SpaceX IPO, Anthropic Fable 5, And Roku | The Brainstorm EP 136

**Sam** (0:00)
SpaceX is finally public, fabled, released, unreleased. Elon saying one trillion in revenue in 2030 Brett, your estimates are about a third of that. Maybe we'll start there, dive in.
What are your thoughts on SpaceX's IPO here and Elon's most recent tweet about one trillion in revenue in 2030 and much more likely in 2031?

**Brett** (0:28)
Yeah, I think that one, the IPO was successful as in they sold a lot of shares to the public and seemed to have listed at something above the IPO price without going totally nuts, which is I think where you want to hit.
There's plenty of shares to unlock over the next 180 days and usually listings are volatile. This one could be particularly so because on a trailing financials basis, this company looks wildly extraordinarily expensive.
Even if you take into account the deals they signed for their data center as a service business with Anthropic and Google, it still looks quite expensive. But if you look forward and say, well, what are they going to do over the next few years, you can understand why it's valued so highly. They're going to take lots of cash, invest in satellites, principally Starlink satellites, loft them into orbit and monetize them. And we think that injecting cash into that particular business is the most efficient use of capital for them in the immediate term. And mostly on the strength of that and with the AI business basically like burning and churning cash, you get to $300 to $400 billion in revenue by 2030 and growing fairly quickly. And Elon says a trillion. So how do you square that circle?
It's pretty clear to me that actually if you inject even more cash into the business, they could accelerate even faster. It could be that we're overestimating kind of like the cost to manufacture the satellites, or he thinks that they can improve their manufacturing and the capital efficiency of the business more than we're expecting. Or it could be that he's thinking that, hey, if we actually merge with Tesla here, they are going to be throwing a lot of cash flow off of Robotaxis, and we can invest that cash flow in accelerating Starlink, but particularly AI orbital data centers in space. That would actually bring on a very rough basis some of his forecast into the realm of possibility as far as I can tell. So, I think that's probably a way to explain it. Particularly, if you're saying a trillion dollars, then it's a trillion dollars of SpaceX. Plus, by the way, there's this Tesla business that's also generating a lot of revenue, then they could easily get there.

**Sam** (2:52)
Okay, can you maybe break down, SpaceX needs money to accelerate here.
Does Tesla not need money to accelerate its RoboTaxi rollout? RoboTaxi has not yet fully scaled here. Why can you throw money at the SpaceX flywheel but not necessarily at the RoboTaxi flywheel?

**Brett** (3:15)
Yeah, I mean, the RoboTaxi scale out should be much less capital intent than SpaceX because the assets have already been built and the assets are being built and basically financed by the customers. So Tesla's Model Y in Austin factory, or across the globally Model Y, they can do like 2 million, remind me here Sam, in production.
So you can transform those into RoboTaxis and they become wildly cash flow generative.
You do need to invest in some charging infrastructure and servicing infrastructure as the model scales, but effectively, I think it would be almost immediately cash flow positive in a very, very meaningful way. Whereas with Starlink and AI satellites, you have to manufacture the satellites, you have to package them in a rocket that you have built, you have to have multiple rocket sites that are going, and then you loft them up, and then the time to manufacture and all of that is like a big cash drag. So the very rough way to think about it is like the 10th Starship rocket launch will cost a half billion dollars cash up front, and it will generate more than a billion dollars in revenue per our forecast. So six month return, but if you raise $75 billion, you can do 150 of those, and then you have to collect the cash and re-inject it, and there's other places they're also deploying cash, including building data centers. So it is a business where you could inject more cash and accelerate. I think Tesla could use cash for Optimus, but I don't think it's cash limited, I think it's just capability limited. You have to have time to work on the software, improve the hardware, throwing $10 billion more at Optimus, would not make the future come faster there, and so then that's why SpaceX isn't, not just SpaceX, a lot of companies, I think are in a spot where suddenly it's like, do you know what, we have a great return on capital opportunity here, we could use more cash. That's why Google came to market and raised equity cash. People were like, what are you doing? It's because investing in data centers is a great business right now, and SpaceX can invest in Starlink satellites and data centers and AI satellites in space and build all the solar panels they need for the AI satellites in space. So there's a lot of like investment they have to put in to accelerate.