I think the point of OP, and of the other magazine article linked in the comments, is that he hasn't 'grossly underestimated' it in the least bit: yes, you can pay peanuts for Iridium and run an already built Iridium network on those use-cases like the military and occasional people going to war zones and NGOs and hikers, and make the business work that way. But you cannot build one costing billions of dollars based on those use-cases, which is why Iridium went bankrupt in the first place - their business plan was based on covering way more than lost hikers and Afghanistan, it was based on the delusional ideas that businessmen in Paris would pay for it. This was delusional, so Iridium went bankrupt, defaulted on its debts, cost its original investors fortunes, and had to be rescued out of bankruptcy for $25m and refocused on the use-cases which actually did work.
There's also a decent rental market for these phones, so one doesn't have to shell out thousands of dollars--you can get one for $50-75/week + airtime, which is a fairly minor expense if you are already going somewhere so remote you need to be flown in/out.
Cellular phones aren't a substitute, because satellite phones are for precisely where there is no cellular coverage, like less developed countries, remote terrain, and war zones.
Nobody expects people in the street to buy them. I would image most customers are military, other governmental and NGOs, maybe some private companies working in remote places, and a few people with private yachts or who like hiking in very remote places.
For all those customers and use cases, the price for the handset and cost per minute is reasonable. Mainly because there's no substitute, but also because these users are big spenders anyway.
And it doesn't matter that they don't work in buildings, because of course you're normally using these in places without any buildings.