Latest / trading places / Anthropic Hits $1T, OpenAI Struggling, Elon Reloads with $60B Cursor Deal | Trading Places Podcast E32

Transcript

Anthropic Hits $1T, OpenAI Struggling, Elon Reloads with $60B Cursor Deal | Trading Places Podcast E32 · trading places

0:00
Do you think that maybe Elon is feeling some pressure to put
0:03
together the AI component of his AI story?
0:07
Yeah, it looks like Anthropic is now passed Open AI, at least in
0:10
terms of the secondary evaluations.
0:12
One thing I will say is Naval was probably one of the smartest
0:16
and best investors on the planet.
0:37
All right, welcome to another episode of Trading Places, the
0:42
world's worst, horrible just absolutely second rate BC
0:47
secondary podcast All right, let's get started with our show
0:50
this week we're going to cover 6 different news articles.
0:53
We're going to talk about lover, the Fed, Iran, a little bit of
0:56
the IPO window. We're going to cover Anthropic
0:59
hitting a trillion dollar evaluation.
1:01
Wow, SpaceX doing a deal with cursor, opening eyes, shipping a
1:06
few new products. Angel List is opening up a
1:10
retail venture capital firm called USBC.
1:13
That's available to investors to put money in without any
1:16
accreditation. Exciting.
1:18
We'll do a catch up on prediction markets and Polymark
1:21
and call sheets. So that's our lineup for this
1:23
week. I'm on getting started with your
1:25
pick, Kevin Wash about to be confirmed.
1:28
Yeah, I think Kevin had a a better week than my girls did
1:32
playing soccer. There were actually two kind of
1:34
updated items around the Fed 1 is he was testifying in front of
1:38
Congress. The testimony was great.
1:41
I think he was thoughtful, disciplined.
1:43
He is a very clearly, he's a former Fed governor.
1:46
He was nominated there in 2006. He really knows knows the job
1:49
and qualified and I think satisfied most questioners that
1:52
he's got the competence and independence to do the job.
1:55
And then maybe more importantly, Jerome Powell, who's the current
1:59
Fed chair and was under a criminal investigation for, I
2:03
guess, among other things, lying to Congress about cost overruns
2:06
on the new Fed building. Trumped.
2:08
Trumped up criminal investigation.
2:09
It's on the word Trump here. Trumped up charges, I don't want
2:14
to go that far, but let's put it this way.
2:16
Tom Tillis, who's a Republican on the Senate Banking Committee,
2:18
basically said, hey, unless that gets cleared up, I'm not I'm not
2:22
putting forward, I'm not voting on anyone regardless of
2:25
qualifications. And those charges were were
2:28
mysteriously dropped this week against Sherman Powell.
2:30
So I guess he had a. Surprise.
2:33
So that that now allows Warsh to get through the conformation.
2:36
The Senate bank committee, I think will approve him, maybe on
2:39
partisan lines, maybe even he'll win over a couple of Democrats.
2:42
And I don't see a obstacle. So by May 15, I think there's
2:45
when, which is when Powell's term is up, that's we should see
2:48
a new Fed chairman in place. But does this mean that we're
2:50
going to see interest rates drop?
2:53
Is is Kevin Warsh in the pocket of Donald Trump or not?
2:58
You know that's a, that's a great question.
3:01
I am going to guess there will be no interest rate changes
3:05
between now and let's say September.
3:08
And why do you think that is? Yeah.
3:10
I do think he is independent minded.
3:12
He has written extensively in previous years about the
3:15
importance of the Fed balance sheet.
3:17
Not so much interest rates, but the balance sheet being tight.
3:21
And I don't think he's going to do anything to compromise his
3:24
credibility with markets. He's also one vote out of
3:28
basically 12 people who vote on interest rate cuts.
3:30
And if he can't convince majority, I don't think it's
3:33
going to go out there to push rates and then the majority
3:34
won't go along. So he's got a little bit of work
3:36
to do, I think between now and let's say September to make sure
3:39
that he's maintaining his credibility, He can build a
3:41
consensus within the group. I do think we will see lower
3:45
rates, but probably, you know, 25 to 50 basis points probably
3:48
at the December meeting. That's that's that's my best
3:51
guess. Well, I guess somewhat driving
3:53
this, you would have to say the Iran war situation and traffic
3:58
in the Straits of Hormuz is a big part of this conversation.
4:01
The Strait of Hormuz. It's I've I've never gone from
4:05
not knowing a thing existed to talking about it constantly that
4:10
fast since the air fryer. It doesn't seem like the war is
4:18
on, but it doesn't seem like the war is off.
4:21
Mutual blockades going on, conversations potentially in
4:24
Pakistan that we're on and now are off again and then we're on
4:28
and then we're off again. I don't know if we can get to
4:30
any sure story on what's going on, but if we go to Polymarket
4:35
and check out what the betting is on the Strait of Hormuz,
4:39
certainly doesn't look like the end of April is going to happen.
4:41
This was earlier at least at 3040%, but now almost at 0.
4:47
If we look at by middle of May, still not very much chance,
4:52
maybe around 20%, that rises to maybe close to 40% by the end of
4:57
May. And then maybe by end of June we
5:00
might get to a better than 50% chance.
5:02
So kind of seems like it's at least another month or two,
5:06
possibly longer before we return to any normal traffic.
5:10
You feel like that's the right read Aman.
5:12
Yeah, you know, there's a ceasefire I guess is the current
5:15
state of play. There's a, it's instead of
5:17
mutual blockade. Last I checked my international
5:20
law textbook, a blockade is an act of war.
5:22
So. So I guess that.
5:24
Means. We are, you know, I guess we're
5:26
texting the at war and it's supposed even the threat of
5:29
disruption is just making, you know, traffic slower in the
5:33
straits. And I I do know that people with
5:35
supply chain dependency in the straits are now looking active,
5:37
looking for other ways to move oil in commodities overland or
5:41
around the straits. So there will be, I think, a
5:43
substantial return to normalcy over the summer, whether it's
5:46
end of June. That still seems a little
5:49
optimistic to me personally. I just, I just don't know how
5:51
these two sides were this far apart on some key issues can
5:55
reach an agreement, but the both sides are very motivated to make
5:58
a deal. So, you know, maybe, maybe
6:00
that's right. I might be thinking this might
6:02
be drags out a little bit longer than the market thinks it does.
