Welcome to the podcast.
I'm your host, Jaden Shafer.
Today on the show I want to talk about the latest news from SpaceX filing some paperwork with the Federal Communications Commission last Friday in what looks like a million satellite data center network right.
So they have filed paperwork to request to be able to shoot a million data center satellites up orbiting the earth.
And at first, I think a lot of people said this is like, oh, you know, this is kind of just Elon.
This is like a theatrical idea that he's just kind of floating out there because You know it makes it sound good to merge SpaceX and XAI and make this 125 trillion company before an IPO.
I think if you look at what's happening, especially as of today this is like a week later I think it is pretty clear that he actually means it.
So today on the show, I want to break down what this would look like, what a AI data center cluster floating around the earth of a million satellites would look like, why they would do it logistically, how this would work.
There is a lot to unpack here, including some really interesting quotes from Elon Musk on a recent podcast about using solar panels and how much more efficient they are in space.
All these different crazy things.
So we're going to dive into all of that.
Before we do, I wanted to share some really exciting news about my startup, AIbox.ai.
We've recently launched some new subscription tiers.
So AIbox essentially allows you to build tools.
Even if you're not a developer, you can describe what you'd like to create and And our AI model links together different AI models, fills in a prompt and it can create a tool that automates your workflows, or just you know, is it generating newsletters or all sorts of things?
People are using it for creating logos and you know anything you use AI for.
But you can link multiple AI models together.
We've just launched a brand new 9 a month tier that gives you access to over 40 of the top AI models in one place.
So imagine, I mean, it looks basically like ChatGPT, except you can switch between talking with ChatGPT and Anthropic and Google, Gemini and Meta's AI model and Perplexity and Quen and XAI.
You also get a bunch of really cool image models like Ideogram and Black Forest Labs and, of course, 11 Labs for audio generation.
So for 9 a month, you get access to all of the top AI models in one place without having to have subscriptions to all of them and without having to log into 10 different accounts.
Juggle all your passwords.
It's all in one place for you.
And we even have a media storage that gives you access to all of the images that you've generated, all of the files you generate.
They're all in a media storage, so you don't have to try to remember what website you're on or what account you're logged into.
It's all in one place.
So if you want to go check that out, it is AIbox.ai.
I'll leave a link to it in the description.
Okay.
So with everything happening with SpaceX and this kind of merger, I think one of the biggest signals we got was on Monday, SpaceX officially merged with XAI and that deal basically is formally tying these two companies together XAI and, of course, the AI Lab.
I think this only really makes sense if both of them are trying to share some sort of infrastructure.
Otherwise, why would you merge them?
And so I think this is why a lot of people were saying oh, this is just you know.
And then Elon comes up with this idea to put data centers in space.
And then people are like oh, this is just him trying to like, consolidate his companies for a bigger IPO.
Um, in reality, i think this might be an actual really, really smart move for both of his companies.
Um, but i think what next happened was that there was an fcc filing and if you look at that fcc filing i've been, you know, reviewing some of the documents i think elon musk is very serious about putting some large-scale ai com compute in orbit.
Um, this is getting a step closer to reality.
On Wednesday, uh, the FCC accepted formally SpaceX's filing and they opened up for public comment.
So I think normally that step is going to attract like not a lot of attention.
Not a lot of people are, you know, whatever.
They're just, this is just kind of a formality.
They're like, Hey, we're opening up for comment.
We've accepted this thing.
I mean, this is just what the FCC does this time right now.
I think the FCC chairman, which is Brendan Carr, publicly shared the filing on X.
Uh, Carr has basically said that he's really willing to support this move right now, as long as Musk is, you know, trying to build things inside of America.
He said that it's not likely to face a lot of regulatory pushback.
Musk also said like so.
Then Musk went on to a podcast which is the Cheeky Pint podcast, which was with Dworkish Patel.
And then also John Collision, who is a co-founder of Stripe.
But anyways, he's breaking it down and explaining it a little bit more.
But he said, quote it's harder to scale on the ground than it is to scale in space, which I think is just a wild quote from this conversation.
One thing he said that I think has a lot of credence is, he said a solar panel gives you about five times more power in space than on Earth.
So it ends up being much cheaper, which is a really fascinating concept, with these data centers, floating satellite data centers around the Earth that don't need to be cooled because there is you know, it's freezing cold in space.
And so that saves a lot of like money on the water that you'd have to use.
It's also more environmentally conscious to not have to use all of this water.
So for people that that's their number one priority, I think this is an interesting option.
And then of course the solar panels can power them.
So they're basically their own operating units independent up there.
I think electricity isn't the only cost for running a data center.
Patel pointed this out during this kind of conversation they had.
But he also brought some really interesting concerns, which is what happens when hardware fails during AI training.
Like, how do you fix it in orbit?
