Radically cheaper launches will still be significantly more expensive than current maintenance contracts. Even if we have infinitely reusable rockets in the fut

Radically cheaper launches will still be significantly more expensive than current maintenance contracts. Even if we have infinitely reusable rockets in the future, you're still looking at millions of $ in fuel costs for a servicing trip. I am personally not as optimistic about "soon". If you send a technician out to the Earth-Sun L2 point (where JWST will be) so that she can throw some grease on a gear or spray WD-40 on a reaction wheel, she'll be 4 times further from Earth than any human has ever gone. Although there may be some synergy with future Lunar/Martian missions, there's a lot of work that needs to be done to figure out how to do that safely. I think L2 is the safest Lagrange point w.r.t radiation, but it's still potentially deadly. >"there is no need to make a big structure capable withstanding 1g gravity". While the overall structure may not need to withstand 1g continuously, you still need to design it to as much as 6 Gs of acceleration during launch (as well as significant lateral acceleration, and occasional shocks in the dozens or hundreds of gs).