these guys are financially motivated
1:23:19
and financially aligned. So Valor Partners is not just
1:23:22
Antonio Gracias, but they have LP's and, and they work for the
1:23:24
LP's. Similarly, double bottom line
1:23:26
partners as stakeholders and IRS can be financially motivated to
1:23:30
make sure this investment pays off.
1:23:31
And obviously Jensen Huang has strategic investors interests as
1:23:34
well, but he also wants to make money on the deal.
1:23:36
Would you bet against Elon Musk, the British man in the world,
1:23:41
People like Antonio and Ira and his friends and Jensen Huang,
1:23:44
who if Elon is not the most successful entrepreneur of our
1:23:47
lifetime, Jensen Huang might might be the guy who was the
1:23:50
most successful lies in this company, in this valuation, in
1:23:55
AI. If the answer is yes, I want to
1:23:57
bet against them because I think they're not aligned, you know,
1:24:00
they wouldn't be one of my interests.
1:24:01
They have other self-serving interests.
1:24:02
I can understand passing on the round, but part of me is like,
1:24:05
if this is the group they've pulled together and they've also
1:24:07
got Sequoia and Andreessen and you know, and Kia and Fidelity
1:24:11
and BlackRock, then then I don't know.
1:24:13
I don't know if I'd want to jump in with my money at this
1:24:15
valuation on this spec. But I'm at least, you know, I
1:24:18
just want to think about the other side of it and and decide
1:24:20
whether do you think there's something really special here
1:24:23
that that's not evident in the current company and the correct
1:24:25
metrics and what they're doing. But a lot of opportunity, a lot
1:24:28
of potential between currently the X dot, the X platform
1:24:31
robotics with Tesla, maybe what SpaceX is doing like if they
1:24:34
could be the AI that. Powers all of that stuff.
1:24:37
Then, you know, then maybe, maybe this works out OK in the
1:24:39
long run. Because I'm, I'm those, those
1:24:42
are different companies, I guess I, I feel like yes, they are.
1:24:46
Well, I'm sort of in your. Camp with.
1:24:48
That. I don't know if I want to bet
1:24:50
against Elon, but I'm not always sure I want to bet on Elon.
1:24:54
Certainly the valuations that are being, I mean, I we've
1:24:58
talked about this before, Tesla's valuation is certainly
1:25:00
not based on the current level of revenue and cash flow from
1:25:03
just selling cars. It's obviously a bet of the
1:25:06
future. I guess it's OK that, you know,
1:25:10
public company investors are speculating on future revenue
1:25:13
and potential revenue streams that don't even exist yet to
1:25:16
justify the valuation for the company.
1:25:19
But it's not a bet that I would probably want to make.
1:25:22
I guess there's plenty of people who've made lots of money
1:25:24
betting on Elon. And, you know, if they feel like
1:25:28
that's a great bet for them, then, you know, go for it.
1:25:31
But but certainly the whole existence of XAI is fairly
1:25:35
recent. This company really didn't even
1:25:37
exist two years ago. It didn't have any real revenue
1:25:41
even a year ago, Arguably still, you know, somewhat limited
1:25:45
revenue. And and even in your analysis,
1:25:48
you're sort of suggesting that, you know, it's dependent upon
1:25:51
potential assets from other companies that Elon owns in
1:25:55
order to justify this valuation. So I got, I think it's fine if
1:25:59
people want to bet on Elon either in public or private
1:26:02
markets, but this is certainly not based on any traditional
1:26:07
fundamentals of valuation analysis that we're used to.
1:26:11
That's exactly where it came out.
1:26:13
It's just based looking at the numbers and what they're doing
1:26:15
and understand if they're different companies and there's
1:26:17
different shareholder groups and it doesn't, it isn't the Pencil
1:26:21
apps based on the live analysis, but it just really is even you
1:26:25
want even the people around them and even the combination of
1:26:29
you'll figure it out, it'll work out some way between this
1:26:31
company and another asset partnership and both.
1:26:35
Aware of that. Well, make these investors.
1:26:38
There's no obvious objective way to value the company.
1:26:40
It really isn't. Bet on and whether you think it
1:26:43
makes sense based on what he wants to do or not.
1:26:46
How about this one? What if I gave you three
1:26:48
options? Which Elon company would you
1:26:51
want to bet on? Would you rather invest in
1:26:54
SpaceX and $800 billion valuation?
1:26:58
Tesla at its current market cap, which I believe is somewhere
1:27:02
north of a billion, billion and 1/2 yeah or XAI at this $230
1:27:09
billion valuation, which which one you would you prefer of
1:27:14
those 3 bets? And maybe the boring companies
1:27:18
have a possibility. I think I've been on the record
1:27:23
saying I, I'm, I'm, I'm a big believer in SpaceX.
1:27:26
I think it'll be more valuable than Tesla someday.
1:27:29
I've said that for five. Years.
1:27:30
Yeah, and I wouldn't I wouldn't change my vote now I'm I'm not
1:27:33
as convicted one or the other on X dot AII can see it really
1:27:38
working out, but I think it would be it would be an
1:27:41
astonishing accomplishment that would require a lot of those
1:27:44
social media assets at X to work in conjunction with AI over over
1:27:47
a longer. Period of the.
1:27:49
Three, I think SpaceX is the is the one that I would pick of
1:27:52
those of those three at that valuation.
1:27:55
Well, good for you. That's the one we have in our
1:27:57
portfolio. So probably, you know, not a bad
1:28:00
bet. I I would prefer it to be a bet
1:28:02
at 400 billion rather than 800 billion, but still maybe a
1:28:06
little less crazy than the other two.
1:28:09
OK, well, thanks for that analysis, Saman.
1:28:11
And I guess good luck to those XAI investors out there and
1:28:16
we'll see what happens. Like you're on your Cape, you're
1:28:25
about to hit the beach with my tie in your hand.