I agree with everything except the first sentence, as that seems to contradict everything else you've said.
Given everything else is true (which I think it is), how do you conclude that BFL is still competitive (even assuming they deliver on time and on spec, which I will wager heavily against them doing.)?
I think they are competitive in a narrow sense... usually pricing (except in these last days of their 65nm product) & maybe in the area of keeping customers on the hook (not that that is a good thing - but it is a kind of fighting spirit). This might be a different kind of competitiveness than you are thinking... there might be a better word for it. Anyway, historically, they have had the best (most efficient) products. Here are a few examples of their competitiveness:
1. BFL's FPGA Singles were the best price for GH of all sources.
2. When Avalon came out with their better pricing than BFL. Shortly thereafter, BFL matched the GH/$. It was a significant change to their price:performance.
3. Of the several companies promising 28nm products, BFL is second in price + performance.
But despite that, I don't think they will be able to keep up with Cointerra. BFL will probably end up in their shadow.
HashFast(HF) also has my attention, their pricing is about double BFL, but HF may be one of the first to deliver.
Oh right, yes, perhaps a slight misunderstanding then. Historically they have been 100% competitive, or rather, the total leader (IMHO) for pretty much the whole post-GPU era. And I also completely agree, in terms of keeping people on the hook (or stringing them along) they are uniquely skilled!
CoinTerra has it all though, and BFL will be crushed pretty soon, I'm sure. The only way they will remain in anyone's shadow is through their ability to string people along. CoinTerra have a working prototype and BFL haven't even taped out yet. CoinTerra have all the experience, BFL have a little, but a history of fuck-ups. It's only the fact that they're getting people to commit now that they will stay in business.
It's an underhanded tactic, but an understandable one. It's a weird situation where people
have to offer preorders or they'll lose market share, and it's a viscious cycle that's keeping it this way. I don't blame BFL for trying to keep their customers with great promises which they'll never be able to deliver on time and on spec/budget, for one simple reason:
fool me once, shame on you; fool me twice, shame on me. BFL customers who switch now can't claim not to have known who they were getting in bed with this time...