6:06
And here's Terry Fernow. She's a clown shoe resoler.
6:09
You know what they say about clouds of big feet.
6:11
Oh yeah. Well, moving on, it seems like
6:17
Anthropic is being valued at a trillion dollars.
6:21
There's rumors of an $800 billion valuation that the
6:23
company turned down, but on secondary markets, it's already
6:27
trading at a trillion dollar valuation and overtaking Opening
6:31
Eye. Opening Eye had raised at an
6:33
$850 million valuation, but probably isn't trading any
6:37
higher than 700 billion. So really quite a a switch of
6:41
fortunes here. Yeah, it looks like Ampropic is
6:43
now past Open AI, at least in terms of the secondary
6:46
evaluations. That's interesting.
6:47
They had a, they did have a pretty active week.
6:49
There were two deals announced this week with investments from
6:53
Google and from Amazon. Google, it says they're putting
6:56
in a $40 billion commitment at a $350 billion evaluation, but
7:01
it's 10 now and then 30 later based on certain milestones.
7:05
And then Amazon is a $25 million commitments.
7:07
They've already put I think 8IN already before.
7:10
So there's another 5 coming from Amazon now and up to 20 in
7:13
future cash based on commercial milestones.
7:16
And again, that valuation is around 3:50 to 380.
7:18
So they're giving better deals to their suppliers and and their
7:21
vendors than where we're seeing at the secondary market.
7:23
And that's just, I guess a way of these cloud for equity deals
7:26
are strategic. And so Anthropic is playing the
7:28
smart game, I think by giving them stake in the company.
7:31
I think both Google and Amazon own between 15 and 18% of the
7:35
company. Now, if I have my cap table math
7:37
right, they're trying to both keep below 20%, which would
7:41
trigger change of control provisions and, and affect how
7:43
this company gets valued going forward and, and their listings.
7:47
So, but both of them are bumping up against 20%.
7:50
And so now those are two companies who really want the
7:53
Anthropic to succeed. I think that sets them up for a
7:55
great IPO in the second-half of the year.
7:57
We talked about the financial success if we do get to 7080 or
8:00
100 billion in ARR by the end of the year, a trillion dollars is
8:03
about 10 times ARR by the end of the year.
8:05
That makes sense to me. That seems like that's the
8:07
valuation on their IPO day. Well, getting back to IP OS,
8:12
Elon announced that there was a deal that SpaceX has made with
8:15
Cursor where they're using Cursor technology for about $10
8:20
billion or they could acquire the company for 60 billion
8:23
outright based on the work together.
8:25
So pretty amazing deal that's valuing Cursor at 60 billion.
8:28
Do you think this is good for SpaceX shareholders about I
8:32
guess we are SpaceX shareholders ourselves.
8:35
We are indeed, yeah, I, I think it probably is.
8:37
So the first point is this is as you said, this is not an
8:40
acquisition. It's actually a, it's an
8:41
exclusive partnership with a, a call option, an exclusive
8:45
option, but. Could be an acquisition.
8:47
Could be between now and the end of the year, they could exercise
8:49
the option. It would then be the biggest
8:51
acquisition of all time, I think for a privately funded company
8:54
with only one exception, that's XAI that 250 billion.
8:57
So it is actually a because it's an option to buy and not and not
9:01
a locked in price. That actually is a benefit.
9:02
I think to basics shareholders that the reporting of the
9:06
weekend made it seem like this is a Theta complete and A and an
9:08
acquisition. But this option is important
9:10
because we know Elon has once before tried to, tried to walk
9:14
away from a deal when he was unhappy with it, right?
9:17
That was Twitter. Anyone working at Twitter knows,
9:20
remembers this, that episode. But I don't think anyone wants
9:23
to walk away. I think the reason why it's a
9:24
it's a call option to buy is basically because the company's
9:29
in their IPO review process and their roadshow.
9:33
And if you have a new company being acquired, you have to
9:35
consolidate the financials, you have to refile your S1, you have
9:39
to redo the risks and disclosures.
9:41
Probably God forbid there any differences in accounting
9:43
practices for the new company. You've got to iron those out.
9:46
And then the SEC takes like 12 weeks to review.
9:48
So if you want to go public in June, you you can't actually buy
9:51
the company. You have to you have to
9:52
structure it this way as a partnership with an exclusive
9:55
right to buy. There is a chance, I think that
9:58
SpaceX walks way if things don't go to plan on the IPO or if they
10:03
have a change of heart. But even even if they do, I
10:05
think the bottom line is SpaceX, you know, seems to they seems to
10:09
they seem to want this fully integrated AI infrastructure to
10:13
get to outer space. And cursor is a a coding agent.
10:18
Coding is like the white hot center in the red hot AI
10:21
universe right now. And and cursor is there the fast
10:26
strong SAS company of all time. There are two billion revenue
10:28
run run rate right now. They're projecting 6 billion by
10:32
the end of the year. Their only obstacle was compute.
10:34
They were hitting a a wall and compute that they couldn't
10:37
compete with the hardware budgets of open AI and entropic.
10:40
And guess, guess what SpaceX and Elon have?
10:43
They've got they've got money. They've got the biggest
10:45
supercomputer in the world, Colossus and in Memphis, TN.
10:49
And so I think it makes sense. I think it's a great marriage
10:52
between these two companies. Do you think that maybe Elon is
10:56
feeling some pressure to put together the AI component of his
11:00
AI story? I, I think with, you know, the
11:03
XAI acquisition, which we speculated was probably out of
11:06
the slightly inflated valuation, some of the talent at XAI headed
11:11
out the door almost immediately after the merger.
11:15
I think you're, you've identified one thing that XAI
11:17
did have, which is they had a budget to spend money on AI
11:22
compute maybe Cursor has pretty interesting technology and user
11:27
base and growing revenue, but needed, you know, that budget
11:31
that was coming both in, you know, the $2 billion that
11:35
they're just raising. But now, you know, Spacex's
11:37
budget, I guess. Yeah, Seems like a marriage of
11:41
convenience, perhaps. Yeah, perhaps it, it is
11:44
interesting because as you point out they spent 250 billion or I
11:48
guess better said they added they diluted the company by 20%
11:51
in order to acquire X dot AI, which included some components
11:55
of what AI needs to be successful in particular.
11:59
So over $10 billion of expense. Bro, did they OK?
12:03
I might. Oh, yeah.
