Fujitsu’s decision to spin off its semiconductor became inevitable as its options for future technology development and manufacturing capability started to close.
Fujitsu’s semiconductor business is too small to afford either technology development, or independent manufacturing.
At the 90/65nm generation, when its CEO was a former process engineer, Toshihiko Ono, the company developed a particularly low-leakage logic process and then managed to get outside partners like Lattice Semiconductor to contribute towards a fab in return for wafers.
Having made a success of the semiconductor business, Ono moved to a corporate role with the parent company taking responsibility for future product definition.
Unfortunately the same strategy could not be followed for the next process generations.
The word in Tokyo last autumn was that the company is late in defining and choosing the materials for its 32nm node, and would be unlikely to be able to repeat the foundry model at a new process generation.
Ono’s successor as CEO of the semiconductor business is a guy with an EDA background.
The Japanese government had long pushed for consolidation. A government-pushed scheme for a joint industry fab foundered on disagreement about whose design rules to support.
Fujitsu’s natural partners were NEC and Toshiba. By far the strongest of the three is Toshiba which has chosen to go to IBM for technology development and to share manufacturing capacity with NEC but not, apparently, with Fujitsu.
With Renesas throwing in its lot for both technology development and manufacturing with Matsushita, that left Fujitsu out on a limb.
The question then was: Does Fujitsu see its semiconductor business as sufficiently important to its computer business to be worth keeping on as a subsidized operation?
The answer to that appears to be: No.
Now, it seems, Fujitsu’s semiconductor division is to be spun off as a separate business entity by March 2008.
Subscribe to this blog's feed