Or perhaps you have a little more smarts, and do like Intel and other manufacturers in the electronics business do, or Pfizer and Merck do in the pharmaceutical business, or Dow does in the chemical business - introduce new items at just as much as you calculate the market will bear to recover your R&D costs and maximize your early profits, then drop the price when either the competition catches up, or you introduce a new product line.

So long as BFL Jalepenos aren't being shipped from off-the-shelf, and there is no other competition, Friedcat can set the price at whatever the market will bear and whatever price maximizes his profits -  which he did, and continues to do.  His profit margins are still over 90%, and he's planning to ship 6 million of these at the lower price, as opposed to the 2 million he sold at the higher price.  And when competition shows up, he'll either just cut the price again, since he has economy-of-scale behind him where a competitor wouldn't, or just quit making them completely if he feels the profit margins have gotten too compressed.

And you really don't get the fact that the chips he uses are binned.  The A-grade chips go into his current generation of blades, and, rather than discarding the rejects, he gets to sell the B-grade and lower parts for massive profits since the buyers of these devices don't care about the few % of hardware errors the lower grade chips end up producing.

Your theory is all well and good, but you have obviously never run an actual business before.  Friedcat is as close to being a sales and marketing genius as I have ever seen in the Bitcoin area.