Have you actually studied the cost to grade-out a site, pour the necessary footings and build the shack, assemble the tower, cable the tower and the building, and supply electricity to it, compared to the cost to build a satellite and launch the satellite?
I bet that for a given number of rural subscribers over a the same area, the cell towers will always be cheaper, especially when long-term maintenance of the tower sites versus the continued construction and launch of not-physically-maintainable sateliites is factored in.
Consider another analogy... 802.11 wireless technology is constantly being revised. Wired technology is mature. One 48-port switch can cover an area of about 200,000 square feet, about 18,500 square meters, assuming that the area is cabled, and can give 1Gb connectivity to all devices, and if the devices follow the usual utilization patterns, a 10Gb trunk would be more than adequate. That switch with fiber transceiver probably costs $10,000, and if there's a high-end router needed, add another $10,000. Drops and infrastructure costs about $150 per drop assuming that one uses higher-cost plenum-rated cable and quality jacks and faceplates. Let's be pessimistic and add some supplies cost and we end up at $7500. So, a generously-high-end setup, probably far more than needed, is $27500 for equipment and installation, and it will run for more than a decade.
To do wireless in that same area you'll probably need more than twelve access points, and at two drops per AP to use 802.11ac you still need the 48-port switch. Now you need to buy APs. You might not need 24 (ie, use every copper port) but you'll spend $300 per AP to cable them, plus $250 per AP itself. You'll spend $2500 for the controller too. The router doesn't change. So, take the original $27,500 and add the $2500, then add the money for the APs.
Now scale-up to most size organizations, which have many more than 48 users in 200,000 square feet. Your switch needs probably will decrease on a wireless solution, but your AP and controller demands will go up if you want to keep the average user count per AP low enough to make the experience good, and as wireless technology keeps changing you're going to regularly re-buy those APs and somewhat regularly re-buy the controller(s).