That's not truth, that is your biased assumption with no data to back it up.
Many people try to model three different outcomes before they make a decision: Worst case, best case, and expected case. Something is a "risk" if the worst case is negative. If the "expected" case is negative, it is not a risk, it is just plain stupid to do it. Your claim is that the "worst" case is not or very slightly negative. Maybe Apple has actually modelled the outcome and came to a result where the "expected" case is not very good. And, as I said before, Apple wouldn't look at the cost and revenue of such a model in isolation, it would consider all effects, like the effect of cheapening the Apple brand, risk of cannibalising its own sales in other models, lost opportunity cost (those developers creating a mid range tower can't work on improving the MacMini), and so on.
Compared to this, the Cube had the positive effect of improving Apple's image, even though it was a commercial failure. There are many people who would have loved to own one, but who believed that it was too expensive to them, especially since it was released just at the start at a recession that was worse than the current one. The positive thing is the "would have loved to own one". Apple sold lots of Macs _because of_ the Cube.
PS and more on topic: MacObserver reports that Apple added claims under the DMCA act against Psystar, which means we might be going from a civil case to criminal charges. A quote from theregister: If the Apple/Psystar brawl were a prizefight, however, the referee would now step in and stop the mayhem before Psystar suffers terminal brain damage.