Google's Wild Plan to Fix Our Energy Crisis (AlIn Space)

Everyone thinks that the most valuable resource in AI is the GPU. It's why

Nvidia is worth $5 trillion, right? But

what if I told you you're wrong? What if

I told you that the most valuable resource in AI is actually energy?

Without energy, you can't power these things. Without energy, the GPU is just

things. Without energy, the GPU is just a worthless piece of [music] metal junk.

And in fact, a lot of AI labs are facing this challenge where they just don't have enough energy to power the tens of billions of dollars of GPUs that they've spent so far. In fact, Satia Nadella of

Microsoft recently said that he has hundreds of millions of dollars worth of NVIDIA GPUs in his data centers collecting dust. Now, there are a bunch

collecting dust. Now, there are a bunch of efforts focused on solving this energy crisis. You've got nuclear

energy crisis. You've got nuclear energy, you've got renewable energy, but one particular solution that's taking up a lot of mind space with Elon Musk and Google is literally out of this world.

I've got six words for you. AI data

centers in outer space. That is a real sentence that I've just said. It's the

concept of putting GPUs in outer space to harness the energy of the sun for AI training and for AI inference. And

whilst it might sound like a far-fetched idea, it's not when you consider that the resources needed on Earth to power up the data centers, which by the way are requiring as much energy as a mid to

large uh city in the US. You might need to look at alternative resources for this. The one being the sun. And the

this. The one being the sun. And the

economics kind of work out in theory.

you have uh up to three to five times more energy efficiency if you operate GPUs above the Earth's atmosphere than below it. So, in this episode, we're

below it. So, in this episode, we're going to dig into all things AI space data data centers. Is this real? Is this

achievable? Or is this kind of like a myth that we're going to laugh about and have pile on our face 6 months from now?

Um Josh, you and I have a tumultuous relationship around this topic. 50% of

us were were optimistic. 50% of us, you were pessimistic about this. What's your

take on all of this?

>> I was the biggest hater in the world.

Space was so stupid. Why are we sending stuff to space when we have it all here on Earth? Well, my mind has changed. I

on Earth? Well, my mind has changed. I

have uh flipped 180°. I'm pro space. I'm

ready to send these things to space tomorrow. I want all the data centers in

tomorrow. I want all the data centers in outer space immediately. And when I was considering why, I did a lot of research, which we're going to get into in this episode. But there's this combination of things where there is a

power grid straining on the US where data centers now are one of the main drivers rising electricity demand. We

can't fulfill all the electricity demand. There are limits on land, water,

demand. There are limits on land, water, permitting, cooling. And I was thinking,

permitting, cooling. And I was thinking, I was like, well, if your house is full of all these gaming PCs and the power keeps going out, well, you start to ask yourself, well, where where else can we put them? And you're like, on [laughter]

put them? And you're like, on [laughter] the roof? No. Why don't we just send

the roof? No. Why don't we just send them off into outer space? And

thankfully, because of new technologies like SpaceX, Starship, we can do this.

So, there are some benefits to space that I want to go over very quickly before we get into who's talking about this and who's doing this. One of them being the most important one, unlimited solar energy. A lot of people don't know

solar energy. A lot of people don't know this, but 99.98% of the entire solar system is the sun.

All the mass of the solar system, it exists in the sun. And the sun is just this huge fusion reactor. And in fact, if you can get access to the sun from outside of our atmosphere, if you leave

Earth's atmosphere and you go into just normal space, it's actually eight times more powerful than on Earth. So if you remove the constraint of Earth's atmosphere, immediately these solar panels get access to eight times more

energy than they would have in the case of the out space. Space is also a vacuum, which means you can radiate heat directly into this empty space where there's nothing to worry about. There's

no land. There's security and sovereignty because let me tell you a jazz it's a lot harder to hack a data center in outer space than it is to do so here on Earth. So there's a lot of perks we're going to go through but I want to hand it back to you to talk

about who's been pro space because there's some some notable names that are very interested in the space race here.

>> Yeah, there's a lot of people um talking about this Josh. Um, before we get into that, I kind of want to like lay out that this is a very real problem and challenge cuz I'm sure a lot of people are listening to this and thinking, well, but like I haven't heard any news

about this. Let me introduce you to

about this. Let me introduce you to some. Um, there are two data centers

some. Um, there are two data centers that got permitted to build these data centers from 2019. By the way, Josh, they're worth hundreds of millions of I think like book value is like $500

million between both of them. They're

collecting dust. There's no lights on in the data centers because they just can't get the energy supply to power these right and I think there's none other than Sachin Nadella of Microsoft which

can explain succinctly what exactly this problem is. I'm going to play a little

problem is. I'm going to play a little clip right here.

