[jasonwatkinspdx]

Because there's only one L2 point, and it'd be very risky to have two units orbiting. So your spare would just sit around costing money in the case the first one works. The folks that do this are pretty dang good at what they do, so they're willing to bet $10 billion on success the first time, vs $20 billion to hedge with a spare. Even with a spare it's not clear to me a malfunctioning first example would have enough Delta-V to get it safely out of the L2 point proximate.

If it fails, they'll learn all they can, then try again with another follow on project, that likely will take advantage of technological improvements since functional requirements on this one were set in stone. Pre-building a spare just doesn't make sense with this kind of project.