Latest / In the Loop AI / SpaceX, OpenAI, and Anthropic IPOs: AI's Reckoning

Transcript

SpaceX, OpenAI, and Anthropic IPOs: AI's Reckoning · In the Loop AI

0:00
Josh: Three of the largest IPOs in history are filing within weeks of each other.
0:03
Josh: We have SpaceX, OpenAI, and as of just this week, Anthropic.
0:06
Josh: And on that same day, Google raised $80 billion of outside capital to fund its own AI buildout.
0:12
Josh: Now, the interesting part here is they're partly funded by each other's money.
0:15
Josh: Everyone is contributing to the other person's balance sheets.
0:18
Josh: And we've even gone so far as in the last few weeks to change the rules protecting
0:22
Josh: passive investors about how they can invest in IPOs.
0:26
Josh: The biggest buildout in the history of
0:29
Josh: capitalism is currently happening and we have
0:32
Josh: to ask the question is there enough money
0:35
Josh: like clearly there's a reason there's correlation as to why each
0:38
Josh: one of these companies are choosing to go public all around the same time and
0:41
Josh: this chart that we're sharing on screen here is pretty incredible i mean between
0:44
Josh: open ai anthropic and spacex the total amount raised during ipos is going to
0:49
Josh: be 180 billion dollars that's more than the entire dot-com bubble combined at 164 billion dollars.
0:56
Josh: That's three years versus three IPOs. So the scale of this is huge.
1:00
Josh: And I mean, we have to answer a few questions. Is this a circular economy moment?
1:04
Josh: Are they running out of money? Have they outgrown private capital?
1:07
Josh: There's a lot to talk about here, Ejaz, I guess, starting with SpaceX.
1:11
Ejaaz: Yeah, so you've got SpaceX open-air and Anthropik. They're having blockbuster IPOs
1:17
Ejaaz: The story isn't really about like each of them individually.
1:20
Ejaaz: It's the fact that they're all happening potentially within weeks of each other.
1:23
Ejaaz: They're all targeting like by latest a Q4 IPO and the combined raise is absolutely
1:29
Ejaaz: massive. We haven't seen anything like this before.
1:30
Ejaaz: So if we break down like what we've seen so far, we had SpaceX,
1:35
Ejaaz: I think it was April 1st that filed for
1:37
Ejaaz: their S1, which basically is their proclamation that they intend to IPO.
1:42
Ejaaz: I think they're targeting around an IPO sometime either this month or at least
1:45
Ejaaz: early July. That's what the rumor is. are saying right now.
1:48
Ejaaz: And then 10 days ago, we had OpenAI reportedly confidentially file their own S1 for their own IPO.
1:56
Ejaaz: And then 10 days later, aka yesterday, Anthropic filed for their confidential
2:02
Ejaaz: S1, which means that now all three of these companies are going for blockbuster
2:06
Ejaaz: IPOs within potentially weeks of each other.
2:08
Ejaaz: And so it begs the question, why is this happening now? Why such a rush?
2:13
Ejaaz: The answer to me is pretty simple.
2:16
Ejaaz: The AI CapEx build-out is becoming costlier than any of these companies could
2:20
Ejaaz: have expected, and they've decided to lean in.
2:23
Ejaaz: They've run out of free cash flow. Up until this point, all of these companies
2:27
Ejaaz: have spent private money, money that they've raised from investors or earned
2:31
Ejaaz: themselves through revenue.
2:33
Ejaaz: And now they're turning to the public and saying, we need more money to build
2:37
Ejaaz: out more data centers and GPUs so that we can train these models and facilitate
2:41
Ejaaz: all the demand that we're seeing.
2:42
Ejaaz: Now, if you talk to each of these companies, none of them say that there is no demand on their side.
2:48
Ejaaz: Google and I believe Amazon, Microsoft and Meta have all reported profitable
2:54
Ejaaz: quarters through all their AI CapEx expenditure.
2:56
Ejaaz: I think combined, actually, those four companies were aiming to spend around
2:59
Ejaaz: $1 trillion this year, but it's not enough.
3:01
Ejaaz: And so they need more money to kind of like fund all of this thing.
3:05
Ejaaz: The criticism that comes to me is like, I don't know whether this is the point
3:09
Ejaaz: of no return. We were chatting about this before we started recording,
3:12
Ejaaz: where there's no going back after this.
3:14
Ejaaz: We are jumping into the abyss, and whether we're going to lever up through debt
3:18
Ejaaz: instruments or whatever that might be, there is no returning from this. It is all in or nothing.
3:24
Josh: Well, and the thing I found most interesting is it's not just the companies that are going all in.