>> The biggest issue we are now having is not a comput but it's a power and it's sort of the ability to get the builds done fast enough close to power. So if

you can't do that, you may actually have a bunch of chips sitting in inventory that I can't plug in. In fact, that is my problem today, right? It's not a supply issue of chips. It's actually uh

the fact that I don't have warm shells to plug into. And so how some supply chain constraints emerge, tough to predict. Uh because the demand is just

predict. Uh because the demand is just going, you know, is tough to predict.

>> So he's not also the only one that's been talking about this, Josh. Um, we've

got uh the likes of SpaceX employees saying by 2045 we're going to need enough compute to the base power of the entire planet to power all kind of like data center efforts. Elon Musk says that

there'll probably be electricity shortages in about 2 years time. Sam

Alman says we need to make a breakthrough whether it's through fusion or something else. Mark Zuckerberg says listen we would build a bigger data center but we just don't have the energy to supply. So this is obviously

to supply. So this is obviously something that is captivating some of the brightest entrepreneurial minds uh um in the world right now when it comes to scaling uh AI and you know we see

this at the infrastructure level as well. You you just heard Satcha talk

well. You you just heard Satcha talk about something called the power shell.

Um what he's referring to here is not even just the energy supply lines that you need to feed electricity into these data centers. There's also like the

data centers. There's also like the manual labor Josh there is the technicians the specialists that you need to kind of like boost this entire thing. There's so many missing gaps here

thing. There's so many missing gaps here that we just don't meet it.

>> It's funny. Um, when I was in school, Boseies and like specialty job training was very popular and that's kind of where you learn to become an electrician or a plumber and a lot of these people, they didn't acrew a lot of debt. They

left school early and they have like wellpaying jobs now. And it turns out in a world of post AI where we have a robots haven't quite made it here. The

people who we need the most are the electricians, are the plumbers, are the people who can do the cooling, the heating, the electricity, wiring. Those

are the most valuable jobs in the world and the people are getting paid a tremendous amount of money. Now before

we get into this next section, I do want to frame a new question that we're going to answer a little bit later, but it's the actual narrative of space because a lot of people are trying to raise a tremendous amount of money which

requires a tremendous amount of optimism from the investors and the public. And

you have to ask the question, is this a real narrative or is this kind of an exponential kind of like moonshot guess?

moonshot like kind of play that they're targeting to get people interested to raise more money to build their AI. I

don't want to answer that yet. We're

going to answer that as we go. But by

answering it, we're going to start with the first company who's actively doing this, which is a project called StarCloud. Now, StarCloud is the first

StarCloud. Now, StarCloud is the first prototype as far as I'm aware, right, that's doing this. Can you tell me a little bit more about it?

>> For sure. Yeah. So, StarCloud is kind of what triggered this entire debate online, Josh. Prior to StarCloud

online, Josh. Prior to StarCloud announcing their prototype of launching a GPU in outer space, everyone was kind of like oblivious to this entire thing and then Google and Elon Musk chimed in after it. So these are these are the

after it. So these are these are the pioneers and they kind of pioneered it with this prototype which should already be launched in November actually which is they launched an Nvidia H100 GPU which is technically the most powerful

computer that's ever existed in outer space into outer space using Elon Musk's SpaceX shuttle. And um they ran a Google

SpaceX shuttle. And um they ran a Google open-source model called Gemma. So it's

currently live and up there running. And

the reason why this is such a big deal is it's mainly a testing phase to see if GPUs can survive in outer space. It's

running on its own. It's training on its own. It's being iterated to on its own

own. It's being iterated to on its own outside in space whilst communicating to their servers on Earth. And what they're really trying to figure out here is number one, is it feasible or will this GPU just die from radiation? We're going

to get into some of the challenges in a second. Uh, and number two, does it make

second. Uh, and number two, does it make economical sense? Like, if we can do

economical sense? Like, if we can do this with one GPU, can we scale this to hundreds of thousands of GPUs? So,

they're the real ones, Josh, that started this trend.

