xAI joins SpaceX • Hacker News Board

Re: xAI joins SpaceX

#301

Post by lifis » Mon, Feb 02, 2026, 10:51 PM UTC

sebzim4500 wrote: post #43

Saline9515 wrote: post #28

I still don't understand the "data center in space" narrative. How are they going to solve the cooling issue?

Cooling a datacenter in space isn't really any harder than cooling a starlink in space, the ratio of solar panels to radiating area will have to be about the same. There is nothing uniquely heat-producing about GPUs, ultimately almost all energy collected by a satellite's solar panels ends up as heat in the satellite. IMO the big problem is the lack of maintainability.

According to Gemini, Earth datacenters cost $7m per MW at the low end (without compute) and solar panel power plants cost $0.5-1.5m per MW, giving $7.5-8.5m per MW overall.

Starlink V2 mini satellites are around 10kW and costs $1-1.5m to launch, for a cost of $100-150m per MW.

So if Gemini is right it seems a datacenter made of Starlinks costs 10-20x more and has a limited lifetime, i.e. it seems unprofitable right now.

In general it seems unlikely to be profitable until there is no more space for solar panels on Earth.

Re: xAI joins SpaceX

#302

Post by SilverElfin » Mon, Feb 02, 2026, 10:51 PM UTC

adastra22 wrote: post #266

SilverElfin wrote: post #157

Earlier quoted context omitted.

Let’s be honest - this is just a way to prop up Twitter/X. It makes SpaceX shareholders subsidize X, and also American taxpayers who are giving contracts to SpaceX for highly sensitive things. The government should ideally refuse to give SpaceX work unless it unwinds this.

Why? The government is paying less for SpaceX than alternatives. It th cheapest and best service.

Because Twitter/X is distorting our politics (with ann unbalanced scheme of censorship / amplification / suppression) and destroying the country by mainstreaming far right supremacist politics. Twitter/X does not deserve a single dollar of taxpayer money. If SpaceX is now part of that machine, it doesn’t deserve a single dollar either. I would rather pay more for alternatives and encourage their growth. I also look at any money given to this company as the equivalent of GOP campaign funding, so I feel it should be treated as illegal under the law.

Re: xAI joins SpaceX

#303

Post by alpha_squared » Mon, Feb 02, 2026, 10:51 PM UTC

The really skeptical take here is that eventually all of Musk's companies merge, or at least the biggest ones, for juicing that market value to get that $1T payout. Looking at Tesla.

Re: xAI joins SpaceX

#304

Post by Culonavirus » Mon, Feb 02, 2026, 10:51 PM UTC

afavour wrote: post #107

rybosworld wrote: post #70

> The basic math is that launching a million tons per year of satellites generating 100 kW of compute power per ton would add 100 gigawatts of AI compute capacity annually, with no ongoing operational or maintenance needs. Ultimately, there is a path to launching 1 TW/year from Earth. > My estimate is that within 2 to 3 years, the lowest cost way to generate AI compute will be in space. This is so obviously false. Fo…

As soon as a statement contains a timeframe estimate by Musk you know to disregard it entirely.

The thing is: at the end of the day, SpaceX takes the "impossible" and makes it "late".

People are going to Tory Bruno the space datacenters until one day their Claude agent swarm's gonna run in space and they'll be wondering "how did we get here"?

Re: xAI joins SpaceX

#305

Post by kortilla » Mon, Feb 02, 2026, 10:52 PM UTC

Buttons840 wrote: post #50

SpaceX is too big to fail. It's important for national security. I wonder if Elon wants to tangle all his businesses into SpaceX so they are all kept afloat by SpaceX's importance.

Being too big to fail is not really a desirable outcome, it’s just better than failure.

Boeing is too important to fail as well but it’s been terrible as a shareholder

Re: xAI joins SpaceX

#306

Post by miltonlost » Mon, Feb 02, 2026, 10:52 PM UTC

CGMthrowaway wrote: post #233

rybosworld wrote: post #70

> The basic math is that launching a million tons per year of satellites generating 100 kW of compute power per ton would add 100 gigawatts of AI compute capacity annually, with no ongoing operational or maintenance needs. Ultimately, there is a path to launching 1 TW/year from Earth. > My estimate is that within 2 to 3 years, the lowest cost way to generate AI compute will be in space. This is so obviously false. Fo…

