I've done some more digging, and found a range of results that suggest average windspeeds at those latitudes and heights between 5 and 20 knots. Which is quite a variation, quite apart from excursions. The UAV manufacturers say "There's no weather up there": I expect if any of them are successful we'll learn a lot more about how much weather there isn't.
As for the feasibility of the whole enterprise, as etudiant says getting the platform up there is one thing, having a useful payload is quite another. I haven't got exact figures, but the radio side of an LTE base station alone takes around 400w (that's the customer-side, ignoring backhaul). LTE cells aren't at FL 600, the Titan Aerospace UAV only generates 100w (and needs all of that to stay aloft) and has a payload of 32kg. Bigger models are planned with higher power budgets, and the basic plan seems to be to have lots and lots of low-power infrastructure rather than a few soaring behemoths, but I rather fear physics (and the lessons of Iridium) are agin them.