3:27
Josh: It's the institutions, and it's the public, and it is pretty much everyone who's going all in.
3:33
Josh: So much so that the largest funds on Wall Street that actually host these IPOs
3:37
Josh: are changing the rules just to accommodate them.
3:39
Josh: This most recent one happened around the SpaceX IPO, where index providers,
3:44
Josh: they waived their profitability requirements and cut the seasoning window from
3:48
Josh: 90 days to five. Basically, what that means is...
3:50
Josh: A company needed to prove that it was a viable company that could be traded
3:55
Josh: publicly before index funds were required to buy it.
3:59
Josh: Now, those index funds can buy it a lot sooner.
4:02
Josh: So a lot of people who might not want to be investing in SpaceX,
4:06
Josh: their retirement funds, their 401ks, the indexes that hold it are going to be
4:10
Josh: able to invest a lot faster than they normally otherwise would have been.
4:13
Ejaaz: Yeah, this has never happened before. So SpaceX is effectively changing the
4:17
Ejaaz: rules of the IPO market in order to facilitate the massive raise that they're going for.
4:22
Ejaaz: So a few sentences here that caught my eye. This forces over $30 trillion in
4:27
Ejaaz: passive 401k, so retirement fund money, to buy SpaceX.
4:31
Ejaaz: It's forcing them to buy it at IPO valuation. So it roughly means that around
4:35
Ejaaz: 24% of SpaceX's IPO supply of shares is going to be absorbed through these passive funds.
4:42
Ejaaz: We've never seen anything like that before. Secondly, typically,
4:45
Ejaaz: if you want to get included in indexes, prestigious indexes,
4:48
Ejaaz: such as the NASDAQ 100 or the Fortune 500, you need to, one,
4:52
Ejaaz: demonstrate that you're earning a ton of revenue, and two,
4:56
Ejaaz: demonstrate that for, I think it was like one or two successive quarters.
5:00
Ejaaz: So we're talking about like three to six months.
5:01
Ejaaz: All of those rules are being waived to days.
5:05
Ejaaz: We're talking like five to 15 days for two separate different indexes that are
5:08
Ejaaz: saying, okay, listen, if you can prove that the IPO price trades above a certain
5:12
Ejaaz: price for this amount, we'll include you the index.
5:14
Ejaaz: Don't worry, we'll forget about it. And we'll get all these pension funds to
5:18
Ejaaz: purchase or buy your stock.
5:19
Ejaaz: So Elon basically needs to keep the stock price up for above a certain level, and he's good to go.
5:25
Josh: Well, it's a little concerning because like for decades, these big indexes,
5:29
Josh: they've had two rules specifically designed to protect ordinary people who just
5:33
Josh: have investment accounts, retirement accounts.
5:35
Josh: And those two rules were like you mentioned, one of them was four straight quarters of profit.
5:39
Josh: And then the other was a minimum float of five to 10%. And actually those were
5:43
Josh: implemented because of the fallout from the dot-com crash.
5:47
Josh: In 1999 to 2000, indexes took a bunch of the super high growth cash losing companies
5:53
Josh: near the top. And that means that 401k holders or retirees, they were also holding
5:58
Josh: those companies and kept holding the bag.
6:00
Josh: So after they got crushed in the bubble burnout, what happened was is that they
6:04
Josh: implemented these rules. Now it feels like we're having deja vu.
6:07
Josh: The rules are getting reversed again. So now you don't actually need to have
6:11
Josh: four straight quarters of gap profit.
6:12
Josh: You just need to have, I believe it's 15 days.
6:16
Josh: So it's like a pretty considerable decrease in the parameters required to allow
6:20
Josh: these index funds to invest.
6:21
Josh: And it's like history doesn't to repeat, but it certainly does rhyme.
6:24
Josh: And it seems like the last time we did this, it wasn't good.
6:27
Josh: Now we're kind of reverting back to those terms.
6:29
Josh: That seems a little bit concerning to me, at least. So that's to be noted.
6:33
Josh: I think it's probably good for the price, right? It's like there's so much buying
6:36
Josh: pressure that's going to be pushing SpaceX up, but the downstream effects of it,
6:41
Josh: not going as well as planned are going to hurt some people who might not want
6:45
Josh: to be buying SpaceX shares in the first place?
6:48
Ejaaz: You know, Josh, I think the main criticism of the SpaceX IPO,
6:51
Ejaaz: and I think honestly, a lot of it is justifiable, is they just haven't proven
6:56
Ejaaz: the revenue model just yet.
6:57
Ejaaz: They're saying like, hey, we'll launch AI data centers in space.