12:05
The costs, the ongoing burn in the business.
12:08
Right, right, right, right. So there's that.
12:11
And you asked yourself, what do they get for it?
12:13
They definitely got Brock and a lot of folks on the team,
12:16
although you said, as you said, maybe some folks there have
12:18
left. And then the, the Twitter data
12:20
repository, there's it is the maybe one of the largest
12:23
depositories, if not the largest repository of data in the world
12:27
by which you can train AI, especially LLMS. But XAI was
12:31
also missing, I think coding, which is more of a recent
12:34
phenomenon and, and that's what CURSOR fills.
12:36
So I guess having made the $250 billion, put it this way, I, I
12:40
feel better about the CURSOR $60 billion bet than I do about the
12:43
$250 billion bet on XAI. I kind of think in retrospect
12:48
they just overpaid for XAI and and CURSOR fills an important
12:52
piece of the puzzle. So maybe that maybe they.
12:54
Overpaid a lot for XAI, but SpaceX shareholders were diluted
12:59
heavily as a result of that transaction.
13:01
You certainly could have bought the Twitter as that's alone for
13:04
maybe only 20 or 30 billion. So paying $250 million for XAI
13:09
in order to acquire over $10 billion a year worth of
13:12
expenses, I'm not sure that was such a great deal.
13:14
And, and, and I agree with you, Cursor in comparison is a great
13:18
deal and only and only $10 billion for the partnership and
13:22
maybe $60 billion for a real company.
13:25
You're paying 1/4 of what you paid for XAI, probably for more
13:28
value there I think. Yeah, for putting a a happy face
13:31
on it. I guess the other point is when,
13:33
when SpaceX made the XAI acquisition, they were valued at
13:37
800 billion and the new and then they got up to a trillion.
13:40
So it was, you know, 20% dilution the if the, if the, if
13:43
SpaceX does go public at a trillion dollars and then trades
13:47
up to 2 trillion, let's hope. And then at that point the this
13:51
deal closes. This is actually 3% dilution for
13:54
SpaceX. So because it's a
13:56
forward-looking contract, it's structured better.
13:58
I think the valuation's better and I think hopefully the, the
14:00
coding piece is an important piece of the puzzle.
14:02
So we'll, we'll all be happy shareholders at the end of the
14:04
year, hopefully. Well.
14:05
Everything becomes a rounding error if you value Elon ET al at
14:09
$2 trillion. God, he's still going.
14:13
Oh, how can he? Oh, he must be feeling really
14:18
stupid. I don't think he's feeling
14:19
anything from the neck down. The third big company in our
14:23
triple big three is open AI and open AI kind of been on the back
14:28
foot lately for a lot of reasons, but they managed to
14:31
ship two major projects out the door.
14:33
Both GPT 5.5 and images 2 Point O were released this past week.
14:38
You know, we should note that Open AI is still the largest
14:41
consumer use AI company out there.
14:43
You know, if we're looking purely from a market share
14:45
perspective, Open AI is, you know, the biggest company in
14:49
consumer AI. Maybe we see Gemini and Grok
14:53
starting to take a little bit of market share, but although a lot
14:56
of stories around Claude, they're really leading number
14:58
one in enterprise but not really in consumer.
15:02
So how do you feel about these Open AI releases?
15:04
1 Is this going to really open up a new opportunity for them to
15:08
create some kind of AI super app for consumers?
15:11
Well, if I set the valuation aside, so on these two products,
15:14
GPT 5.5 was released on April 23rd.
15:17
It's only four days ago, but I've seen on the early reviews
15:20
is it does have significant breakthroughs in agentic coding
15:25
versus like autonomous research. This is really I think we're to
15:28
set Claude apart and where cloud has become such a big deal.
15:31
So we're talking about agents now as opposed to like finding
15:34
the answer. It's like actually do stuff for
15:36
me. And this is their first salvo in
15:39
that territory, which is clearly the future of AI.
15:42
And it seems like so far the reviews are pretty positive.
15:44
So that's good. And then ChatGPT images was a
15:48
that was a massive update to image generation, which I think
15:50
it's just been an area where even where Claude has struggled
15:52
to bet. So I think there's I think that
15:54
is a big step function in the right direction either in the
15:57
conversation, now they're in the game.
15:58
I don't know that they are, you know, going to perform better
16:00
than Anthropic or Gemini, but they're definitely in the game
16:04
on both images and on agents. That's clearly where the puck is
16:08
headed to, to quote my favorite hockey player of all time, Wayne
16:11
Gretzky. And so I feel like it's his
16:13
legitimate 3 horse race. And maybe with Brock being, you
16:16
know, with cursor, maybe it's a four horse race.
16:19
So just being in the game I think is is great.
16:22
I don't know that they are 850 billion in terms of that
16:25
valuations based on the fans. So I don't know that that
16:27
necessarily makes sense. That may still require a reset,
16:30
but from everything I'm hearing comes early reviews and and
16:32
buddies of mine who are engineers are trying them out.
16:34
They're already impressed with ChatGPT 5.5 and with images.
16:39
So I, I, I guess I would say there's still a ton of value on
16:42
the consumer AI usage standpoint.
16:45
You know, maybe you're concerned that Open AI hasn't been growing
16:48
their market share, but hopefully they've at least been
16:50
keeping market share. You've got, there's a chart
16:53
showing Dows or daily active users and, and especially in the
16:56
mobile instantiation, it looks like, and I think on that, you
16:59
know, this is just exploded and open the eye is, is gaining,
17:02
everyone else is gaining a little bit faster, but that's
17:04
just how as the market matures. But they haven't done as well as
17:07
crack the enterprise market and there's a lot more money in
17:09
enterprise than in consumer. And so the Dows don't translate
17:13
directly into money, but I think that's just more of a valuation
17:17
consideration. On the consumer side, they're
17:19
definitely in the lead. They have a great product that's
17:22
getting better. I think that is a very valuable
17:24
franchise that will also translate into success on the on
17:27
the enterprise side. But on the enterprise side, I
17:29
think Claude is now the clear market share winner there, open
17:31
as the clear market share in the consumer side.
17:33
So if those are two big prizes they're going after.
17:39
Something's bothering him. I think we've learned that in a
17:41
photo shoot at an Indian gaming casino and it didn't work for
17:44
him. Here.
