To: Hightechhooper who wrote (13013 ) 6/13/1998 7:38:00 PM From: Grand Poobah Read Replies (2) | Respond to of 25814
K, Sorry if I'm getting into this discussion a little bit late... Re:Design expertise can be bought or cloned much easier than most people want to believe. Not true. It is very difficult for companies to develop, accumulate, or buy and integrate design expertise. There is much, much more hue and cry in the media about companies moving into new markets than there ever are good product offerings that follow from the announcements. Witness INTC's ballyhooed acquisition of Chips and Technologies and how the whole graphics chip industry was supposedly in imminent danger of INTC dominance. Now a few months or a year later we are hearing how INTC's i740 graphics chip just isn't up to snuff with the competition and they are having to dump inventory. If INTC can't just buy into a new market, who can? Also witness all the Japanese DRAM makers who talked so big about getting out of the low-margin DRAM business and jumping into SOC (or system LSI as they call it over there). Now we hear that that is not so easily accomplished (something about their problems was posted on this thread, I think). It is true that there are successful acquisitions and that design groups leave companies and make successful new startups, but the failures of those kind of ventures greatly outnumber the successes. There is just too much involved these days in making a successful new design. I posted a summary a while back on what was involved in the IC design process, so I won't rehash that, but suffice it to say that the company infrastructure and R&D processes have a huge impact on the success of any design. Contrary to your claim, I would say that design of today's 500k gate designs is not as easy or portable as it is often made out to be in the press.Process technology is the real differentiator because it is more art and less science than most people believe (that is why INTC kicks butt, not design but manufacturing). Here I would agree with you but would perhaps put a different spin on it. Process technology is more engineering and less science than most people believe. I have seen a lot of process engineers make stupid decisions because they didn't know how to make good engineering decisions. Processing is really not that mystical. You get yourself into trouble when you are running the process because it works and you don't know why. What happens then when something in your process drifts and you start losing yield? You have to go back and do the engineering you should have done in the first place. I have seen this happen more times than I wish. INTC kicks butt because they are a good engineering organization from top to bottom in everything they do, whether it is processing, design, or marketing. Sometimes they can be a little arrogant about it and they certainly don't like to admit when they've made a mistake, but despite that I have enjoyed the times I've worked with them because you know they will do a good job of engineering. I think LSI is also good in this regard, but INTC of course is at the head of the class. G.P.