7:00
Ejaaz: We'll do it on our rockets. But they haven't like proven that model,
7:03
Ejaaz: at least even like a proof of concept right now, like they're launching space rockets up there.
7:07
Ejaaz: But like, we don't have GPUs in space that are currently training Frontier models.
7:10
Ejaaz: So it's still kind of like, trust me, bro, in a sense.
7:13
Ejaaz: Now, one company that has proven
7:16
Ejaaz: a lot of revenue and has gone on the craziest story arc is Anthropic.
7:20
Ejaaz: And yesterday, they filed for a draft S1 registration statement of the SEC,
7:26
Ejaaz: which basically states that they plan to IPO in the coming months. Now,
7:31
Ejaaz: The story is very different with Anthropic in a few different ways.
7:35
Ejaaz: Josh, if you remember, at the end of last year, we did an episode, right?
7:38
Ejaaz: And we mentioned something about the CFO, Krishna Rao, of Anthropic.
7:42
Ejaaz: And he said, we have no immediate plans to IPO. We're taking our time.
7:48
Ejaaz: Back then, they had achieved $9 billion of annual recurring revenue.
7:51
Ejaaz: And they had estimated for the entirety of 2026 that they would hit something
7:56
Ejaaz: along the lines of $20 billion of ARR. They hit that in the first month and a half of 2026.
8:03
Ejaaz: And most recently, they just hit $45 billion of ARR.
8:08
Ejaaz: That's because of the success of Claude Code, Claude Cowork,
8:11
Ejaaz: and a host of other enterprise contracts, which they are signing.
8:14
Ejaaz: They're doing a bunch of different JVs. They're raising money from Blackstone.
8:16
Ejaaz: They're doing a ton of different things.
8:18
Ejaaz: So in contrast to SpaceX, they are actually earning a lot of money.
8:22
Ejaaz: So it makes sense that they now want to take this a step further.
8:25
Ejaaz: They're acquiring as much compute as they can. and they're in a race with mainly
8:29
Ejaaz: OpenAI to serve their frontier models, to train their frontier models and to
8:34
Ejaaz: provide it to as many people as they can.
8:36
Ejaaz: That's why I think they're going forward with this IPO. Now,
8:38
Ejaaz: it's important to state that there was no details released, was there?
8:41
Ejaaz: This is just kind of like a mandatory statement that they had to make.
8:44
Ejaaz: OpenAI, I don't think, had to
8:45
Ejaaz: do this, but Anthropic, in the efforts of transparency, decided to do it.
8:50
Ejaaz: I have a few reasons why I think they might be doing it, but I don't know.
8:53
Ejaaz: Do you have any thoughts? Yeah.
8:55
Josh: It's funny. It's a confidential disclosure of a confidential filing.
8:59
Josh: It's pretty ironic, but I think this took a lot of people by shock.
9:03
Josh: Everyone was kind of surprised at the rate. I was looking at Polymarket and
9:06
Josh: everyone expected OpenAI to IPO prior to Anthropoc.
9:09
Josh: So to see this news shocked it and it kind of flipped the margins on their heads
9:13
Josh: there in terms of the Polymarket.
9:14
Josh: But what I'm seeing here with the information post, and this actually looks
9:18
Josh: outdated because the rumors I've
9:19
Josh: been seeing is that Anthropoc is actually growing even faster than this.
9:22
Josh: And it's showing that Anthropoc just has this unbelievable growth trajectory
9:26
Josh: that has been driven by real value add, mostly on the enterprise side in creating
9:31
Josh: these like unbelievable models.
9:32
Josh: I know, I mean, when you think about Mythos now, they announced it,
9:35
Josh: what, two months ago, which means it finished training even before that.
9:39
Josh: And like, they just have these unbelievably powerful models.
9:42
Josh: I feel like they're feeling really confident going into this.
9:45
Josh: And it gets into the conversation about how much money actually is there for the market to absorb.
9:49
Josh: Is there actually a race to go out and collect capital
9:52
Josh: because we know spacex is going first the rumor is maybe around june 12th that's
9:57
Josh: coming in the next like two weeks so that's going to absorb god maybe a hundred
10:01
Josh: billion dollars of capital it looks like they're trying to raise 75 but i'm
10:04
Josh: sure it's going to exceed that how much money can anthropic then pull from that
10:07
Josh: and then if open ai is after them how much money is going to be left for them
10:11
Josh: and it feels like we're just,
10:12
Josh: demanding a lot from public markets and like we started the private markets
10:16
Josh: maybe they're tapped out, maybe they're not, but we're going to quickly learn how fast,
10:21
Josh: public market reserves can get tapped out because this is a tremendous amount
10:24
Josh: of capital everyone's raising.
