If it wasn't a vise investment, publishers and developers of AAA titles wouldn't invest in it. But they do.
I would suspect that it probably offers some reduction in casual piracy around release time (as compared to a game with no DRM that you can install anywhere you want as many times as you want with no need to work around anything), though that's probably hard to measure since there's no control group for AAA titles - most of them have some sort of DRM.
The downside is it probably causes significant damage to customer goodwill and willingness to buy future titles, but from the companies' perspective, that's both really hard to measure and it applies to financial results for future quarters, not this quarter, so it might as well not exist.
I agree that it's overwhelmingly likely that someone ran the numbers about using DRM at these companies. But it's not unlikely that they over- or under-estimated certain factors. They will have more information than we do, but certainly not perfect information (see also above).
I wouldn't even go that far. It's very rare for anybody to run the numbers. Humans are too lazy. At most they'll eyeball it. Unless somebody requires hard numbers and they start digging up your process so that it's easier to actually do it than to fake it, no "running the numbers" will happen.
None have done so. Ergo, I am quite sure no one has "run the numbers"; they have, as mentioned, eyeballed them and used them to justify their own biases, as humans are wont to do. And either do not have enough of an analysis to release, or realize it is so easy to poke holes in that they shouldn't.
These numbers are, of course, wildly optimistic but it’s harder to say exactly how far off they are.
Fundamentally, they're spending other people's money as insurance to avoid losing their job following a flop that can be spun as due to piracy.