[downrightmike]

During the irrational slap-'AI'-on-everything to juice the stock.

Their current plan is the same as that shoe company that tried to rent out their warehouse space for AI compute.

Same plan as Meta, other than scammers using meta AI to scam our parents out of their money. (Meta makes billions off of this use case and really shouldn't)

[dlcarrier]

At least there's a real aerospace company in there. The companies pivoting will have no value when the market corrects, but affordable aerospace launches will keep on keeping on.

If anything, the anti-datacenter NIMBYism that the AI craze is generating will far outlast the AI craze, and NIMBYism is the only thing that would make launching datacenters into space cost effective, so Space X may get a long-term boost out of the whole ordeal, whether or not they have internal success in the AI market.

[ben_w]

> At least there's a real aerospace company in there. The companies pivoting will have no value when the market corrects, but affordable aerospace launches will keep on keeping on.

Perhaps, but the size of the launch market is still unknown at any given price point. Space hotels? Really expensive point-to-point 45 minute deliveries in a vehicle hard to distinguish from an ICBM on an early warning radar? High-value-low-volume free-fall manufacturing? All comes down to what the price tag is.

(And right now, don't even look at Starship until that actually delivers, look at Falcon 9: SpaceX has had so little confidence in the Starship hardware that despite 13 launches it has yet to be given a circularised orbit, this suggests development is much slower and more difficult than Falcon was).

> If anything, the anti-datacenter NIMBYism that the AI craze is generating will far outlast the AI craze, and NIMBYism is the only thing that would make launching datacenters into space cost effective, so Space X may get a long-term boost out of the whole ordeal, whether or not they have internal success in the AI market.

Could be, but space is hard even compared to otherwise hard options like the middle of the Sahara, and solutions to cooling in space also work in the Sahara.

[downrightmike]

except you're forgetting that the taxpayers have subsidized it to this point and it can't stand on it's own.

And then there's all the competition

[dnemmers]

Renting data center capacity to AI providers might be the best business model out there, currently related to AI. Surpassed possibly by hardware manufacturers.

[torh]

I'm not so sure. Is there really a demand for more AI data centers? Has anyone been turned down by OpenAI or Antrophic because they are out of capacity?

[benoau]

I think the demand is more about what's not using AI yet, plus half the world isn't really dependent on online services and platforms yet, two huge areas of growth.

[duskwuff]

There's certainly demand for more data centers. Whether there's demand for the services those data centers would provide remains in question.

[Zigurd]

Any jackass <cough>Kevin O'Leary</cough> can announce a data center project. Then you have the pivoting crypto bros. The competent data center builders are the hyper scalers who have been building and running data centers for decades.

Low margins, lots of garbage participants, and a few competent players who are likely to crush anyone less efficient. That doesn't look very remunerative.