10:25
Ejaaz: I mean, the way I think about it, right, is you have a subset of investors that
10:30
Ejaaz: are purchasing or will purchase the stocks of these different companies for
10:34
Ejaaz: one particular thesis, right, which is like, I'm bullish on AI.
10:38
Ejaaz: But then you have a whole retail subsection that are like, I just use Claude
10:41
Ejaaz: every day, and it's super helpful to me.
10:43
Ejaaz: So obviously, I'm going to buy that thing, right? I use ChatGPT every day, right?
10:46
Ejaaz: So there's these two subsects, and they're all going to result in the same thing,
10:49
Ejaaz: which is like, we're going to purchase stock of these companies.
10:53
Ejaaz: Now, if you're on the company side, you're using it for one thing.
10:56
Ejaaz: Krishna Rao, CFO of Anthropik, has said it.
10:58
Ejaaz: Sarah Fryer, CFO of OpenAI, has said it. And Elon has said it.
11:02
Ejaaz: We just need more compute because more compute translates into better AIs,
11:05
Ejaaz: which translates into better products, which means that we can serve more customers,
11:08
Ejaaz: which means that we can earn more money.
11:09
Ejaaz: Now, in Anthropik's case particularly, I think the rumors around their AGI-like
11:15
Ejaaz: model called Mythos is very real.
11:17
Ejaaz: We've got like breaking news from today that I have up on the screen that they're
11:20
Ejaaz: rolling out their Project Glasswing, which is kind of like the bucketed sandboxed
11:25
Ejaaz: version of release for Claude Mythos to 150 additional organizations across the entire world.
11:31
Ejaaz: And they even said on a previous statement recently that they're going to be
11:34
Ejaaz: releasing it publicly to people over the next couple of weeks.
11:37
Ejaaz: So I think this all comes at a very coincidental, intentionally coincidental time.
11:42
Ejaaz: And the other thing that's different with Anthropik is they're estimated to
11:45
Ejaaz: be profitable by the end of this month to the tune of $550 million, which, listen,
11:51
Ejaaz: is a drop in the ocean compared to like the trillions of dollars that are being
11:54
Ejaaz: spent on CapEx, but they'll be the first major AI lab to do so.
11:58
Ejaaz: And so it's just this massive run rate. I think of all these IPOs,
12:01
Ejaaz: I'm probably most bullish on the anthropic thing, but they're all different
12:05
Ejaaz: to their own kind of lane.
12:07
Josh: Now, when we were looking at our bubble meter in a few episodes ago,
12:11
Josh: and we were kind of figuring out at what point do things get scary?
12:14
Josh: It was largely around the idea that these large mega companies like Google will
12:20
Josh: start to spend more than their revenues allow them to.
12:22
Josh: Basically, they'll start going into debt to fund these build outs.
12:24
Josh: And we have some indication that we're kind of slowly turning into that territory
12:29
Josh: where Google has ran out of revenue on their balance sheet and are now looking
12:34
Josh: to raise external capital.
12:35
Josh: This is not IPO related as you know Google has been
12:38
Josh: public for a very long time but they need more money so
12:41
Josh: what did they do they went out and raised 80 billion dollars to fund the AI
12:47
Josh: build out that is a tremendous amount of capital I can't remember the total
12:50
Josh: capex spend that they guaranteed but I imagine it's close to like 30 40 percent
12:54
Josh: of the total that they were planning to spend annually and they just went and raised it noteworthy,
12:59
Josh: Berkshire Hathaway, Warren Buffett's old company, wrote a $10 billion check
13:03
Josh: to get in. And it's a pretty huge deal.
13:06
Josh: It's $30 billion in underwritten public offerings, $40 billion through an at-the-market
13:11
Josh: stock program starting in the third quarter of this year, and then $10 billion
13:14
Josh: through private placement to Berkshire, which we just talked about. So it's a huge deal.
13:18
Josh: And he's just I'm curious your take on this, actually, because we did a deep
13:23
Josh: dive into Google's balance sheet.
13:24
Josh: We saw how much money they made. We saw how much they were spending. They were in the green.
13:28
Josh: Now it looks like they're intending to. Are they going to go in the red or are
13:31
Josh: they just offsetting? Are they softening the blow so that they still have a
13:34
Josh: lot of cushion just in case?
13:36
Ejaaz: Yeah, I think they're leaning all in and they'll eventually have books in the red.
13:41
Ejaaz: Larry and Sergey Brin, the founders of Google, said very explicitly,
13:48
Ejaaz: I think it was a year and a half ago, that they are willing to lose it all versus
13:52
Ejaaz: like lose the actual AI race.
13:54
Ejaaz: So they're going to spend, spend, spend until they create enough compute.
