zlacker

[parent] [thread] 5 comments
1. jiggaw+(OP)[view] [source] 2025-10-22 11:52:16
This is absolute nonsense.

The first thing to consider is that this thing won’t be stationary!

Geosynchronous orbit is much more expensive to reach per kg launched, even for Starship… when it starts working properly.

Lower orbits… aren’t stationary. Who wants a data centre that’s “over the horizon” from the owning country most of the time!?

If you think AWS egress costs are bad? Just add some zeroes! No, more zeroes than that…

replies(3): >>non-+N >>action+r1 >>esafak+Aj
2. non-+N[view] [source] 2025-10-22 11:58:10
>>jiggaw+(OP)
Nothing stopping the satellite data center from communicating back to homebase via Starlink network right?

Would probably need to negotiate for a huge amount of dedicated priority bandwidth, but latency shouldn't actually be that bad.

replies(1): >>ben_w+U4
3. action+r1[view] [source] 2025-10-22 12:03:21
>>jiggaw+(OP)
It will soon come back from over the other horizon. :)

Of all the things insane about this proposal, I'm not very bothered about this one. It could be high availability and distributed by default. Like having redundant datacenters with eventual consistency on all continents. Except the continents are spinning really fast above you...

The animation is wild... 5GW concentrated up there at the top of a field of solar panels - it's not a Starcloud, it's an electric Starfurnace.

◧◩
4. ben_w+U4[view] [source] [discussion] 2025-10-22 12:26:04
>>non-+N
Round-trip to GEO will add 238.7 milliseconds to whatever other infra you have over the last 200 km vertically* and whatever along the ground. It's probably fine for some things, but not for everything.

* while there could, in principle, be no extra infra in the last 200 km vertically, that means someone on the ground is talking directly to GEO. As per similar discussion about big PV space stations beaming power to the ground, your minimum ground spot size for a transmitter this big and this far away is still tens of km, which limits the other parts of your overall system design.

5. esafak+Aj[view] [source] 2025-10-22 13:33:12
>>jiggaw+(OP)
Why can't it be geostationary? Laser communication can get you gigabit speeds today. That would take a month to transmit GPT-5's estimated 280TB training corpus, which is acceptable. Latency does not matter.
replies(1): >>Gazoch+Iy
◧◩
6. Gazoch+Iy[view] [source] [discussion] 2025-10-22 14:41:21
>>esafak+Aj
With geostationary orbit you won't ever get less than 200ms round-trip latency from the ground (at the speed of light).

Fine for some applications, but a massive regression from modern fiber infrastructure and definitely not suitable for everything (just think how slow the modern web is even with 15ms connections to datacenters). There's a reason why Starlink & co are trying to set up communication satellites closer to the ground.

[go to top]