This is not exactly straightforward.
Elon said that He predicted by 2028, that's going to be kind of the big turning point.
He said you can mark my words.
In about 30 to 36 months, the most cost effective place to run AI will be space.
He then went on to say that within five years, more AI compute power is going to be launched and run in orbit each year than the total amount operating on Earth.
I think, for like context.
Global data center capacity right now is expected to reach about 200 gigawatts by 2030.
And that's, you know, close to a trillion dollars in kind of ground based infrastructure.
So if Elon Musk is predicting that, he's effectively I mean he's basically arguing that there's going to be a massive chunk of that future investment that is going to not even be on earth.
Maybe, you know, half a trillion dollars a year is going to be spent on on data centers in space.
I think obviously this is a vision that fits really nicely into SpaceX's business model, right?
They make money by launching things into space.
And so now, if they own an AI company that needs massive amounts of compute power, launching things into space is kind of the best way for them to go about doing that.
But according to Bloomberg, this whole SpaceX-XAI merger values them at about 125 trillion, and it makes it the most valuable private company in the world.
SpaceX has reportedly been preparing for an IPO, which might happen in June, but Elon Musk didn't say whether the merger is going to change those plans in any way.
I think it is interesting that on this podcast recently, when he was talking about all this, he was there with the co-founder of Stripe, which is one of the other biggest privately held companies which a lot of rumors about when they might IPO are circulating out there.
Something that's really interesting is there was a memo that was kind of announcing this deal.
And Elon Musk basically framed the merger almost entirely around this kind of space-based data center.
And that's basically why he said they needed to merge.
And this is a quote from that memo.
He said Global electricity demand for AI cannot be met on Earth alone without putting pressure on communities and the environment.
And that's basically his argument, which of course is a little bit complicated.
Right now we know that XAI has been accused of straining local infrastructure at their data center in Memphis, Tennessee.
I mean, this is probably where he's getting a lot of this from.
I think he's gotten a lot of flack for that.
And so I think he's like look, if we're going to scale this up, how do we get it so it's not directly impacting communities?
But this is, you know, absolutely massive merger, bringing some really huge companies together.
XAI right now is burning about a billion dollars a month, according to Bloomberg.
So 12 billion a year.
SpaceX, on the other hand, earns about 80 of their revenue from launching their own Starlink satellite.
So it's not just I think.
A lot of people think, like you know, SpaceX is making their money from just government contracts or putting satellites up for other companies.
80 of the revenue is just from their own product, which is Starlink, and you know, basically global internet and if you look at a report from routers that came out last year xai, when they were acquired, x um, apparently in that deal musk said that the combined value was about 113 billion dollars and since then, with this merger, they valued xai at about 250 billion.
So since that merger uh, the company has, you know, gone up over 100 billion dollars and Elon said that this kind of space-based data centers are gonna have, they're gonna need a lot of satellites.
It's basically getting to like a steady flow of satellites.
He didn't say how many.
I think, basically the requirement is gonna guarantee though, that there's this ongoing launch demand for SpaceX.
So this is, SpaceX is never going to see the demand drop once this is kind of a new revenue stream.
And I think this is especially since the FCC rules now require satellites to be brought down after five years.
So if they're up, they have to basically have plans to deorbit all satellites after five years, which means that I think even these data centers are going to are going to have to do that as well.
So you're gonna have to bring things up, bring them down.
And I think SpaceX has just enormous amounts of business that are going to come from all of this.
For basically I think, the long term SpaceX and XAI are going to have some.
They definitely have very different short term challenges.
And they kind of have the same, I mean, this now combined long-term vision.
I think SpaceX is still focused on proving that their Starship rockets can actually bring people to the moon.
And then eventually Mars.
XAI, of course, is competing directly with OpenAI and Google.
So it's interesting.
There feels like there's these different areas.
So, putting all of the speculations on the reasons for this aside, I think you know, based on the regulatory filings that are being filed right now, I think and the merger is officially complete.
The timelines are attached.
I think this whole push to move AI infrastructure into space has kind of moved from this it feels like a science fiction kind of thing into something that is going to be very real and also very expensive.
And it looks like it is actually going to happen.
I'll definitely keep you up to date on everything rolling out with that.
Thanks so much for tuning into the podcast.
If you learned anything new, if this was an interesting show for you.
It would really help my podcast out a lot if you could leave a rating or review wherever you get your podcast.
If you're on Apple, drop a comment, drop some stars.
I read them all and I really appreciate them.
If you're over on Spotify, it's the About tab that you can hit.
So thank you so much.
If you'd be able to do that, if you haven't already, I'd super, super appreciate it.
And as always, make sure to go check out AI box.ai.
We have our brand new $9 a month tier that gets you access to everything for less.
If you want to go check it out, there's a link in description to AI box.ai.