There's clearly rhetorical hyperbole happening there. But assuming that thermal rejection is good in space, & launch costs continue falling, as earth-based data centers become power/grid-constrained, there is a viable path for space power gen. The craziest part of those statements is "100 kW per ton." IDK what math he is doing there or future assumptions, but today we can't even sniff at 10 kW per ton. iROSA [1] on t…

"There's clearly rhetorical hyperbole happening there" in a business paper is called lying

Re: xAI joins SpaceX

#307

Post by bastawhiz » Mon, Feb 02, 2026, 10:52 PM UTC

CGMthrowaway wrote: post #233

rybosworld wrote: post #70

> The basic math is that launching a million tons per year of satellites generating 100 kW of compute power per ton would add 100 gigawatts of AI compute capacity annually, with no ongoing operational or maintenance needs. Ultimately, there is a path to launching 1 TW/year from Earth. > My estimate is that within 2 to 3 years, the lowest cost way to generate AI compute will be in space. This is so obviously false. Fo…

There's clearly rhetorical hyperbole happening there. But assuming that thermal rejection is good in space, & launch costs continue falling, as earth-based data centers become power/grid-constrained, there is a viable path for space power gen. The craziest part of those statements is "100 kW per ton." IDK what math he is doing there or future assumptions, but today we can't even sniff at 10 kW per ton. iROSA [1] on t…

Maybe I'm just out of the loop, but is solar substantially more efficient in space? I assume the satellites won't orbit in a way that follows the sun. And presumably the arrays of panels they can attach to a satellite don't exceed the size of the panels you could slap on and around a data center (at least without being insanely expensive).

Re: xAI joins SpaceX

#308

Post by fragmede » Mon, Feb 02, 2026, 10:52 PM UTC

Veserv wrote: post #184

wongarsu wrote: post #117

Earlier quoted context omitted.

You operate them like Microsoft's submerged data center project: you don't do maintenance, whatever fails fails. You start with enough redundancy in critical components like power and networking and accept that compute resources will slowly decrease as nodes fail No operational needs is obviously ... simplified. You still need to manage downlink capacity, station keeping, collision avoidance, etc. But for a large con…

You mean you operate them like Microsoft's failed submerged data center project [1]. When pointing at validating past examples you are generally supposed to point at successes. [1] https://en.wikipedia.org/wiki/Project_Natick

Did we read the same Wikipedia page? It doesn't say the word "failed" anywhere on it.

Re: xAI joins SpaceX

#309

Post by cosmicgadget » Mon, Feb 02, 2026, 10:52 PM UTC

consumer451 wrote: post #251

rybosworld wrote: post #70

> The basic math is that launching a million tons per year of satellites generating 100 kW of compute power per ton would add 100 gigawatts of AI compute capacity annually, with no ongoing operational or maintenance needs. Ultimately, there is a path to launching 1 TW/year from Earth. > My estimate is that within 2 to 3 years, the lowest cost way to generate AI compute will be in space. This is so obviously false. Fo…

Here is my main question: Musk is on record as being concerned about runaway "evil AI." I used to write that off as sci-fi thinking. For one thing, just unplug it. So, let's accept that Musk's concern of evil runaway AI is a real problem. In that case, is there anything more concerning than a distributed solar powered orbital platform for AI inference? Elon Musk appears to be his own nemesis.

He just says stuff to convince people of things that benefit him. Internal consistency was never the plan.

Re: xAI joins SpaceX

#310

Post by TacticalCoder » Mon, Feb 02, 2026, 10:52 PM UTC

Zigurd wrote: post #96

Buttons840 wrote: post #50

SpaceX is too big to fail. It's important for national security. I wonder if Elon wants to tangle all his businesses into SpaceX so they are all kept afloat by SpaceX's importance.

Starship has a large number of critical milestones coming: Can it land and quickly reuse the upper stage? If not, it can't make refueling flights without building a dozen or two starships. Can it carry the full specified payload? If not, it can't even try to refuel in orbit. If it can't refuel in orbit, it can't go beyond earth orbit. Etc. Everything has to go right or it will be irrelevant before it works.

> Everything has to go right or it will be irrelevant before it works.

Starship is not all of SpaceX. Saying, maybe because one hates Musk, that SpaceX is going to become irrelevant is wishful thinking.

In 2025 SpaceX launched more rockets into space than the entire world ever sent in a year up to 2022, something crazy like that.

Then out of, what, 14 000 active satellites in space more than half have been launched by SpaceX.

SpaceX is, so far, the biggest space success story of the history of the human race (and GP is right in saying that SpaceX is now a national security matter for the US).