17:45
He's fucked up, like you said. Oh, and he takes a full load
17:48
from Herbie. He doesn't know whether to
17:49
scream or smile. And our next story is about
17:56
Angel List launching a private AI fund for access to retail
18:01
investors. This was announced just a few
18:04
days ago. In fact, Naval Ravikant, the
18:06
founder of Angel List, announced on his Twitter account.
18:10
USVC is a allowing you to buy the future, and some of the
18:14
companies that they have in their portfolio immediately out
18:17
of the gate include Anthropic, Open the Eye, Sierra Purcell,
18:22
Crusoe and Lagora, and Xai. There's quite a bit of
18:25
announcements that were made and then quite a bit of fuss over
18:29
the cost of what those, what Angelus was charging for here.
18:34
If you look at the disclosures, it sounds like there's a little
18:37
bit more than a 3% fee structure there and, and no direct carry
18:42
being charged by Angelus, but the ways that they're acquiring
18:45
those companies might incur underlying carry.
18:48
Some of the folks who were weighing in on that controversy
18:52
suggested that, you know, that 3.6% gross annual expense ratio
18:57
may be capped at 2.5% net. This might be a little bit
19:00
expensive. I, I don't know if it's any
19:03
different than other retail funds that are available to
19:06
people, notably, I guess Kathy Woods ARC fund and maybe the
19:10
Dusty XYZ fund. And I think Fundrise is another
19:13
fund. Then we had Robin Hood's RVI
19:17
fund that just came out recently.
19:19
So it seems like there's quite a bit of retail access now to
19:23
venture capital, but I think even with this one, there was
19:27
some confusion about what those actual fees involved.
19:30
What do you think about this structure amount is, is that
19:34
expensive or cheap or how much is it really being paid by the
19:38
retail market? Yeah, well, let's, let's help
19:41
Angel List a little bit. Let's try to clear up the
19:44
confusion and their fees, shall we?
19:45
And then we can talk a bit about the other side.
19:47
But just like what do you get for it and if they have a a good
19:49
strategy and access. So here's what we know from USVC
19:53
and an Angel List filing. They have, they have reported a
19:56
1% management fee, that is the base fee paid to Angel List.
20:02
Some of the confusions on them, well, what about these other
20:05
fees? There was a .95% acquired fund
20:09
fee that this is a pass through cost from underlying venture
20:12
funds like you mentioned. So management fees charged by
20:15
the VCs who actually hold the shares that then get sent to
20:17
Angel that get paid by Angel List and then sent through to
20:19
their investors so. Yes, you the investor have to
20:23
pay the underlying fees on fees and then you're, you're paying
20:25
the 1% Angel list. And then there are these other
20:28
expenses. They, the good news about Angel
20:31
list, I guess is they have a, they're open to a minimum of
20:34
only $500 and dealing with non accredited investors.
20:37
But that's a big compliance burden.
20:39
And to be a regulated SEC fund requires a lot of legal and
20:42
operational costs and all those costs add up.
20:45
And so the total expense ratio or TER that you mentioned is
20:49
3.61%. That sounds like a lot if if
20:53
what you're used to a mutual fund world or hedge fund worlds
20:56
is like, you know, you can, you can buy an index fund, the S&P
20:59
500 index fund for 1/4 percent load, half a percent TER.
21:03
So this sounds like a lot, 3.61%.
21:06
As I said, Angeles has capped it through a contractual waiver to
21:09
low cap the ratio 2.5% through October of this year.
21:13
But if the gross number stays in 3.6%, that might get passed on
21:16
to the investor. On the flip side, this will be
21:19
brand new news to a lot of our VC listeners.
21:21
There's no carry, no carried interest.
21:23
You might ask why aren't they charging carry?
21:25
Like don't funds charged 2 and 2020% carry.
21:28
How do they make money? The SEC rules around this Public
21:31
listed vehicles are very weird. They don't.
21:33
They don't allow you to charge carry.
21:34
You actually have a preference that you charge fees and
21:36
disclose them and not not do a profit share.
21:39
Which to me is like backwards. I would actually want a manager
21:41
to not charge me that much in fees and then to charge me
21:44
profits share and carry. If they do well, I want to do
21:46
well. Just seems like better alignment
21:48
to me. But for whatever reason, the SEC
21:50
based on 1930s rules doesn't doesn't allow carried interest.
21:54
So you're paying about 3.6% all in as a management fee, no
21:56
carry. And so I guess that's a little
21:58
on the expensive side. If your benchmark is the S&P 500
22:01
index fund, I think it's OK. If you're used to the VC 2 and
22:05
20 cost structure, you probably end up in the same place, maybe
22:08
more upfront fees and less carried interest.
22:11
The question I guess is, are they going to be good investors
22:13
and how do they how do they select and get access to deals
22:16
and that you know, I don't know, Dave, what do you think?
22:18
Would you, you know naval better than I do?
22:20
Do you feel like Angel List is set up to get into quality deals
22:23
and good prices that will make that that will justify the fee
22:26
especially with this no carry? Or do you feel like they might
22:28
be challenged getting in on the best deals?
22:31
Well, I I think it's a little bit of a mixed bag here, but one
22:34
thing I will say is Naval was probably one of the smartest and
22:37
best investors on the planet. I started investing with Naval
22:41
and Angel List over 15 years ago.
22:43
In fact, when I was running 500, we were an investor in Angel
22:45
List itself. The reason it's called Angel
22:48
List was originally it was just a list, an e-mail list.
22:53
And I think there was about 10 or 15 of us that started on that
22:56
list with Naval. Gosh, I think this was back in
22:59
2010 or 11. And it really came out of, you
23:02
know, the same audience that Naval and and Nivi were running
23:06
with Venture Hacks way back in the day.
23:08
So Naval has always had, you know, ground floor access to a
23:11
whole bunch of amazing companies.
23:12
He was early in Uber, he was early in Crypto, Just one of the
23:17
really smart guys there. I mean, on the website here
23:20
we're we're seeing pictures of Naval along with Steve Jobs and
23:24
also with Elon. But I actually would say that
23:27
that that's not an egotistical comparison.
23:30
Naval really is a pretty smart guy.
23:32
But I think the story that's being presented here, they're
23:36
talking about 8 or 9 very high profile names in venture and
23:40
providing retail access to that. I think that's probably what
23:43
they lead with. But I think the ongoing story is
23:46
going to be the vast amount of fund managers and companies that
23:49
Angel list has that really aren't like, you know, top ten
23:53
visible names. You'd see the tag line here that
23:55
says back the company's building the future before it's obvious.
