Post

The PR hype in the lead up to the IPO will be studied by historians.

Aren’t the GPU’s only good for about five years? Then you have to launch replacements up again.

Like mortgage brokers in the early aughts it's all about volume not quality, so we click more and get stupider and stupider and stupider. We should read more books. Some of my faves. 👇

genuinely, truly, comically awful And they're not even the worst of the sector

They are "journalists" in the way my dog's vomit is "art". Paid PR shills passing off advertising as writing. they need to be at least a little bit of a sociopath to go along with this tbh

Oh God they thought Palantir was too subtle didn't they

And our retirement security is going to be locked up in this madness.

🤔 PSP, the Belgian entrepreneur Thomas Leysen and the media company Antenna Group are among its new investors. Semafor launched with $25 million in financing from wealthy investors, including media industry executives and the now-disgraced cryptocurrency tycoon Sam Bankman-Fried.Jan 7, 2026

AKA people who cheated through high school and college and only care about numbers. Here's what I care about.

SpaceX IPO journalism is “Fast Company” for rockets.

In the future, all stocks will be meme stocks.

There's a trillion reasons this dystopia is teetering towards post apocalypse.

Musk announced that it has a 10% chance the company will be successful. Why would anyone buy stock that has a 90% chance of failure.

it's a big club and you ain't in it

It's PR for tech bros.

Feels very similar to collateralized debt obligation squared

Remembering all the "journalism" going on at CNBC during the frothy days of the dotcom bubble and the housing bubble. Every generation has to learn anew it seems.

Literally

Jfc these people think sci-fi is current day real life

This technology that's still being developed and hasn't been tested yet that's specifically for unloading supplies is a sure bet to remove broken hardware from a data center and replace it.

Replacing chips and fixing broken hardware are difficult to do in space. The solution? Robots! Robots that are also difficult to maintain and fix in space! Brilliant!

Not just a flight failure, but a failure to address heat.

That is priceless! OMG Icarus Robotics.

they will test unloading bags of supplies next year ——> able to replace a RAM module in a server speaking as someone who has not only worked in a server lab, but also studied robotics, yes, those are TOTALLY the same difficulty level 😆

Yes, probably because making robots that do stuff remotely in space has precedent. It's not revolutionary or disruptive technology being developed. Even getting parts up there to replace old parts is likely trivial with unmanned missions. The ignorance of heat management up there is more pressing.

100%. Daedalus might have been a better choice, but they probably have no idea. It’s like the Palantir guys.

Sounds like 90s and aughts @wired.com.

“Another common argument is that this is impossible, but what if we all close our eyes and imagine.”

As long as they don't fly too close to the sun...

It has no natural protection against radiation. It isnt manufactured on any of the silicon processes like FD-SOI that you use when youre building a satellite or space vehicle. Even if physically shielded, it still will face constant bit flips and fry outs from cosmic rays, solar flares etc.

Might as well start building data centers on the moon while we’re at it. AI has gotten so out of control.

'What do you mean 'no heat transfer via convection in a vacuum' and 'exposure to solar radiation'?'

Data centers in space is one of the dumbest ideas I've ever heard out of fElon - and he's full of really incompetent ideas. They fail every single test of a data center: - Power - Cooling - Upgrades - High speed, low latency I/O You couldn't pick a worse location for a data center if you tried.

Nah I think it's totally plausible that the cost of a self sufficient orbital data centre is about 4x the cost of a big warehouse in Alabama. A fully loaded server rack only costs about 3 million dollars to launch into low earth orbit, and of course putting it in a warehouse in Alabama costs $750k

We don’t need this.

This is just a way for AI companies to justify asking for more investors.

In college I lived with b school students (at Ross/Umich). As a science major, it absolutely blew my mind when they told me they are trained to take financial statements at face value. Like, what, do you also write your assignments in crayon? Anyway they all have way more money than me now.

It's always been that way. Hype sells, skepticism does not. That's how musk became known as some sort of tech genius while in reality he's just a mediocre con artist.

The market is being run by carnival barkers with no sense of truth. It's being run by "Trust me" bro's. FFS, just got done watching the COO of SpaceX on CNBC say we could have 1 million people on Mars by 2035 (maybe 2040)... GTFOH.

I tried reading the SemiAnalysis piece, which *seemed* to say that all of the typical arguments for data centers in space are bullshit, BUT, if you are willing to believe in infinite demand for compute, well, then off to space we go

The real question is, why do we need data centers, what are they processing, who is benefiting Ask those too

If I wanted to find an outlet to help me understand whether this is feasible… let’s just say Semafor would not be at the top of the list

Cost-effective = paid for by taxpayers

As if they need more space junk to collect. Doesn't Musk have a Tesla up there somewhere?

Welcome to the social media feedback loop, 2026.

Reminds me of the Theranos hype. Nobody is asking the hard questions about HOW this will work. It's all mania and vibes.

' They may as well claim the "cost-effectiveness" of putting the Easter Bunny in space. With enough buzzwords and acronyms, you can baffle pretty much anyone....

correct

I don't understand why Elon's isn't building them at his Mars colony with his fully autonomous Teslas? 🤔🤔🤔 www.dw.com/en/elon-musk...

I want you to notice that We The People have no idea, no definition, of what AI really is. We are not privy to what is going on Behind Closed Doors. In space or on the earth. Congress & State lawmakers refuse to regulate & disclose. It is NOT just robots. Good vs Evil?

I feel like a Richard Feynman type pithy analysis is lurking out there to be said. Physicists, could you?

any day now!