>> The initial thing that I found funny about StarCloud was the 16 km heat diffuser. That was kind of where I got

diffuser. That was kind of where I got caught up on where 5 G of energy means 16 km of metal that it took to diffuse this. Because again, in space there is

this. Because again, in space there is no atmosphere. there is no place to cool

no atmosphere. there is no place to cool it besides shooting the radiation off into deep space. It's not like you can run like cold air over this because there is no air. Uh so I think as I've

kind of learned more about the StarCloud project, I've started to realize like, okay, that's that's a that's kind of within reason and that's an early prototype in how they're going to go about actually dissipating heat. So six

square miles of solar panels, they fold it into a rocket, they could send it out on a starship. It's like it's pretty cool. It seems like it makes sense. you

cool. It seems like it makes sense. you

Jazz, is there anything that I'm missing here that technically doesn't make a whole lot of sense based on these posts?

>> Yeah, I mean, you nailed uh technical challenge number one, which is like these deployable structures are not only so large for the amount of compute that it's supposedly meant to be supporting.

Like these things, like you said, are 16 km. That is huge to fuel. Do you know

km. That is huge to fuel. Do you know how much power it fuels actually, Josh?

It's like what's it one gawatt or is it less than that? Like maybe 100 kow.

Well, it's showing 5 GW now from the data center.

>> Yeah.

>> Which is pretty right. 4x4. Yeah. So, so

that's technical challenge number one.

And then technical challenge number two down here is the radiation that you mentioned which is basically you know you have all these kind of like nuclear and atoms flying around which can really

mess with the composure and structure of these data centers. Um radiation is a very real issue in space. That's why you have all these different space suits and insulation. So if you expose hardware to

insulation. So if you expose hardware to this and it kind of messes up with any kind of the machinery of it's doing its job then it becomes more and more implausible. But these are things that

implausible. But these are things that they we're kind of at like the frontier testing in terms of research. We haven't

really found a resolution just yet. Um

there are solutions like kind of like cooling that you mentioned but otherwise there's no clearcut solution.

>> Yeah. Okay. But there is another company that we're all familiar with that does have a seemingly more clear-cut solution and a seemingly more clear road map which um I believe the CEO goes by the

name of Elon Musk if I'm not mistaken and Elon and SpaceX have a strategy. Uh

we have a clip here that I think you wanted to share. Maybe we could start there and we'll get into how exactly this is all going to work.

>> In in order to harness a non-trivial amount of the energy of the sun, you have to move to solar powered AI satellites in deep space. um which

somewhat is a confluence of Tesla expertise and SpaceX expertise um and XAI on the the AI front.

>> So there's three critical pillars that Elon mentioned in that little summary.

There's Tesla, SpaceX, and XAI. SpaceX

is the most the single most important pillar in that because without the ability to get mass to orbit, it is impossible to build these data centers in the first place. And the reason why the window is just barely starting to

open is because we're able to get the cost per kilogram down low enough to make it economically viable because when we started with the space shuttle E jazz, it cost $60,000

per kilogram to get something into orbit. Nothing is nothing is going to

orbit. Nothing is nothing is going to make that viable if it cost $60,000 per per kilogram because these data centers are very very heavy and so are the solar panels. So are the batteries. So all of

panels. So are the batteries. So all of that. Um, Falcon Heavy, which is

that. Um, Falcon Heavy, which is SpaceX's newer rocket, got that cost down to $1,500 per kilogram. So, we've already seen

per kilogram. So, we've already seen like a few orders of magnitude decrease.

And Starship, which we've covered in previous episodes, that is planning to go to Mars and beyond. It's still not working, but they're planning to get it fully reusable by next year. That brings

the cost per kilogram down to $10 per kg, which is the window that enables this conversation to happen. Because

without in the absence of Starship, we cannot send mass to orbit at scale. With

Starship working, we can actually talk about sending these data centers into space. And that's a hugely powerful

space. And that's a hugely powerful unlock that only exists because of SpaceX, which is why I think it's important to pay attention to Elon when he talks about this stuff because he's the one who's enabled this to even be a

possible conversation. Um, Earth gets

possible conversation. Um, Earth gets only one to two billionth of the sun's energy. So if you go a million times

energy. So if you go a million times Earth's capacity, you must go into space is kind of the thinking where in order to scale AI to the size that we want in order for true AGI for for AI that

scales infinitely with us, we need to go past Earth's ability to generate energy and we need to go out into the stars and that's kind of the the thesis with Elon and the XAI thing.

>> Got it. Um I mean I was just looking at this other post Josh um where you know StarCloud is kind of laying out the economics for their kind of projections and they say something pretty similar

which is like um on Earth it costs near like $200 million to kind of run a 40 megawatt data center for 10 years. Uh

technically or theoretically in orbit it would be worth $8.2 million which is 20 times cheaper. So so what I'm hearing

times cheaper. So so what I'm hearing from you and the Elon case is that well number one he's going to be the toll master. He's gonna basically own the

master. He's gonna basically own the highway in space.