24:00
And I think the shoe that really hasn't dropped there is the fact
24:03
that they have access to a ton of companies that aren't, you
24:06
know, mega billion dollar, you know, visible companies.
24:10
And over time, I bet you they start introducing more of those
24:13
companies into the portfolio when there's still maybe pre
24:16
IPO, you know, much earlier stage, maybe series B&C stage
24:20
companies, not necessarily these mega $100 billion, you know,
24:24
sort of stories. Well, if they can get in on
24:27
those companies at a reasonable, I guess the other question is
24:29
the evaluation and what is the what, what is the evaluation
24:31
they're getting in on, especially for those not, not
24:34
the top 8 or 10, I guess we're leading with, but for the next
24:36
100 or so, if they can get in and on good valuations and
24:39
pricing, then then I think I can see them, you know, doing well.
24:43
And then the, and then those fees are going to seem like a
24:46
bargain if that fund can return, you know, 15 to 20% net IRRS,
24:50
which a high quality manager in that space should be able to do,
24:54
especially when it compares from itself to Steve Jobs and Elon
24:56
Musk. Try the.
25:01
Buffet Yeah, I'll try. Here's Jody Macaroni, a lunch
25:03
lady counselor from Boise Unified School District that
25:09
calls for an NXC. Impact.
25:12
Our last story today is going to be on prediction markets and
25:15
Polymarket and Kaushi, which we featured a couple times before.
25:19
Kaushi and Polymarket were attempting to raise a $20
25:21
billion valuations. This is a Wall Street Journal
25:24
article from a little over a month ago.
25:27
And then we saw Kaushi raise at $22 billion maybe at the end of
25:33
March that was already reported. And Polymarket now is in talks
25:38
to raise money at about a $15 billion valuation, a little bit
25:40
lower, but still both of these companies have been on a tear
25:44
and raising money and generating revenue.
25:47
I'm on what do you, what do you think there's a difference in
25:49
the valuation of Polymarket and Calci?
25:52
Are they, you know, different in terms of revenue or growth or
25:55
are there other things going on here?
25:58
Polymark had a little bit of a later start in the in the US
26:01
market. Kalchi is probably over a
26:03
billion in revenue for 2026. I don't think Polymarket is
26:07
going to be quite that high. They're probably between 1 and
26:11
400 million just based on their more recent entry to the US The
26:15
they charge lower fees than Kalchi by a little bit.
26:18
They're a more recent entry into the US because Kalchi's been
26:21
here through a they're ACFTCA regulated exchange and have been
26:23
here operating legally in the US for a couple years.
26:26
Polymarket just got into it end of end of last year, excuse me,
26:30
end of 2024, early 2025 S beginning of last year, 2025.
26:34
There are technologies different there.
26:36
Polymarket's kind of built on Polygon blockchain.
26:39
They accept USDC. Calcia is just US cash and you
26:43
know, more like a traditional exchange and off chain.
26:46
But I think there are because there may be, you know, 2X on
26:48
revenues and have been in the US longer.
26:50
They're probably getting a little bit of a premium in the
26:53
in the current valuation. But I think I think in the next
26:55
couple years both will be, you know, big players, they're
26:57
definitely taking share from from like FanDuel and sports
27:01
betters in general and casino. So I think both of these are
27:04
going to be really interesting stories do.
27:07
You think there's any risk of legislation from people trying
27:12
to restrict prediction markets and access like, I guess
27:14
there's, you know, been sort of a, a ban against sports betting
27:18
in many states, but that's all of a sudden started to go away
27:21
with access to prediction markets and access to other
27:24
traditional platforms? Yeah, for sure.
27:27
I think all, I think a bunch of states are looking at it now
27:29
there it this been a slow roll towards legalizing sports
27:32
betting on a state by state basis.
27:34
Their major, a bunch of major states are still holding out and
27:37
there's a legalizing sports betting in general, but
27:39
Polymarket and CFTC both got CFTC approval.
27:42
So they're able to use federal law to operate and now the
27:46
states will come back and, and look at it.
27:47
So I think there will be legislation probably around age
27:51
verification. We talked a bit before about the
27:55
court case in California against social media, where for the
27:58
first time a social media company was found liable for
28:00
creating quote UN quote addictive products.
28:04
And so there's going to be some liability around that, which I
28:06
think is going to affect probably how they go to market
28:08
and how they roll out and do age verification.
28:10
Not I think the legality of it. I think most in most states
28:13
there seem to be sympathetic majorities on allowing it.
28:17
But how it gets done and some of the legal liability around it, I
28:20
think are definitely unwritten questions, which just means
28:23
these companies are going to be a little bit.
28:25
It's tricky to invest in these companies unless that stuff is
28:27
cleared out. I thought there was some recent
28:30
tussle between the CFTC and the state of New York and the
28:34
ability to block, regulate oversight of prediction markets.
28:39
Do you have a feeling on who's going to gain control of that
28:42
situation? I don't CFTC of is a federal
28:46
regulator, the state of New York just based on the supremacy
28:48
clause and how federal and state laws interact.
28:51
I think the feds will will probably, you know, have the
28:55
ability to, if it's a CFTC regulator product and you're
28:57
complying with CFTC rules. I think New York can be limited
29:00
in terms of what they can do, but I don't I can see this, you
29:03
know, I can see this being an ongoing legal Gray area for for
29:06
quite a while. And different states select
29:07
different solutions. All right.
29:09
Well, it's been quite a week in news, Aman, you'll be covering
29:13
cursor in our valuation corner later today.
29:15
That's right. And here's Fred Kaczynski, the
29:20
e-mail bomber. Oh, Fred, you've got jail.
29:31
All right, Professor Virgie, we're back with another edition
29:34
of Valuation Corner, and this week we're going to be covering
29:36
Cursor. Cursor's in the news lately,
29:38
isn't it? It is.
29:40
It's a pretty amazing company. We're going to be talking about
29:42
the company as a standalone. This is a $50 billion evaluation
29:47
based on an April 2026 Series E that they were raising.
29:52
And I'll tell you why I love the company.
29:54
I didn't love the valuation at $50 billion.
