Getting it to work with one end stationary first sounds like a reasonable development plan. LEO adds a lot of complexity, but with huge benefits.
OTOH the number of engineers that focus on throughput over latency is quite staggering.
I guess if your goal is just to stream aircraft telemetry and black box like recordings then latency may not be high on the agenda.
Leo seems easier to me. Geostationary is really far away. Leo is much, much closer. It's easier to hit a buck thats running right past you than to hit a stationary target across the Atlantic.
Especially if you yourself are on a moving platform