13:57
Ejaaz: It's founder mode. They've gone back to founder mode.
13:59
Ejaaz: Sergey Brin came back to Google to do exactly this, to kind of lock in.
14:03
Ejaaz: This is my favorite IPO story of the week. And it's not even an IPO,
14:08
Ejaaz: right? So you have Google, public company, raising $80 billion.
14:11
Ejaaz: Now, the question becomes, what are they using that $80 billion for?
14:14
Ejaaz: Now, the headline is, we're using this $80 billion to build out more AI CapEx.
14:18
Ejaaz: We're going to build more TPUs. We're
14:22
Ejaaz: A story that a lot of people missed is, I think it's like 30 billion of this
14:27
Ejaaz: 80 billion is being used to pay off tax obligations for their employees who
14:32
Ejaaz: are cashing out stock over the next couple of months.
14:34
Ejaaz: All right. So like, I just want to be clear here, like, a large chunk of this
14:39
Ejaaz: raise isn't actually going to be used for the AI cap expand.
14:42
Ejaaz: Now, putting that aside, Google isn't a bad actor, in my opinion.
14:48
Ejaaz: They've been pretty transparent as they can be with how much they're spending
14:51
Ejaaz: on all this AI stuff and what they intend to do with all of this.
14:54
Ejaaz: And they're giving it their best shot.
14:55
Ejaaz: Now, what this gives me reminiscence of is, Josh, do you remember at the end
14:59
Ejaaz: of last year, we were kind of saying how OpenAI is kind of distracted.
15:03
Ejaaz: They're kind of focusing on like random AI products and they're like missing
15:08
Ejaaz: the whole coding AI thing and all that kind of stuff.
15:11
Ejaaz: And then they decided to call a code red and lock in. I feel like Google has
15:16
Ejaaz: drifted into that kind of like broad spectrum of things.
15:19
Ejaaz: They're not really focused. They're like building out an agent,
15:22
Ejaaz: but then they're trying to build out a better coding model.
15:24
Ejaaz: They're trying to build out a better generalized LLM, but then they're building
15:27
Ejaaz: out the infrastructure TPUs, but then they're selling those TPUs to competitors,
15:30
Ejaaz: which means that they don't have enough compute to train their own Gemini models,
15:33
Ejaaz: which means that Gemini falls behind, and it has fallen behind.
15:35
Ejaaz: Their recent Gemini 3.5 flash is behind Frontier models despite all this money
15:39
Ejaaz: that they had, and now they need to raise all this money to train a better model.
15:43
Ejaaz: It just seems like they just need to lock in and focus.
15:46
Ejaaz: Now, when I look at this structure, $80 billion, that's a lot of money.
15:50
Ejaaz: It's effectively their own IPO to be able to fund their own build out.
15:54
Ejaaz: I'm not convinced that spending $30 billion of that to write off tax obligations
15:58
Ejaaz: is the best way to do it. It kind of wreaks a little bit of desperation.
16:01
Ejaaz: But I am optimistic because the last four times that we've had this kind of
16:07
Ejaaz: public raise from a major company, Berkshire Hathaway has put in a $10 billion check.
16:12
Ejaaz: And all of those companies have done pretty well. So I'm hoping the same thing
16:15
Ejaaz: is going to happen with Google. But it's just an interesting story.
16:18
Josh: Yeah. I mean, in Berkshire, we trust, right? Like they have been right in the
16:21
Josh: past. They are very disciplined.
16:22
Josh: We're going to hope they continue. It's also nowhere that Google owns a large
16:26
Josh: percentage of a lot of these companies going public.
16:29
Josh: I mean, Google is one of the largest private stakeholders in SpaceX and also Anthropic.
16:34
Josh: So they stand to benefit from the upside of this in a pretty considerable way.
16:38
Josh: It's also a little bit scary how large these numbers are getting.
16:41
Josh: It feels like we're getting numb to the hundreds of billions of dollars number.
16:46
Josh: I mean, Google at $180 to $190 billion in CapEx this year is unbelievable to
16:52
Josh: imagine just a few years ago.
16:53
Josh: So for them to go all in, it's all in at a scale that we have never seen before.
16:57
Josh: And I think that's one of the themes of this episode is we're really experiencing,
17:01
Josh: like this is a very special time in history because never in the history of
17:04
Josh: this country and capitalism has this amount of money and value been concentrated
17:10
Josh: on pursuing a single idea.
17:12
Josh: And the outcomes of that are going to be pretty big.
17:16
Josh: I mean, bubble or not, we're building real value, real intelligence being built.