29:57
The only way you make money investing at $50 billion is you
30:00
got to find someone who's going to pay more than 50 billion for
30:03
the company, who's got deep pockets, who needs their
30:06
technology, maybe who has a compliant board of directors who
30:10
will do what they want. And guess what?
30:12
Just this week they found it. They found the perfect partner
30:15
in SpaceX and Elon Musk who now wants to pay more than $50
30:19
billion for the company. So the that round is now off the
30:22
table. We actually have friends were
30:24
trying to wire into that round last week and we're shut out.
30:27
They were like, hey, where's the bank wire instructions?
30:28
Where do we send the money? And it didn't happen because if
30:31
the announced merger with SpaceX.
30:33
So, so we'll talk a bit about the company and then about why
30:36
SpaceX wanted to buy them and maybe what this means for SpaceX
30:40
and for SpaceX shareholders. So level set on cursor.
30:44
What is cursor? They're an AI coding agent.
30:47
They basically integrate with AI with human input to automate
30:50
coding tasks. They're like the user interface
30:52
that they used to work with underlying AI like Quad is
30:57
Anthropics version of this Codex is open a is version of this.
31:00
That's what cursor is now where AI is evolving.
31:03
You can use it for a lot more than coding.
31:05
Although that was their their history.
31:07
It is the fastest growing SAS company of all time.
31:10
They're at a revenue run right now of $2 billion.
31:12
They were. And that's actually just as if
31:15
the start of this year, they're planning on being as high as $6
31:17
billion by the end of the year. So explosive growth in the
31:20
company. It's a company founded by
31:22
basically a bunch of young MITI think they're all dropouts.
31:26
I don't think any one of these folks has a degree from MIT.
31:31
Mitchell is the CEO and founder. He was an intern at Google.
31:34
I think it's the most senior job that he had before deciding to
31:38
start his own company. Great educational credentials.
31:40
He went to Horace Mann High School in the Bronx in New York.
31:44
That's where my daughter would have gone if we hadn't moved
31:46
from New York to the Bay Area. The other founders and members
31:51
of the management team, Arvid A Monsinger and then Swallow, are
31:54
all currently with the team. They all went to MIT.
31:57
And again, I don't think any one of them finished their degree,
31:59
although I'm sure MIT will allow them back on campus if they if
32:02
the net want to come back. So what does Cursor do and how
32:05
do they compare with the competition?
32:06
It is a very heated space. I mentioned ChatGPT, clod and
32:10
Opus from Anthropic. This is just based on their
32:13
website. This is from any Spheres
32:15
website. Any spheres, the parent company
32:17
of Cursor. They're showing that they are.
32:20
Their performance is 1 notch behind GPT 5.4, but a little bit
32:25
ahead of Opus and their first Gen. of their product called
32:28
Composer 1.5. So they're in the conversation
32:31
for 3rd generation proprietary AI models.
32:34
There's a little bit of self reporting going on here.
32:36
I've seen independent research that puts them in the
32:38
conversation. This is from their own website.
32:40
So, you know, when people do self assessments, sometimes they
32:42
have an inflated sense of how they do, but it seems like from
32:46
most third party evaluations, they're in the conversation and
32:49
and as good as or close to GPT 5.4.
32:52
Just this week and Open the Eye launched 5.5, which might
32:56
improve on these benchmarks, but they're definitely a high
32:58
quality product and they're in the conversation.
33:00
Customers seem to love them. As I said, they're ending this
33:03
2026. Their projected run rate right
33:05
now is $6 billion in revenue. So the trajectory implies to
33:09
tripling the revenues basically every every year.
33:12
So that's from 2 billion and then the last year under 1
33:14
billion a year before that. So very rapid growth in the in
33:17
the company. So that was the standalone
33:20
company. Let's talk about why SpaceX
33:21
wants to buy Cursor. It seems like SpaceX and Elon
33:24
have determined that they need a full stack AI infrastructure to
33:29
go to outer space whether they need to or not.
33:31
Let's pen that for a second, but let's assume that they do.
33:34
SpaceX, you might remember, acquired Grok and XAI at the
33:39
start of this year. But Grok and XAI have a lot of
33:42
capabilities that you need to make AI successful.
33:45
There's a lot of data on Twitter.
33:47
There's the LLM foundational model that Grok has.
33:51
There's a lot of computing capacity that Grok has.
33:54
However, they don't have the coding, which has now become the
33:56
very, very hot space and AI. And if you had to, you can
34:00
either, I guess, build your own coding agent.
34:01
But if you had to acquire someone who's out there here
34:04
would be the candidates copilot would be probably the market
34:07
share leader. Right now they're owned by
34:09
GitHub and and Microsoft. Not really available.
34:12
Claude code is Anthropic. I don't think the SpaceX wants
34:15
to work with Anthropic. They might, they could on a
34:18
licensed basis, but there seems like there's enough competitive
34:21
issue there that that's that's not going to be something that
34:23
Elon wants to do. And similarly for Open AI and
34:26
Windsurf there, it's a very high quality products, maybe better
34:29
in some in some ways than what cursor is, but again, not really
34:32
an independent player. So if you don't want to partner
34:35
with or work with Anthropic or Open AI, you can't really use or
34:39
don't want to use Copilot, which is owned by GitHub and
34:42
Microsoft, which is an open AI ally, then Cursor was the
34:46
logical place to go and to acquire that company.
34:49
Elon has been pretty forthcoming, actually.
34:51
Remarkable for a founder to go out and say, hey, the first time
34:53
that I built something, we didn't do it right.
34:56
So we have to rebuild from the ground up.
34:57
Most, you know, most founders wouldn't be this candid, but
35:00
Elon amazingly is. And he said the same thing about
35:02
Tesla. So he's trying to now rebuild
35:04
XAI from the ground up. And the Cursor team is
35:07
fantastic. What they do is, is terrific.
35:10
And they have an emphasis in coding that wasn't there in the
35:14
original instantiation for XAI. Second reason, I guess why
35:18
SpaceX wanted a cursor. I alluded to it.
35:20
I mean, they basically think AI is going to be helpful to go to
35:22
outer space. And that's what SpaceX wants to
35:24
do. Do you really need AI to go to
35:27
outer space? I mean, I don't know.
35:28
I guess the Apollo missions proved that you can get to outer
35:31
space without AI, but certainly if you want to go past the moon
35:34
and on to Mars, navigation, telemetry, orbital navigation,
35:39
just avoiding space debris, AI is is helpful, may at least
35:43
helpful and may be critical for doing all the above.