If you use solar energy then you also HEAT your system big time and it already has no sensible way to get rid of all the heat from the chips - vacuum flasks keep things hot, and space is a very large vacuum flask.

They have the electrolytes CPUs crave!

Radiation as food. Mmmm, who taught you science, Dijon?

Neil deGrasse Tyson recently talked about this concept. All they had to do was listen to Star Talk & they’d have plenty of questions

I have only read a few Semafor articles and listened to the Editor talk about trump for a while in an interview. Not good, not good at all.

2030, well aren't they optimistic. 😂

Calling your satellite company Icarus Robotics is like calling your cosmetics company Orc Skin.

Lol It's literally the worst place to build a data center. You can't really vent heat in a vacuum.

If I based my business investment decisions on whatever describes that kind of article I'd be broke more often than Trump.

Cobalt mining on Jupiter. Microchip processing on the Moon. Tesla Diner on Mars.

there is no case for data centers in space watch me prove it, with penameters to spare if you go up against enrico fermi you will lose just sit back and watch things unravel instead

Hey at least Aluminum Oxynitride is real.

Usually don't pay attention to @semafor.com. But this one was so bad I had to go over there: bsky.app/profile/bjke...

Pish posh. Thermodynamics don't exist in their world, nor does bandwidth constraints.

@semaforben.bsky.social @semafor.com

After speaking to six different sources about this, the conclusion I came to was that you could argue they're feasible, but to do so at the scale Musk has suggested would see SpaceX monopolize one of the Earth's most important orbits to the great harm of the planet. www.engadget.com/ai/orbital-a...

Space elevator

and to curry favor with jet fuel companies

The “profitability” of Starlink is only because they categorize all the launch costs under the space vertical. I doubt very much it’s actually profitable if you took into account the cost of getting the satellites into orbit

CEO said perpetual motion and ain't one reporter asked for a receipt.

It's that at 40x the scale, directly tethered to the economy and retirement savings

I just don’t know why they’d do it. Venting heat is a much bigger challenge in orbit than I think most people appreciate.,

"Unbiased Media" means every claim has the same odds as a coinflip: either it's true or it's not, with equal weight to both options.

I guess they’re also afraid that questioning tech claims will make them look stupid? No one wants to be the person who said that the internet or the iPhones would just be short-lived fads.

genius man is so full of knowledge and capacity that you do not have to worry about whether claims made about genius man are true....genius man transcends humanity and is going to bring bounty to fools who will use it to further destroy the planet and humanity.... c u later...

"Do the laws of thermodynamics allow this cost-benefit analysis" should really be one of those fundamental questions, before even considering human laws. "Elon Musk and Marc Andreesen say if governments legalize time travel, it can save civilization." Great. Has the universe legalized it?

😡

Most interesting part of watching the episode of 100 Question with Tom Simmons featuring Brian Cox was when he talked about all these funds allocated to making us an "inter-planetary species". That hardly anything, in comparison, goes to mitigating threats of asteroids like trajectory shifting tech.

The bullish business case is effectively a nested "if" statement: "The cost difference could fall to 30% within five years, and if demand zooms upward, chip supply increases, and public backlash against building data centers on the ground grows..."

Timing tied in to the stock offering for SpaceX?

but Elon said it...

I live very close to where Microsoft is building three datacenters in Brazil, like, they are bizarrely well located(There is a supermarket close to one of them), the idea that they are going to be able to build them in space is dumb. Even if energy and cooling were really cheaper in space.

Data centers are massive. How the fuck would we put one in orbit. They require shit tons of power where would they get it? Space may be cold but it's also a vacuum so heat can't escape quickly

It’s the absolute fantasy of the whole thing. Elon’s handmaidens pretending ANY of his ventures are worth investing in.

Journalism malpractice by @semafor.com

Hi. I'm an idiot. And even I know that building ANYTHING remotely like this will NOT, in any way, be remotely cost effective by the 2030s. It's patently ridiculous! It will only become cost effective once Elon is on Mars.

America was founded by hucksters selling you lead and cocaine filled cough syrup and everything since has been an iteration of that

"journalists".

It was inevitable that PR would become an outsourcing model. Nearly all forms of expository media have been perverted to Marketing in order to stoke unconstrained consumption perpetually. Combined with idolizing techbros as superheroes, too many are helpless and adrift within a systematic con.

'Dylan Byers of Puck News wrote in 2025 that Semafor has "morphed into an events business oriented almost entirely around sponsorship deals." And with Sam Bankman-Fried as one of their largest original investors and the plan hatched at Davos...well

Thanks for that info. Now I'm getting to wonder if a meteor came in and what kind of chain reaction there would be if it hit one of Musk's things and then things exploded all over the place... It's a dream...lol!

Clearly a promotional marketing piece.

finance.yahoo.com/markets/stoc...

Data centers in heaven. That's a great place for them.

The whole premise is la-la fantasy, as if the only issue with datacentres in space is launch cost.

Please tell me they spent at least 3-4 paragraphs talking about the engineering nightmare of cooling circuits in space. We must have made some super progress on converting heat pollution to infrared light that I was unaware of!

i’m no expert but seems to me they keep imagining more and more elaborate schemes to get around the serious problems with these things which can ONLY make the whole thing MORE expensive and destructive.

I mean, if we can get the tech bros to go to space to build them. Without suits.

How did we go from we need 8 kajillion data centers the size of Manhattan to we can fit a data center on a rocket.