>> He's the Jensen Wong of space.

>> That he's the Jensen Wong of space.

That's that's perfect. We we have to trademark that term. He's the Jensen Wong of space. That's that's crazy good.

So, he's going to not only own that toll. Um but the other thing, Josh, um

toll. Um but the other thing, Josh, um which I know we both love the idea of and which he actually commented on uh in in this tweet. He says simply scaling up Starlink V3 satellites which are the

main satellites that he deploys from his SpaceX launches um will have high-speed laser links which would work and what he's referring to here is work in terms of like a interconnected network. So

these satellites would beam lasers to each other 1 terab per second speeds to be able to train or inference AI models up at that level. So imagine like a

satellite network kind of in space orbiting Earth that you can use to beam terabs worth of data at super rapid speeds between each other satellites,

train your AI model and then beam it back down to Earth, which just sounds so nuts at this point, Josh, that I I can't quite wrap my head around. Um, and then the third thing that he's talked about,

Josh, I I don't know if you saw this, he's talking about quantum computing on the dark craters of the moon. Can can

you sorry can you help me wrap my head around this? What what the hell is this?

around this? What what the hell is this?

What's this?

>> I'm pulling a number out of thin air. I

I believe it's right though where the craters there are craters in in the moon. So there is no dark side of the

moon. So there is no dark side of the moon but there are craters that never see the sun. Those craters exist at a perpetual 200°. They are literally

perpetual 200°. They are literally frozen solid. There is no heat that gets

frozen solid. There is no heat that gets in. And if if you can place a gigantic

in. And if if you can place a gigantic data center in one of those craters at -200°C or whatever the exact amount is, uh you have a pretty good cooling

system. Like it's not that bad. And

system. Like it's not that bad. And

each, as you'll often see with quantum computers, there's this huge mechanism on top of them and then there's this tiny little chip at the bottom. The

entire mechanism on top of it is just to keep it cool. The actual chip is is the size of a normal chip. But because it's so important and so necessary to get quantum computing chips at solid zero so

that the cubits don't move around too much that that's why they need it. So

there is a world in which quantum computing makes sense if we can get it into outer space. There is some data and energy notes that Elon has that I wrote down that I want to share briefly cuz

this is also fascinating. and he said Starship should be able to deliver around 300 gawatt per year of solar powered AI satellites to orbit maybe 500

gawatt. So if this is true within a few

gawatt. So if this is true within a few years of full rate Starship launches orbital AI could rival or exceed entire countries electrical consumption. Um so

the scale of this is outrageous and granted this is all very optimistic.

This is all very um very forwardlooking, very sci-fi, but in the case that is even directionally correct, the next 10 to 20 years of these launches, the next

10 to 20 years of compute can scale in a really exciting way. And we were talking to to Luke who helps produce the show right before this and he he made a great point that said in order to get exponential gains in progress, you need

exponential gains at technology or something along those lines. It's so

true where if we really want to build AI that changes the trajectory of the universe, you need to do something novel, something new. And doing this all in space and unlocking this new energy source feels like the natural progression given where we are in terms

of technological capability back here on Earth. It's also equally fair to say

Earth. It's also equally fair to say that Elon sometimes gets ahead of himself. Like he can see the vision,

himself. Like he can see the vision, right? But his projections are way too

right? But his projections are way too near-term to what it actually requires to get to that point. And I wish I could critique him, Josh, except that there's also another person which is bulling

about AI data centers in space. And it

is Sunda from Google, the CEO of Google, who in pretty much every single interview, this clip that I'm showing you on on the screen right now is from an interview he did literally yesterday, which has got nothing to do with

Google's AI space efforts specifically, but he had to shell Google's project called Project Suncatcher, where they are also aiming to put Google's TPUs, which is their GPU equivalent, into

space so that they can harness uh the power of the sun. There are too many important people talking about these things, Josh, where like my inner kind of like half my brain is kind of fighting against this, saying this is absolutely ridiculous, but the other

half I think is overwhelmingly now in favor of this thing might actually work.

Yeah, it's uh it's their moonshot. It's

another Google moonshot. And again, I'm going to tease the question, is this just a narrative thing to get people excited to invest more in their companies or is this real technology?

