All they need is access to some R&D and pilot facilities.
I think you are right. Maybe due to IBM's tie with the Common Platform, that's possible. Then, again, I would still have to come back to the question: What's in it for IBM to continue its semiconductor R&D now?
IBM is a big think-tank and huge IP company and will grow this way into the future. They don't need to maintain their own semiconductor facilities! Way too costly! All they need is access to some R&D and pilot facilities which they do already.
I am wondering... how can they keep the research without access to fab facilities? This goes together in my opinion. I dont know of anyone else in the industry doing similar research without access to a fab.
Furthermore, selling the semiconductors division does it mean they go fabless or that they sell (or outsource) everything chip related? Both are problematic. For many different standpoints.
The biggest question that IBM's CEO would have to answer is if those fat software and services margins will remain as fat in the case of loosing the unique characteristics of the systems IBM provides. What is the main selling point of IBM if they loose the grip that they have on customers because of their proprietary systems?
I think that IBM should work hard on trying to make the most of its hardware division instead of trying to get rid of it. The whole thing is how far you look at the issue: short-term or long-term. For wall street, long term is totally meaningless. For any other business that wants to stay relevant (instead of becoming opportunistic) it is the other way around.
Adam Sichko, a reporter at the Albany Business Review, coincidentally posted a story about changes to the GlobalFoundries' plans for its Technology Development Center. Abu Dhabi will invest 8-9 billion in GlobalFoundries over the next two years, including an increase in the number of tools planned for the spacious TDC. "One labor union, in a newsletter posted online, informed members that 'the TDC building is being redesigned for a new use, so only limited work can be done on that site till the drawings are redone and coordination takes place,'" Sichko reported.
Transferring manufacturing to GlobalFoundries and sharing R&D makes so much sense for all concerned, including the very able R&D teams at both IBM and GlobalFoundries, that I hope it happens.
@ssliva, thanks for pointing out other IBM chip facilities I failed to mention.
I think the genius of this deal, if any deal is ever made, will be how they structure it. As you mentioned, they can sell it in a piece meal manner; or sell just fabs and keep IPs and R&D team; or go all the way in.
I suppose there are many ways to skin the cat, instead of taking an all or nothing approach.
But the question still remains. Even if IBM decides to sell just fabs, what for would they keep the semiconductor R&D then?
IBM has to decide to either invest in a business that it doesn't think has high enough margins to deserve investment, or sell it. They're currently spending hundreds of millions of dollars each year laying off workers in microelectronics, shrinking or folding efforts in foundry/ASIC/EDA that perhaps could be sold, but in the IBM fold they're treated as red-headed stepchildren.
They're not going to find anything they want to invest in, though. Time to sell. Doesn't mean they *won't* let it wither, though. Or sell some of it piecemeal. The biggest sticking point is going to be control of research and IP, but they'll look to Google's sale of Motorola Mobility as an example of selling the business but keeping the patent IP. They can't keep all the inventors, though.
IBM microelectronics also has a 200mm fab in Burlington, VT and packaging/test in Bromont, QC as well as the 300mm fab in East Fishkill mentioned.
My Mom the Radio Star Max MaxfieldPost a comment I've said it before and I'll say it again -- it's a funny old world when you come to think about it. Last Friday lunchtime, for example, I received an email from Tim Levell, the editor for ...
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...