17:19
Josh: And I think like the actual underlying civilizational shift is going to start
17:24
Josh: to be felt as the outcome of all the spending starts to make its way into the market.
17:29
Josh: And I think it's like it's easy to get lost in the numbers that are huge and large.
17:33
Josh: It actually is going to result in real valuable tools. And we make fun of Google.
17:37
Josh: They haven't come out with a new frontier model in a while.
17:39
Josh: But I mean, I'm just using their tools and services now. And they're noticeably
17:43
Josh: getting smarter and smarter and smarter.
17:44
Josh: One thing I'm excited about is WWDC next week. We're going to be covering that.
17:47
Josh: We'll see how Apple plans to actually roll out these tools and make them smarter
17:50
Josh: and smarter. So we'll see where all the CapEx has been going.
17:52
Josh: But there is one more IPO that's coming that we haven't mentioned, which is OpenAI.
17:58
Ejaaz: Yeah, exactly. So we mentioned it earlier, but 10 days before Anthropic Fall
18:02
Ejaaz: for that, potential IPO. OpenAI did the very same thing.
18:06
Ejaaz: Confidentially confidential filing, by the way. And so it was reported...
18:10
Ejaaz: They get this one confidential.
18:10
Josh: Even though it leaks. It's like, if it's all going to leak, you might as well
18:13
Josh: just publish it yourself.
18:15
Ejaaz: Yeah. What was curious is the polymarket market on this particular IPO spiked
18:20
Ejaaz: up before the Bloomberg and Financial Times report.
18:22
Ejaaz: So there was definitely some insider stuff going on. But basically,
18:25
Ejaaz: Goldman Sachs and Morgan Stanley have reportedly been behind the scenes helping
18:29
Ejaaz: OpenAI and prepare for their IPO.
18:31
Ejaaz: If you're wondering why they're doing it, it's for all the same reasons.
18:33
Ejaaz: They wanna raise more money to build out a bunch of different data centers.
18:37
Ejaaz: I think they broke ground on a new data center literally a few days ago, so it's all in.
18:42
Ejaaz: A comment that I have is, and maybe this is like a potentially hot take, Josh, is,
18:48
Ejaaz: I think all of this money that is being raised and is about to be spent is ultimately
18:53
Ejaaz: going to result in a good thing.
18:56
Ejaaz: I think that it's ultimately not going to be bubbly in its nature,
19:00
Ejaaz: and it's going to result in the necessary infrastructure that is going to be
19:03
Ejaaz: onshore in the West or in the U.S.
19:05
Ejaaz: Specifically to create that next bedrock of foundation for technological innovation. You need compute.
19:12
Ejaaz: You need the electrical lines. You need all the substrates to make these GPUs
19:16
Ejaaz: and silicon chips actually work to serve it to customers. You need all of this.
19:20
Ejaaz: And we are currently constrained by physical atoms.
19:24
Ejaaz: If that sounds like a vague statement, it is not. I literally mean that.
19:27
Ejaaz: I sit on the Gavin Baker's camp of opinion here, which is no matter how much
19:31
Ejaaz: money you want to lever up and spend,
19:33
Ejaaz: you can't actually spend the money because you are limited by how slow it is
19:38
Ejaaz: to overcome regulations and build the actual, you know, brick and mortar of
19:43
Ejaaz: the data centers and then create the silicon chips. There's only one ASML.
19:47
Ejaaz: There's only one NVIDIA. There's only one TSMC. And it's hard to scale the physical
19:51
Ejaaz: infrastructure side of AI.
19:53
Ejaaz: So even if you want to lever up and create all these like weird debt structurings,
19:57
Ejaaz: it does not matter because there is nowhere to spend it on. You're limited by physical atoms.
20:02
Ejaaz: And until I see that block get unblocked we're not in a bubble
20:06
Josh: This is my take yeah i mean to talk about gavin
20:08
Josh: baker's point again he always mentions the idea of dark fiber comparing it to
20:11
Josh: the internet bubble era where there was a tremendous amount of fiber optics
20:15
Josh: laid for the internet but there simply wasn't enough use cases for the internet
20:19
Josh: to go through the fibers there was a lot of dark fiber a lot of the build-out
20:23
Josh: didn't get used it didn't generate revenue everything fell apart in this case,
20:27
Josh: I mean, we talk about this all the time, but the GPUs from four or five years
20:29
Josh: ago are being rented for more money than they were four or five years ago.
20:33
Josh: Yeah, they're worth more.
20:34
Josh: They're worth more. And it's funny, like Michael Burry, the big short guy,
20:37
Josh: he was like, no, no, no, that's wrong. Everything's going to crash.
20:40
Josh: He could not have been more incorrect.