Maybe we'll see. He is partnering with Planet, a company named Planet, and their plan is to launch two prototype satellites by early 2027 to test CPUs and optical links to orbit and see how that works. In the previous example with

that works. In the previous example with SpaceX, you mentioned that there is these like space lasers that can talk to each other. A cool thing is that you can

each other. A cool thing is that you can communicate at the speed of light. There

is no kind of interference in space. So

the time that it takes for a photon to get from one laser to or one satellite to the next is the latency between the two, which is very very fast. So there

is a world in which theoretically if you're looking at this from first principles based on there are no laws of physics that are being violated in order to do this. So therefore it's a matter of technical capability prowess um and

our intellectual abilities to actually figure out and solve these problems. It is it's funny seeing Google and Elon kind of converge on this. I assume like you mentioned earlier a lot of other companies are not far behind. They just

don't quite have the infrastructure or the ability to to tackle a problem like this. But it seems very real and They're

this. But it seems very real and They're planning to do this as soon as the next 24 months with with the new TPUs that we've been talking about so so often.

>> Sundar said that they're already prototyping a bunch of radiation measures uh on ground here at Earth to make sure that their TPUs can survive.

So the next natural step there is to launch these babies into space probably using Elon Musk SpaceX because he is the Tom he's the >> to that point actually the uh the economics and the launch cost they actually we have a quote here that says

uh their modeling says if launch costs get under $200 per kilogram to orbit by the mid 2030s then spacebased compute could be comparable to earth data centers on a per kow per year basis. So

if we can get 20 times more than what Elon's projecting, so if the cost is pretty high. So if it's 20 times more

pretty high. So if it's 20 times more expensive than what Elon's projecting in a decade from now, then we'll have par with Earth. So

with Earth. So >> it gives you a testament to like the the timelines of this. And maybe this is a good time to go into the pros and cons of what what is good about this, what is bad about this, and maybe maybe humble

people a little bit. EZ, if you want to start with the cons as to why this isn't as rosy as we think it might be. like

maybe maybe bring us back down to earth.

In fact, >> I that's that's hilarious. I didn't did not expect that pun, Josh. Okay. Right.

So, I'm gonna start off with the bare case here, which you've mentioned a few times, which is this is just a narrative play by some of the biggest companies to boost their valuations. And let me

explain the problem to you. Right? Up

until today, speaking about these data centers in space, the focus has been on GPUs. GPUs is the new gold, right? it's

GPUs. GPUs is the new gold, right? it's

what you need to kind of train your AI.

And then we found out that they don't have enough energy to kind of like fuel these things. Now, the reason why this

these things. Now, the reason why this is so important, um, as this post explains, is a lot of these companies, Sunda's company, Google, Elon Musk's companies are valued based on the amount

of GPUs they've been purchasing. And the

Wall Street analysts have basically assumed, okay, the amount of GPUs that this company buys is equal to the economic output. I can do a similar I

economic output. I can do a similar I can do a simple formula, right? if

they're buying buying these GPUs, they're going live. What they haven't factored in is that these GPUs aren't going live. They're just sitting

going live. They're just sitting collecting dust. And so that's less ROI

collecting dust. And so that's less ROI per quarter than they initially projected, which means that the valuations of these companies should probably be lowered. So what's the countermeasure to this, Josh? Well,

obviously it's infinite energy.

Obviously, it's what was that 99.98% of the solar system is the sun. It's all

that energy. So if you can miraculously Wow. now figure out a way to harness

Wow. now figure out a way to harness that energy. You can now maintain that

that energy. You can now maintain that valuation. Dare I say boost the

valuation. Dare I say boost the valuations, right? Share prices go up

valuations, right? Share prices go up after you you hear Sundar talking about this. Uh Google's on an absolute run. So

this. Uh Google's on an absolute run. So

that's my skeptics's take. I think that they might be kind of like fueling this narrative which they know is a decade out from any kind of real uh prototype being out there that scales. Um so

that's my bet is do do you agree? Do you

disagree with this? Yeah, I I think setting expectations on timing is probably going to be fairly important because again like Satia is saying mid 2030s. I mean Elon's probably saying

2030s. I mean Elon's probably saying tomorrow, but the reality is it's probably mid 2030s. It's going to take a while to get all this stuff to orbit and do it at scale that's comparable to the United States. Now when I I think about

United States. Now when I I think about that, I think about what does the worlds of AI look like in the mid 2030s? Like

