There are more, hardly mentioned reasons to hang on to a none successful CPU business which carries RELATIVE high fixed costs compared to the overall cost of a low revenue company. It is simply to have a practical testing field for their "paper" IP work, to respond faster to mistakes, to optimize based on real live applications, to show the IP really works to prospective licensing partners, etc..
But Hey, who said the CPU business will fail???
s.
Market Data 
Markets 
AI
