This.  If Bitcoin never gets any bigger then it doesn't really matter what ASICMiner share is but if/when Bitcoin rises significantly in valuation the global mining revenue will bring in new players with deeper pockets.  Global mining revenue is ~$130M annually today.  That is the pie that ASIC developers are fighting over.  If they build hardware to mine themselves or build hardware to sell it doesn't really matter.  The pool of potential profit is $130M annually. 

Now like I said if Bitcoin doesn't get larger then all this is just a tempest in a teacup.  Bitcoin will be an experiment that limps along and eventually is replaced by something else.  Under that scenario it doesn't matter if ASICMiner goes bankrupt or runs 100% of the network.  However lets say Bitcoin increases by an order of magnitude in value.  Suddenly ASIC developers are fighting over $1B potential revenue.  That is going to bring in companies with deeper pockets.  While ASICMiner can maintain both high ROI% and high marketshare fighting the incompetent competition of today it is improbable they can against better run and capitalized companies.  Forget AMD/Intel there are hundreds of companies which make specialized ASICs for all types of industries who already have the expertise, staff, and industry contracts to roll out SHA256 ASICS.  However a slice of $100M at very high risk is simply not something they are interested in yet.  $1B+ well that is a different story and if Bitcoin is still around in a year or two the risk of it failing will be lower so higher reward and lower risk = more competitors.