>> what what are we what's what's it going to look like here? Like we are moving so fast and to project 10 years into the future. My god I I don't know what the

future. My god I I don't know what the world looks like. So

>> our energy bill is going to be 5x Josh may or maybe we have people like Isaiah Taylor who was on the show last week solving nuclear reactors right here on home and we actually have energy

abundance that's enough for our midterm goals prior to extending out to the long-term goals. And maybe this is a fun

long-term goals. And maybe this is a fun time to talk briefly about the cardv scale which is a sci-fi thing and there's three tiers to it. The first one being you can harness the entire energy of your planet and then it's the energy

of your sun and then the energy of your solar system. We have been able to

solar system. We have been able to collect perhaps a very small fraction of Carter level one which is of the planet the energy that hits our planet. So

there's a very long way to go but I think this is kind of like the grand sci-fi vision that gets people excited gets people motivated. If I'm going to work at a company like SpaceX I'm stoked at the idea of building super

intelligence in outer space. That's a

cool mission to have. And like so many of the other missions, it's not physics constrained. It is possible. You can do

constrained. It is possible. You can do this thing if you can figure out how to create the technology to do it. So I

love the ambition. Is it a narrative play to get more money? Maybe, probably.

But like in the case of Google and and Tesla and SpaceX, it doesn't seem that way. If this was coming from someone

way. If this was coming from someone like Open AAI or Perplexity or Anthropic, I would feel a little bit differently about it. But Google's

always had a moonshot division since the beginning of time. I mean, that's the reason why AI exists in the current state it does today is because they spend so many resources on research and development. And if you remember from

development. And if you remember from our episode yesterday, EJZ, they they developed the first neural network 15 years ago. So, it's been under

years ago. So, it's been under development for a long time. And this

very much feels like a similar project where these timelines are going to be 10 to 15 years, but the research starts now. And it's very important research

now. And it's very important research because directionally it feels like this is correct. Directionally, we're going

is correct. Directionally, we're going to need more energy than we can have on Earth. even if we solve nuclear power

Earth. even if we solve nuclear power because of the hope that it's going to require so much energy to power super intelligence that will be smarter than the smartest thing on Earth. And it's

just it's a fun science experiment.

Maybe it's a narrative violation because they're getting a little ahead of the gun. But, you know, it could it could

gun. But, you know, it could it could be. And I like this future. So do I. And

be. And I like this future. So do I. And

I think my optimistic take on this is all of these things um nuclear fision, nuclear power coming into like a scalable means, renewable energy and you

know harnessing energy from the sun all comes collectively at a time in the mid 2030s where our energy bills are going through the roof and we cannot afford anything else. We need to rely on

anything else. We need to rely on another energy source and I think it's all going to neatly come at the right time. Josh, I don't know what makes me

time. Josh, I don't know what makes me think that this might actually happen, but I think we are well on our way to do that, but to your point, there's a bunch of challenges, right? So, we've

mentioned kind of like some engineering challenges around protecting against radiation and building and deploying these infrastructures out in space.

Right now, it's too expensive, right?

But the idea is presumably using SpaceX's um I'll call it the tunnel to space or the highway to space, it'll eventually reduce the cost down uh by uh you know, maximally. So, we're able to

kind of achieve this. The other thing is kind of like maintenance, right? So it's

like a lot of these things require a lot of attention. Like we mentioned earlier

of attention. Like we mentioned earlier that, you know, supplying energy is just one thing. You need to have technicians

one thing. You need to have technicians and experts on there that can like kind of connect the right plugs between satellites and make sure that they speak to each other or GPUs between each other. Are we going to have astronauts

other. Are we going to have astronauts roaming around these data centers on space? That is kind of harder to wrap my

space? That is kind of harder to wrap my head around because it just seems infeasible. We don't do that right now.

infeasible. We don't do that right now.

Why would it be in 10 years time? Maybe

it will be. I'm being optimistic. I have

to maintain optimism here.

>> Maybe robots. robots in 10 years.

Probably going to be pretty good.

>> Oh Josh, I need to buy more Tesla. The

humanoid.

>> I need to buy more Tesla. You've just

The humanoids. Oh god, we are getting way too bullish on the section. On the

con section. Okay, hang on. Wait, hang

on. We got We got to remain tethered to Earth. Uh, speaking of tethering

Earth. Uh, speaking of tethering [laughter] to Earth, actually, >> how do we get this data and compute that we generate in space down back to Earth?

Presumably, the majority of the humans are still going to be on Earth, right Josh? So, a question that's in my mind

Josh? So, a question that's in my mind is like, what is the receiving infrastructure of this look like? And is

that economically feasible? Um Elon

seems to claim so in that in that thread that we showed earlier and Sunda has said multiple times that yeah we can just beam data down. We can just beam some of the learnings the compute down.

I don't know if anyone's actually proven that and it's because we need to do step one first which is launch the data center. So there there's so much that

center. So there there's so much that this relies on that needs to be kind of solved via frontier research that we're kind of hoping for but we haven't got any definitive proof if that makes sense. So if you're investing in this

sense. So if you're investing in this stuff, it's definitely a longer horizon type thing.

