Comment by TimorousBestie
17 hours ago
Starlink could do laser comms with other satellites, probably, assuming they shared positioning information and whatever proprietary laser sauce is keeping their intraconstellation links up.
But their radar link to ground is Ku band, which is already getting congested (partly by Starlink itself!). That’s why everyone’s talking about moving to Ka, despite the worse attenuation and higher cost. Much more bandwidth is available at Ka band, as well.
Edit: I’m slightly out of date, Starlink now also operates at least partially in the Ka band.
Starlink has already demonstrated the capability to do laser communication with a maneuvering spacecraft in a very different orbit from the constellation, during the Polaris Dawn mission. And Starlink has way more than enough downlink capacity for the "staggering" numbers in the article. 85 TB/day is a drop in the bucket compared to the traffic Starlink sees. Starlink will need to expand their downlink capacity in the future, but not because of demand coming from space. Maybe unless space GPUs actually happen.
Yes, but I was referring to Starlink potentially interoperating with a satellite that they do not own or operate. To my knowledge they have not done this yet, but clearly my Starlink knowledge is a little out of date so if you know of a mission that did this I’d be interested.
> And Starlink has way more than enough downlink capacity for the "staggering" numbers in the article.
This is the only independent analysis of Starlink’s downlink capacity that I know of: https://thexlab.org/wp-content/uploads/2025/07/Starlink_Anal... (July 2025).
Of course, the new V2’s and V3’s will change the math here considerably.
Starlink is primarily trying to act like a "bent pipe", with the downlink being physically close to the end user terminal in order to reduce both latency and bandwidth requirements. This means having to place your downlink terminals in areas with lots of humans - which of course means the aether is going to be a bit crowded.
On the other hand, an earth observation satellite doesn't need a sub-second-latency link. If you're relaying data via an earth-covering network, nothing is stopping you from placing a bunch of downlink terminals in remote areas of Australia. No need to fight over frequencies when there's at most a single-digit number of humans in a 100km radius!