35:46
And then the last point to make is cursor right now is that is
35:49
that a $2 billion revenue run rate at a $60 billion
35:51
evaluation. They're being acquired at a 30X
35:54
revenue multiple. SpaceX right now trades at 100
35:57
multiple. And so I think it's a stupid
36:00
reason to buy something just 'cause you're worth 100X
36:02
revenue. And some this is just a the
36:05
finance guy and me like that's just a dumb reason to leak.
36:07
But a lot of financial analysts and some very smart people, I
36:10
heard Jason Calacanis say this over the weekend.
36:12
Jason Lemkin said this over the weekend. 2 very, very smart
36:15
people, even smart people say weird things sometimes.
36:18
Both of them are making the point that if you were valued
36:19
AT100X revenues, find something for 30X revenues is a creative
36:25
mathematically true, that's a stupid reason to buy something.
36:28
But the point is, I think yes, that you know, the financially
36:31
this is a creative for the company in advance of an IPO.
36:33
And so maybe there's some merit to that.
36:36
Why did Cursor want SpaceX? That's the other side of so
36:39
SpaceX. We know why they want to swipe
36:40
right. They need the technology.
36:42
It's maybe some financial accretion.
36:44
It differs vice the business strengthens the AI story.
36:47
What does Cursor get out of the deal?
36:49
Happy marriages are only successful if both partners are
36:51
happy. I think for Cursor, it's a no
36:53
brainer. They were about to run out of
36:55
money, which is why they're working on the $50 billion
36:58
evaluation and the new 5 to $10 billion round.
37:01
They were also running out of compute.
37:02
They cannot compete against the big hardware budgets of open AI
37:05
and Enthropic. Open AI has started all this
37:07
money now from SoftBank and Microsoft.
37:10
Anthropic just got money over the last week from Google and
37:14
Amazon. Both are raising prolific
37:16
amounts of monies from from investors.
37:18
And so Cursor was going to was going to have a very, very hard
37:21
time competing against those budgets.
37:23
It turns out that SpaceX, in addition to having a great
37:25
leader like Elon Musk and money and equity, also has something
37:29
called Colossus. If you don't know what Colossus
37:32
is, you should check it out. It's amazing.
37:33
It's a it's a massive AI supercomputer.
37:35
It's actually located in in headquarter in Memphis, TN.
37:38
They've got capabilities in Tennessee and and other S
37:41
southwestern states down in Texas.
37:43
It's basically over a million each 100 NVIDIA clusters.
37:47
And so you can see on this chart that they had basically once
37:50
they'd made the XAI acquisition and had launched the product at
37:53
the end of 2025, there was a lot of usage for rock open router
37:57
usage open way, way up at the end of 2025.
38:00
Then all of a sudden and they're giving away tokens at a very
38:03
discount priced encourage usage. Over the last year or so, usage
38:06
has dropped fairly significant leakage compared to the compute.
38:09
So there's a lot of excess compute capacity that SpaceX has
38:11
and Colossus and Cursor can have benefit from.
38:15
So it sort of solves the cursor problem with lack of compute and
38:18
the the lack of money in terms of the valuation on cursor and
38:22
some of the early investors, this chart shows from Sacra that
38:26
they were at a $2 billion revenue run rate.
38:28
Right now. They're they're planning on
38:30
tripling that by the end of the year.
38:31
I think with Spacex's compute and money, they'll be able to do
38:33
that. Some of the early investors, by
38:35
the way, who have made a lot of money with Cursor now Thrive
38:39
Capital along with Accel and Andreessen.
38:41
They were in at the Series C. That'll be a four or 5X multiple
38:45
for them. They were in at a slightly under
38:46
$10 billion evaluation. So really good investment
38:49
obviously and good credit to those fantastic investors.
38:52
And then check this out. Guess who led the Series B at
38:55
the early part of 2025? Thrive Capital and in recent, so
39:00
the same companies are leading and pricing subsequent rounds.
39:03
You don't often see that a company coming in at AB round
39:05
and then doubling down at AC round.
39:07
That's a real sign of conviction from AVC who knows the business
39:11
on that check Thrive and in recent we'll get a 30X multiple.
39:15
Guess who was in on the A round in 2024?
39:19
Thrive Capital and in recent Lev BA that that check will be over
39:22
100 X multiple. So they were in early, they were
39:25
in late stage. They were they were in late.
39:28
It's an impressive feat for a couple of VCs to be in early to
39:30
double down, to use their per router and follow on rights and
39:34
to take that company all the way to what is now going to be the
39:36
largest private a position by a venture back company of all
39:39
time. The only other one that that you
39:41
can say it's in the conversation is XAI, which wasn't really
39:44
venture back. It was a take private and then
39:46
had a lot of other banks and other debt.
39:48
And so this really is the largest private actors of all
39:50
time. Dave, you'll get a kick out of
39:51
this. There was one other there's one
39:53
other investor who was in very early.
39:56
Yeah, you recognize this guy? Sam Bankward, Freed, SPF.
40:00
Our old friend SPF Sam Brankenfig, Elevator Research Co
40:04
LED. Investors of all time, I would
40:07
argue. You say that like it's a joke
40:11
but he owns it's. Not a joke, it's not.
40:13
A joke. 5% of the company. He invested $200,000 to 5% of
40:17
the company. This could be a 300040005000X
40:21
multiple multi billion dollar win for elevator research.
40:26
Yeah. So it was a huge investor in
40:28
Entropic early which in work actually.
40:31
In Empropic he's a great investor, but Dave, guess what
40:34
has multiple on this investment is the is going to be.
40:38
Oh my gosh, I don't know, 500, something like that.
40:41
It is not going to be 5000 X. You know why 'cause
40:45
unfortunately for Sam, in his reputation while he's in jail.
40:48
Oh my. Gosh, you're kidding.
40:51
FTX went bankrupt and they did and so they had all the stuff on
40:55
their balance sheet and the liquidator did with the
40:58
liquidator it's supposed to do and they sold the stake for
41:01
basically 1X. So Sam Bankman, Freed and
41:04
Elevator Research will get a 1X return multiple on their on
41:08
their early stage seed stage Any sphere 2022 investment of
41:13
$200,000. That's unfortunate because it
41:16
was, it was actually a great bet, but Sam and his investors
41:19
at FTX are not going to be the beneficiaries.
