Hiring another staff member to create a second piece of DLC doubles all of that. A third will triple it, etc.
Basically like ThatMarkGuy says, that (very theoretical, as you said) concept only works if you ignore REVENUE. If we released four songs instead of two, but our sales per week didn't double, our margin is now worse or in the red, right? If Rock Band stopped, no one would be saying, "But dang, those few months of higher output felt good, glad that happened," they'd be going, "Boo. No more? Boo." Aaron said a few times (mostly around the Amp Kickstarter) that Rock Band on 360/PS3 basically just paid for itself. Consider everything you know about the popularity, how many of your friends were playing it, how much DLC was sold. And it was still run to the equivalent of paycheck-to-paycheck. You presumably do not live paycheck-to-paycheck, you save where you can. Because living paycheck-to-paycheck is dangerous.
We've got over a decade of metrics, sales, patterns, habits to draw from. We can see where the money is. We can see where the paradox of choice takes over. I've got spreadsheets for lifetime RBN sales, a spreadsheet of songs that haven't made their costs back, a spreadsheet of RB4 DLC sales (that's like two months out of date, really need a new one...), I invited myself to take over the rockband.com/request page to cross that data with sales patterns. I'm not even on the finance team, I quietly ran around and collected all this so I could make arguments for whats and hows and whens for DLC because I've felt Rock Band DLC is like, the coolest thing in games since I was 14. But I ain't winning any in-house debates with, "I really wanna!" so I gotta put in the work.
Rock Band is making more than it is spending, which makes it a smart business decision to keep supporting.