With GlobalFoundries bringing up its Fab 8 in Malta, New York, and their long mutual history, it would
make a great deal of sense to bundle up a lot of IBM's researchers, R&D,
patents and its 300-mm wafer fab at East Fishkill and hand them over.
One sticking point might be the price.
Lenovo paid about $1.75 billion for IBM's PC business back in 2005. But
at how much would you value IBM's chip business and semiconductor
research? Is $1 billion too much? Is $2 billion not enough?
Two other sticking points might be U.S. national security and national pride.
its investment decisions and with public authority support IBM has helped
revitalize the northeast of the United States and New York State in particular. It has
helped to bring GlobalFoundries, Sematech, and the Global 450
Consortium to come and work in New York. Only last September IBM pledged to
contribute $3.6 billion to a $4.4 billion five-year spend on the future
of chipmaking, backed by the state of New York.
While IBM must do the right thing by shareholders,
it is in a position to co-operate with United States' strategic
objectives, be they economic or related to the strategic ability to
manufacture leading-edge chips. If IBM's chip business fell under the control of Abu Dhabi
that would no longer be certain.
I do know that IBM is a foundry supplier of chips. But it is in fact relatively small lying just behind TowerJazz in a 2011 ranking.
As to its role producing processors for games consoles I could argue that is largely historical and will diminish going forward.
IBM will never sell Chip business 100%. It is our National interest. Which country we can trust. These business people can say anything but we have $1T defense budget. Do you think we should sell DoD because it is not profitable? Oracle will not buy Sun. Why Sun? Ask these questions.
Just more idle speculation from poseurs in a country ( ARM included ) with not much skin or credibility in semiconductor technology or business.
For the last 10 years or so IBM's tactic of amortizing its semiconductor R&D expenses by licensing its process technology first to AMD and then newbies like Samsung & the Foundries that cater to ARM licensees has been a pain in the neck for Intel. If IBM was really in the market to cash out of its semiconductor activity and its still considerable R&D and IP goodies then Intel would probably outbid Abu Dhabi to acquire it. Would be more effective than even finFET and the US Govt. would be happy to see IBM technology stay within the US ( Intel still builds its leading Fabs in the US in order to slow down IP theft ).
I agree that such a deal could face too many political hurdles to be viable given the US gov't investments in New York and IBM's military business, but IBM might surprise us and make the rumors come true.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.