That's not for lack of trying by those of us in the space robotics community... There's a chicken and egg problem at work here. The robotics technology to do o

That's not for lack of trying by those of us in the space robotics community... There's a chicken and egg problem at work here. The robotics technology to do orbital construction exists, and in fact has existed for a couple of decades now. Nobody wants to fly it because the older technology - which amounts to “spend lots of money to fit your payload in a single rocket fairing” - has worked so far. The James Web Space Telescope will have a 6.5 meter primary optic. It will require something like 70 separate deployments in order to deploy the primary optic and the sunshield. At some point, custom designing hundreds of deployment mechanisms for every single mission becomes both more expensive and higher risk than just flying a damned robot arm. As does “just build a bigger rocket”. FYI, robotic refueling on orbit has been demonstrated a couple of times (Orbital Express and RRM).