20:41
Josh: So, so far, all of the guidance, all the signals that we're seeing are green.
20:46
Josh: They're all positive. Everything looks good. It seems like CapEx is actually returning real value.
20:51
Josh: I mean, there was those rumors about Anthropic maybe turning up profits sometime
20:55
Josh: soon, as you mentioned earlier.
20:55
Josh: That's a really huge deal because if they could actually
20:58
Josh: absorb all of this um fundraising money
21:02
Josh: and then spend it in a way that's capital effective enough where
21:05
Josh: they can return revenue that's like huge and open ai i
21:07
Josh: don't believe they were planning on doing that for at least till the end of
21:10
Josh: this year so all these signs are green the one thing i'm looking out for personally
21:15
Josh: is companies who are cutting back on spending i know there's some news in the
21:19
Josh: headlines about a half a trillion dollar bill that came up for amazon that was
21:23
Josh: unintentional i don't know how true these things are but.
21:26
Ejaaz: I think so you know
21:27
Josh: Yeah so so long as companies are still extracting value from these ai systems
21:32
Josh: then they'll continue to spend because it will just increase the revenue and
21:37
Josh: increase the margins in their own business and so far so good fingers crossed it used to go that way did.
21:43
Ejaaz: You see the uh the crazy stat that christian rao
21:46
Ejaaz: dropped on i think it was like the invest like the best podcast
21:49
Ejaaz: um anthropic for the fortune 10 so like 9 out of 10 of the fortune 10 so the
21:55
Ejaaz: top 10 companies in the world basically use anthropic thought code specifically
21:59
Ejaaz: and their net dollar retention rate which is basically like the budget that
22:04
Ejaaz: they started with january 1st versus how much they estimate to spend by the end of the year
22:09
Ejaaz: is increased by 500%. So they're planning to spend 5x more.
22:13
Ejaaz: But the reason why they're doing that is not because they have to,
22:16
Ejaaz: it's because they want to, because they're getting so much ROI on the back end.
22:20
Ejaaz: So the point is, if this continues to trend in the right way,
22:25
Ejaaz: the definition of a bubble is it's over-levered, and there's not enough demand on the buy side.
22:30
Ejaaz: We're seeing the opposite happen, and we actually can't supply enough of the
22:33
Ejaaz: silicon and compute to serve the buy side. That's why Google's raising $80 billion.
22:37
Ejaaz: That's why all these companies are IPO-ing at massive valuations.
22:40
Ejaaz: It's so that they can serve that.
22:41
Ejaaz: And maybe I'm drinking my own Kool-Aid, but that's my take.
22:44
Josh: Well, and also think about the downstream effects of that. They're huge.
22:46
Josh: Now we have all this infrastructure that we're building out in the United States.
22:49
Josh: We have the ability to build it out.
22:51
Josh: And soon that's going to transition away from software as well.
22:54
Josh: Like software will continue to be important, but I just saw a post from OpenAI
22:57
Josh: recently. They're hiring for a robotics division. They're going to start building robots.
23:01
Josh: SpaceX is building their Optimus robot. We're going to get a reveal probably
23:04
Josh: around the time of the IPO, I would assume to gather a little more attention.
23:09
Josh: And I think what we're seeing here is this like really huge shift of industrializing
23:13
Josh: the United States in a big way.
23:15
Josh: Like we're going to get satellites into space. We're going to send data centers into orbit.
23:19
Josh: This takes a huge amount of capital, but think about the value it delivers.
23:22
Josh: Think about the value of cloning the internet, but putting it into low earth orbit.
23:26
Josh: So that way it's, it's nation state proof. It can never go offline. You never have outages.
23:30
Josh: It's just this unbelievable technology that we're building funded by
23:34
Josh: this capex build out this huge spending and now the public has the opportunity
23:38
Josh: to actually get involved so is it going to be good or bad for the public we
23:41
Josh: will see i'm feeling pretty bullish i think this is where i want to end the
23:44
Josh: show actually ejaz is asking you which of these three if any are you most excited
23:49
Josh: about and most likely to invest in at the ipo.
23:53
Ejaaz: Oh, damn, I'm excited about all of them, dude. Okay, if I had to rank them,
23:58
Ejaaz: it's Anthropic, SpaceX, and OpenAI, but like it's within like a hair's distance
24:03
Ejaaz: between each other. It's so tough.
24:05
Ejaaz: They're all building amazing things. And I do truly believe that these three
24:09
Ejaaz: companies will create products and services that will become the bedrock of
24:14
Ejaaz: everything and anything that we build future businesses on.
24:17
Ejaaz: And so the question is, how much would you value that kind of a company, right?