>> Yeah, conceptually it all makes sense.

We're learning a lot from Starlink actually in terms of latency and sending things back to Earth where if you're sending data that's timely, it's a little difficult because there are tens to hundreds of milliseconds that are added in the latency between the two.

But if you're batch training, if you're just training me mega amounts of AI and then you're beaming them in batches down to Earth, there's enough bandwidth uh if you disregard latency to send quite a bit of data pretty quickly. So, I I

suspect there's some healthy middle ground there where they'll talk to each other very fast and they'll beam data down kind of slow and that will be okay.

Um, there are pros. There's lots of pros. We talked about a few. Energy

pros. We talked about a few. Energy

abundance being one, cooling being one where you can just create mega radiators. You could send things on the

radiators. You could send things on the dark side of the moon, put them in craters where it's really cold. Um, the

environmental thing, there's you don't need permits or you don't need to use water or you don't need. we're just

unobstructed by AI, which is kind of an interesting thing because currently AI is is very intrusive. Um, it uses a tremendous amount of data and resources and energy and just human labor. Um,

which maybe is a good thing to some extent, but that's another thing. And

then we have just like the sovereignty and resilience point where when stuff is in space, it is pretty antifragile. No

one's really going up there to mess with it. Uh, it is very secure. You kind of

it. Uh, it is very secure. You kind of know what you're getting when it's up there. You set it, you forget it. It's a

there. You set it, you forget it. It's a

nice thing. So that's kind of how I think we would weigh the pros and cons.

>> Okay. So if if those are the pros and those are the cons, the next natural question for me, Josh, is like if I was an analyst looking at this over the next say like 5 to 10 years, what kind of

milestones should I be looking for? Like

do you have any idea or like what's like what's sensor out there or is it a couple of GPUs?

What is it?

>> Yeah. So because there are a few companies that are actively working on this problem, we can just kind of evaluate their progress as I think a way of of tracking things. So we could set a few milestones starting with the

Suncatcher prototypes um who are planned to go up in 2027. So if you're trying to map out trajectories, we have Project Suncatcher. We could ask the question,

Suncatcher. We could ask the question, did the 2027 Suncatcher prototypes launch on time? Did they actually even make it to outer space? Is Starship

working? Are they able to carry the correct payloads? Is the payload even

correct payloads? Is the payload even able to get into outer space? Then once

it's in outer space, we can ask the question, well, does it actually work?

Is it able to shield itself from radiation? Is it able to power itself

radiation? Is it able to power itself on? Is it able to cool itself down? Um,

on? Is it able to cool itself down? Um,

do they show working optical links and TPUs in orbit? Can they actually make that work? So, we have the prototypes of

that work? So, we have the prototypes of Suncatcher. Then we want to look at

Suncatcher. Then we want to look at Starship cadence. Is Starship actually

Starship cadence. Is Starship actually getting the cost to orbit low enough to make this economically viable? That I

think actually is the single most important thing to look out for is the Starship program. If you are not

Starship program. If you are not watching every Starship launch with us on this show, you are missing out because we're going to cover every single one. and it's the most important

single one. and it's the most important thing to the progress of this project and we are going to closely monitor how the cost per kilogram to orbit is going.

As long as that number is going down, as long as Starships start blowing up less, that's going to be a huge win and that's going to make it at least plausible. And

then the third thing I guess to monitor is the power crisis on Earth. We want to see how much power we're able to generate outside of these moonshot ideas. So, are people like Isaiah Taylor

ideas. So, are people like Isaiah Taylor of Valor able to create modular nuclear reactors to power gigawatts of these data centers? and eventually terowatts.

data centers? and eventually terowatts.

Are we able to use natural gas turbines and other forms of renewable energy to not only power data centers but our day-to-day lives? Can we do so without

day-to-day lives? Can we do so without making the electricity bills run too high? So, I think the convergence of

high? So, I think the convergence of these three things are are things to look out for that we can monitor uh that will probably give us a better idea of the road map how we're going in the trajectory. And again, 10 years from

trajectory. And again, 10 years from now, who knows what the world looks like, but that's what I'm going to be looking out for at least. Yeah, I think that last point is actually uh one of the most important things because maybe

this is a hot take, but I think the AI data centers in space hype dies down if things like nuclear and renewables can scale on Earth because >> people aren't going to focus on things

in outer space if they can just kind of fix the thing on problem on Earth. I

think this is more of a philosoph philosophical semantic thing due to human nature versus kind of doing something in the sky where they haven't spent that much time out there, right?