41:22
They will get the money back in the liquidation and so anyone
41:24
who has bought that in the secondary market will be the
41:26
beneficiary, but maybe the most successful unsuccessful seed
41:30
stage investment of all time. One last point on SpaceX.
41:34
So this is just how now I think SpaceX is thinking about it.
41:37
So this is the PNL for SpaceX 2025 based on the financial,
41:42
SpaceX was about 18.6 billion in revenue. 12 of that came from
41:46
Starlink, 4 of that came from the launch business, and they
41:49
had another 3 from Twitter and Grok.
41:52
So XAI, Cursor, it's, it's not going to add to those revenues
41:57
and expect 2/20/2026 revenues. We'll see finally what their S1
42:01
looks like and what their projections look like.
42:03
But I think they're on track for around $20 billion from Starlink
42:06
in 2026, probably 9 to 12 billion on a launch business.
42:10
We'll break that one down in a second.
42:12
It looks like we'll get 3 to 4 from XAI and now 6 from Cursor.
42:15
So you add it up, it looks like SpaceX is now a $40 billion
42:19
revenue business that'll go on the road and or is on the road
42:22
right now. They'll go public later on this
42:23
year. They'll be targeting probably
42:25
$100 billion over the next three years.
42:27
How that $50 billion, you know it probably 50 will be from
42:31
Starlink, another 50 will be from other sources of what that
42:33
looks like, I don't really know. Some combination of the launch
42:36
business, XAI now Cursor and then defense with NASA and and
42:41
more of a more of a space based defense approach.
42:44
So that looks like the PNL for SpaceX and that looks pretty
42:47
exciting, right? Pretty compelling. 40 billion in
42:49
revenue growing at almost 100% year over year.
42:51
Now multiple revenue sources and and it's a more compelling IPO
42:56
story than it probably was, you know, two to three months ago.
42:59
In terms of the launch business, I was doing some math on whether
43:02
that 11 to $12 billion launch revenue line looks reasonable or
43:06
not. This is, this is kind of my map
43:09
of what it should look like just based on number of launches,
43:12
which ship they're using, the Falcon 9, the Falcon Heavy or
43:15
their new Starship. And then, you know, which
43:17
customers are going after and how they're going to price the
43:19
launch. I think 1112 billion, if I'm off
43:22
by, you know, a little bit, I might be off for a little bit.
43:24
Maybe they don't get to this full revenue picture in 2026.
43:26
Maybe it's more of a run rate by the end of the year.
43:29
But it looks like they're, they've got a launch business
43:31
that is settling into about a profitable $10 billion revenue
43:34
stream on top of a profitable Starlink business.
43:37
They're definitely losing money on XAI and on on CURSOR and that
43:41
that combination. But that's going to be a very
43:44
another very fast growing, call it $10 billion revenue stream
43:46
between those two pieces of the business, which will have the
43:50
ability to get to profitability in the next couple of years and
43:52
empowers the rest of their business.
43:53
So I think it's a great move actually for Cursor.
43:56
I think it's a pretty good move for SpaceX if you accept that
43:58
they need to be in AI to win. Cursor's filling in a very
44:01
important piece of the puzzle with this coding agent.
44:03
As a stand alone business, it it seems to be very financially
44:05
successful with a lot of very exciting customers.
44:08
So that's the rationale for the SpaceX Cursor acquisition, and
44:12
it looks like that'll that'll happen to the in the second-half
44:15
of this year as the IPO closes. All right.
44:18
Well, thanks a lot, Aman. And I guess we'll look forward
44:20
to seeing what happens after the IPO.
44:23
Absolutely. Thank you, Dave.
44:27
Did you folks know that more than half of all employee stock
44:30
options expire worthless? Every year?
44:33
Billions of dollars that are left on the table.
44:35
Opportunity for employees to realize some gains in their
44:38
efforts that they lose. We found a company that's
44:42
turning this into a business. It's a company called Equity B.
44:45
They've been on our show previously and we just LED a
44:48
safe round and Equity B, we put $2,000,000 into the company.
44:51
There's still another million that's available for our Co
44:53
investors that we're going to show you and give you an
44:56
opportunity to invest it. So again, just to remind people,
44:59
Equity B is a marketplace that connects startup employees who
45:01
can't afford to exercise their options and match them with
45:05
masters who can fund that exercise in exchange for a piece
45:08
of the upside. These are non recourse forward
45:10
contracts. The employee keeps the shares,
45:12
the investor gets paid at exit. Equity B has already done the
45:15
hard work of building this platform.
45:17
They're a FINRA registered broker dealer.
45:19
They're the only ones doing this at scale, and they've done
45:22
transactions with over 800 companies over 1000
45:25
transactions. They've deployed over 100
45:27
million, over $150 million in the last few years.
45:31
The opportunities they have are in companies like SpaceX and
45:33
Stripe and Databricks, and actually a whole bunch of other
45:36
companies that you might not have heard of before.
45:38
Unlike other secondary marketplaces, they don't sell
45:41
these shares. Equity B Finance is the
45:42
exercise. So it's kind of a different
45:44
game. It's not the same as Forge or
45:46
Hive or Equity Zed. And from investor perspective,
45:48
these are way better economics. We decided to invest because
45:52
this gave us a great opportunity to invest in secondary alongside
45:56
them, as well as invest in the company that's providing the
45:58
platform. Oren, the founder and CEO has
46:01
built a great company over the last few years.
46:03
They, they went through all the ups and downs of the Zerp era
46:06
and the nuclear winter. And now the company's really
46:08
been accelerating and growing over 100% the last couple of
46:11
years. And we think it's a great
46:12
investment and we'd like to share that with our Co
46:14
investors. So if you'd like to check out
46:16
more, you can go to the URL here that we're showing that's
46:20
practical vc.com/SPV slash equity B.
46:24
There you can learn a little bit more about the company.
46:27
You can download A1 pager that gives you some information on
46:29
how the company's doing. And if you'd like to learn more
46:32
about it, you can reach out to us and we'll be happy to share
46:34
more info. We're raising another
46:35
$1,000,000. We've already raised 2 million.
46:38
This SPV will be available till the end of May, and we hope you
46:41
can participate and get in on a great company that's going to
46:44
create a lot of value both for us as investors and for the
46:47
employees, investors that participate on their platform.
46:50
Thanks a lot, Gary.