These are not good discussion posts:
A question was asked. Platform-biased beliefs have no value in an answer.
Not wild assumptions but shared ideas from people who don't know to people who don't know. That's just normal discussion. If the interest is only in a super-informed answer, a general discussion board like this is no use and the OP should ask suitable accounting experts on Stack Exchange or something.
This shouldn't be viewed as Sony versus MS with some sort of winner, but a mystery, like a murder mystery. Everyone contributing should be happy to share their theories, have them criticised, and be open to alternative theories. If we're lucky, we'll find an answer that everyone likes. More likely we'll share contrasting theories without any real understanding of the truth until we get bored of the conversation and move on.
So the two real questions
1) Is there really a $100+ physical loss on each console sold?
2a) If so, what accounts for that?
2b) If not, how does the accounting work?
Regards 2b, the suggestion R&D cost is spread across a flat prediction of units sold means there'll never be a time when the console is not loss-leading. That's contrasted with XBOne where it was announced the consoles would be break even break a profit:
After months of policy reversals from Microsoft, there's only one major difference between the Xbox One and the PS4 at this point, the price differential, and the subsequent reason for the price differential. The PS4 retails at $399 while the Xbox One is $499, and comes with the Kinect sensor, [...]
Of course, the statements culd be using different numbers in different contexts. Contrasting with Yusuf Mehdi telling us XBO would be proiftable and XB360 was very proftable, we have Lori Wright saying MS has never made a profit on any console hardware sold. I think that's one for the lawyers to argue over!