And then the other thing is like to your point, we got to reduce the cost of space travel. Uh which I fully believe

space travel. Uh which I fully believe Elon's going to do, but he has to prove it before people get really bullish and confident that this can be a thing. Um

but that's it. Those are the pros and cons and and that is the thesis and the pitch the the for and against for AI data centers in outer space. Um, it's so

funny seeing uh our evolution and reaction to this Josh given that 6 weeks ago we were absolutely laughing at this when StarCloud announced that they were launching an H100 GPU in space and now

here we are opining about um, you know, Elon being the toll master, it being cheap enough to launch uh, data centers in space and then figuring out how to beam down terabytes of data to Earth.

Just an insane thing that I can't quite wrap my head around. Uh, but it's been a super fun episode. Um, and just like the trajectory of putting data centers in

outer orbit, there's something else that is reaching the levels that is exiting the stratosphere. Josh, it is the

the stratosphere. Josh, it is the limitless follower page. It is our pages. It is our YouTube pages. We have

pages. It is our YouTube pages. We have

gained I think what was it 3,000 subscribers over the last couple of more? How about how about tell me

more? How about how about tell me >> over 4K >> and that's not enough because 80% of those people are still not subscribed who are listening to the episode. So if

you're listening on YouTube please make sure not only click the subscribe button but there's a little bell next to it which means you could turn on notifications to get notified whenever a new episode comes out. Happens three

times a week. They're all pretty good.

They normally range from 25 to 35 minutes. And if you enjoy them with your

minutes. And if you enjoy them with your friends you should share them along because it really makes a big difference to the growth of this podcast. And we're

trying to grow. We're trying to make this the biggest thing ever. We've been

stalling out a little on the audio front. So, if you do prefer to listen to

front. So, if you do prefer to listen to it as a podcast, perhaps go find it on your favorite podcast player like Apple Podcast or Pocketcast is my preferred one. Those are good places to to find

one. Those are good places to to find out and hang out with us. But yeah, I think that's mostly adjing words cuz I actually I have one parting story.

>> So, you go first.

>> Okay. Well, one final thing is we also have a newsletter and we have about 70,000 of you tuning in every single week. Josh, either Josh and I write an

week. Josh, either Josh and I write an essay or a thesis on what we think the future of AI is going to look like. Our

last one is a juicy bullc case on anthropic. You go check it out. And we

anthropic. You go check it out. And we

give you the five highlights of the week. So, we are anywhere and everywhere

week. So, we are anywhere and everywhere that you could possibly ingest information about AI and Frontier Tech.

Follow, subscribe, give us your email list. Josh, what is your parting words?

list. Josh, what is your parting words?

>> There's actually a lot of alpha in there. Google, um, EJ, you wrote the

there. Google, um, EJ, you wrote the Google bull case like four or five weeks ago and the price has gone off like 20% since then. though if you're not watch

since then. though if you're not watch if you're not listening to the or reading the newsletter I would advise I wanted to leave with one little bit of sci-fi cuz when I I do episodes like this I'm always reminded of sci-fi like when I look at the future I'm always

reminded of sci-fi books and a lot of them get it wrong but I think you could almost always find a sci-fi story that will map to the reality that we're having currently and you kind of see this with Black Mirror one of them that I would encourage everyone to read it's

called the last question by Isaac Azimov and it's it's where I see this going as we move AI into space where he kind of like tracks the progress of humanity over millions of years And each time there is a smarter and smarter and

smarter AI and it is an all knowing form of intelligence and every time it gets smarter because they've been able to capture more and more energy and eventually it leads to the last question which is the last question humanity will ever need to answer. Um which I will

leave that for you guys to explore and to uncover for yourself. But it is like it's so exciting that we get to live in this world that feels like a sci-fi novel where we're actually talking about

sending AI into outer space, building artificial general intelligence, doing so using space rockets and satellites.

It's like it's really cool and it's not here today, but there is a very clear trajectory to getting there in the future. So I think that's mostly what we

future. So I think that's mostly what we are excited about covering and I think the people who have been here with us long enough understand that and they're excited to be on the journey with us.

For all of you who have made it this far through our fairly long episode today, thank you. I hope you enjoyed the space

thank you. I hope you enjoyed the space race. Um, our AGI in space episode. I

race. Um, our AGI in space episode. I

hope you left informed, excited, optimistic, or just downright frustrated that we are missing the mark so much and AI in space is stupid. But whichever one it is, let us know in the comments down below and we will see you guys in the

next one. Thank you so much for

next one. Thank you so much for watching.