Re: ASICMINER: Entering the Future of ASIC Mining by Inventing It
So it's wishful thinking to have next gen ASIC chips on par with Intel's Skylake/Skymont? 
Only if Friedcat wants to disrupt Intel next!
Seriously, it's a very interesting question. At any stage of this arms race competing chip makers have to choose between higher upfront costs and lead times against have a chip with a competitive edge. It's inefficient and risky to go too big too soon. There's no point in being too big a share of the network, so even if you can build something miles better than everyone else perhaps you shouldn't. So I guess it's a question of considering:
- who else is making chips;
- in what volume can they get them;
- what performance will they get with them; and
- when do I think they will hit the network, or will they go broke first
Companies that survive will have to revisit these questions and get good answers every time they commit to a next generation chip.
If ASICMINER builds hashpower that will put it over 50% of the current network speed, it can always sell the excess hashrate to keep below a certain percent of the network. There's always incentive to build something faster with better technology as you raise the barrier to entry and thus can keep the profit margin.
Did the USB stick officially goes on sale? That would generate revenue of 0.05 BTC / share if sold out, a pretty juicy dividend / good reserve for further development.