24:20
Ejaaz: We've never seen anything like this before, right? We've never seen a type of
24:23
Ejaaz: technological disruption which permeates every single industry that you could
24:28
Ejaaz: possibly think of, including the hardware side of things, right?
24:31
Ejaaz: What do you think happens when robotics scales up?
24:33
Ejaaz: You're going to need a robotics model. You're going to need data to do that.
24:36
Ejaaz: That's what these AR labs are going to do.
24:38
Ejaaz: And you're going to need infinite compute, potentially in space.
24:42
Ejaaz: You're going to need really smart models, potentially trained on RL from Anthropic
24:46
Ejaaz: or from OpenAI that is able to do this.
24:49
Ejaaz: And these three companies are the clear bet. So if I take a long-term perspective,
24:52
Ejaaz: which is typically how I invest, I'm going to buy an IPO and we'll see what happens, you know?
24:58
Josh: Yeah, I'm pretty stoked for all of them. I think for me, SpaceX is most interesting.
25:02
Josh: I've been a fan of them for as long as I can remember.
25:05
Josh: They are just such a badass company. How long have you been tracking SpaceX.
25:08
Ejaaz: Josh? You've got to flex that. Come on.
25:10
Josh: Well over a decade, probably close to 15 years.
25:12
Ejaaz: That's it.
25:13
Josh: Like it was as the beginning of the Falcon 9 program started because I remember
25:18
Josh: watching all the live streams. And in fact, my YouTube channel is like,
25:21
Josh: I think, 14, 15 years old or something.
25:23
Josh: The first video I ever posted was like a handheld camera of my screen of the Falcon 9 launch.
25:29
Josh: It was like, it's pretty amazing because it's been so cool to watch them go
25:32
Josh: from the Falcon 1, which is a single rocket thruster to 9 to now Starship,
25:37
Josh: which has like between the booster and the first second stage, like.
25:40
Josh: 30, 940 engines. It's like unbelievable how much progress they've made.
25:45
Josh: And it's so exciting because it opens up this awe-inspiring opportunity and
25:49
Josh: having the opportunity to invest in that company to participate is really exciting to me.
25:53
Josh: I know people are upset that it's trading so high on a relative basis,
25:57
Josh: but when you consider what the future looks like in the case they succeed,
26:00
Josh: and when you consider what the team requirements need to look like,
26:04
Josh: there's really no better chance.
26:07
Josh: You can't assemble a better team.
26:08
Josh: You can't assemble a better company than this to take a shot on goal.
26:12
Josh: So it's at least worth getting excited about and supporting that shot on goal,
26:16
Josh: that like civilization is going to do something we've never done before in a
26:19
Josh: really cool and exciting way that stands to benefit everyone else.
26:21
Josh: And that's why I'm excited about SpaceX.
26:22
Josh: I will try to participate. Hopefully it doesn't launch at like a $4 trillion valuation.
26:27
Josh: Hopefully it stays kind of close to the share price it launches at.
26:30
Josh: I think that's another thing we'll have to look out for is like how high the
26:33
Josh: premiums actually are once they start trading, because we know there's going to be a ton of hype.
26:36
Josh: But yeah, I think that's just about everything for this episode. Any final thoughts?
26:40
Ejaaz: Nope, that's pretty much it. I just want to remind the audience that's listening
26:44
Ejaaz: to this that Josh and I are delusionally optimistic about a lot of these different things.
26:49
Ejaaz: We do try to ground ourselves occasionally, but overall, we take an optimistic
26:52
Ejaaz: approach to AI and the frontier of tech.
26:54
Ejaaz: So none of this is investment advice. But if you did enjoy this episode,
26:58
Ejaaz: or if there's anything that you actually disagree with, please let us know in
27:01
Ejaaz: the comments. You guys are very vocal.
27:03
Ejaaz: Wherever you listen to us, if it's on YouTube, if it's on Spotify,
27:05
Ejaaz: if it's on Apple Music or wherever, please subscribe, give us a rating,
27:08
Ejaaz: leave us a review, leave us a comment.
27:10
Ejaaz: We want to hear from you guys about what you guys think. Are you enjoying these kinds of episodes?
27:14
Ejaaz: Are there other topics that we could cover that you might be more interested in hearing more about?
27:18
Ejaaz: Let us know. But Josh, do you have any parting thoughts?
27:21
Josh: That's it. Share this with a friend if you enjoyed. Let them know what they
27:24
Josh: got to know about these IPOs that are incoming.
27:27
Josh: Let us know which ones you're going to be participating in. And as always,
27:30
Josh: we will see you guys in the next episode. Thank you guys so much for watching.
27:32
Ejaaz: See you guys.