1) Hospitals generally get paid (by someone) for treating patients. So they want to run at 100% capacity, even with covid patients.
2) You can't really just "build an ICU." You can stand up a field hospital that's able to provide similar, degraded services. But it's not going to be the same.
Imagine if someone said "set up a datacenter in a field," and then a customer had a choice to colo their server in a purpose-built DC or your field DC. Which do you think they'd pick?
3) The units that staffed these facilities were heavily augmented by military reserves.
Corpsmen may have non-medical civilian jobs, but what do you think military reserve critical care nurses and doctors do for a living?
If these field hospitals were packed to the brim, all of those people would be locked in place. Possibly while they were needed in their civilian positions, in their home cities (although the military seems to have done a pretty good job about preferentially pulling non-Covid-civilian personnel).
4) Exponential growth with unknown but high R_t. +1 week before we locked down would have looked very different.
... all of that to essentially say that these facilities were designed as surge capacity, in the event they were needed. We intentionally filled up hospitals first.
USD$660M seems like a reasonable price to pay, strictly as insurance, given the very real risk that hospitals would have been overfilled and covid-related mortality would have spiked due to lack of care.
I think this is under-appreciated. In contrast to the $0.6 billion spent on field hospitals, the US spent some $2000 billion on the CARES act to mitigate economic harms of the virus and response to it.
At least some of the reason it didn't treat many patients is that they made it hard to become a patient.
Military hospitals (including ship-based hospitals) have been planning and preparing to deal with biological warfare (i.e., communicable diseases) for many, many decades, and even before that they were well acquainted with communicable diseases. It was not uncommon in the past for communicable diseases to kill more soldiers than died in actual combat.
That is one reason the U.S. Public Health Service is a uniformed service -- so their personnel could operate in war zones without the risk of being shot as spies if they were captured.
Small detail about the navy ships: they are setup for battlefield injuries. The wards are all open (think what you see on old episodes of MASH) with little to no separation between patients. That is fine for traumatic injuries (earthquakes, fires, war etc) but less than useful for control of infectious disease. The plan was for them to take non-covid patients to relieve hospitals but it never really worked out. The risk of a ship becoming a hotspot should one covid infection slip through into an open ward was too great.