But here we are 12-15 years later and it's still not operational.
This would require it being safe for and having facilities for crew. Tool and machinery for service and recovery. The vessel would need to handle re-entry stress with return payloads on board. The heat tiles would need to support the extra mass. The vehicle capture system. Etc. Etc.
Could it all be built on the starship "platform"? Maybe.. One day. But that day is not this day.
I don't see this as "space shuttle done right" as a vehicle or "program" without the ability to carry out maintenance and recovery missions. As expensive as it was, for various reasons, it was actually used for those missions.
For example I don't think "space data centers" are going to work I'd the upgrade path for the hardware is vaporize is all and replace it vs recover and recycle/resell.
The more cost effective approach is now repair-in-place. Startups in early orbital testing include Starfish Space (autonomous satellite life extension, orbital repositioning/inclination changes, and active debris disposal) and Katalyst (Multi-arm robotic capture, hardware upgrades, and orbital repositioning).
Old-space satellites are expensive, so on-orbit repair theoretically makes sense here. The largest obstacle to cost-effective repair in that domain is the one-off nature of these vehicles. This means each repair mission requires R&D, manufacturing, and testing before launch. So for this method to be effective, the repair mission needs time before any faults become lethal. I'm not super confident about resuscitation either; off-nominal thermal/power can cook electronics beyond repair. Plus, any RPO mission risks accidental damage.
And on the other hand, new-age satellites like Starlink are build cheap(er) and designed for replacement. Why bother with repair when you can just build and launch another?
It seems like you're also a satellite repair nerd, so I'd love to hear any thoughts you might have :D
As for whether it's worth it, it depends a lot on the satellite and the orbit. Starlink: probably not worth upgrading. Geostationary satellites: there's a reason why what goes up there commands oldspace prices (long way up, intensive radiation environment) and that reason also justifies serviceability. Hypothetical LEO datacentre with most of the mass being solar panels and radiators that last a lot longer than the chips are commercially relevant: yeah, probably makes sense to have the chip payload swappable and the thrusters refuellable.
A mission for one satellite to repair a lot of others is complex (it's basically an extra-hard version of the travelling salesman problem) but when the alternative is inserting many entirely new satellites into n orbits, sending a tiny fraction of the mass to n orbits via a reusable servicer starts to make a lot of sense...
If starship launches ever get anywhere near the 2m-10m claim.. Suddenly fixing or retrieving that 500m-2bn satellite. Or rotating the compute modules on that 150bn space datacenter constellation..
Honestly I could envision regulation requiring it at some point.
The cost to launch may get to the point that it is cheaper to replace than repair, and we can build 10 of a satellite instead of 1 bespoke unit per program.
I see a need for maintenance crews and large payload recovery in the future.
Starship doesn't seem like the tool for that, and no tool is in development for it TMK?