TOKYO--The merger of two ailing semiconductor businesses at Fujitsu and Panasonic, reported but not confirmed last week, will be a nonstarter. It will neither save Japan’s semiconductor industry nor improve Japan’s standing in system-on-chips (SoCs), which outside of Japan is virtually nil.
"Negative one plus negative one makes negative three," Takashi Yunogami, a director of Fine Processing Institute, dryly said in an interview with EE Times here Monday (Feb. 4). Yunogami was critical of Japanese bureaucrats and corporate executives who have been busy in recent years structuring mergers among struggling Japanese companies and calling it “synergy.”
Yunogami, formerly an engineering expert in dry etching at Hitachi, is now author of books on the Japanese semiconductor industry and a lecturer at several universities.
There is mounting evidence to support Yunogami’s argument.
Back in 2000, Japan had five separate chip companies--excluding captives. They were: Toshiba, Hitachi, NEC, Fujitsu and Mitsubishi. Hitachi spun off its semiconductor division, which became Renesas Electronics through subsequent mergers--first with Mitsubishi and then with NEC Electronics. Fujitsu also spun out its chip operation--now Fujitsu Semiconductor. Toshiba is the only company that neither separated its semiconductor business as an independent company nor forged a merger with another Japanese chip company. Twelve years later, Renesas’ revenue is continuing a spectacular downward spiral and Fujitsu’s chip business is beginning to slip. Only Toshiba remains on a growth track.
In any culture, mergers are never easy. But it’s clear that the prolonged process of merging chip operations at Hitachi, Mitsubishi and NEC sucked the life out of Renesas. No "synergy" is evident.
So, if you muttered "Not again," when you heard about the latest speculation about a Fujitsu-Panasonic deal, you are not alone.
This time around, though, the intriguing feature about the report was the absence of Renesas from the current round of rumor. The originally envisioned joint venture was supposed to be a three-way merger among SoC businesses at Fujitsu, Panasonic and Renesas.
As late as last June, Fujitsu President Masami Yamamoto was quoted in the Japanese media saying that Fujitsu still wants a system chip merger with Renesas and Panasonic. Yamamoto said in his interview with the Asahi newspaper that such a merger would only be considered if Renesas carries out an aggressive turnaround plan.
The absence of Renesas in the latest speculation implies two things: 1) Renesas is still far from actually completing its turnaround plan, with no additional bandwidth left to save its already feeble SoC business; 2) Renesas Mobile, Renesas’s mobile SoC subsidiary, sees no advantage in merging with either Fujitsu or Panasonic.
Japanese electronics industry is too wedded to manufacturing. Japan focused on the science of manufacturing and were very successful. Unfortunately, there are diminishing returns to this strategy, especially when standards of living improve and costs rise. When it became clear that creative distruction was necessary for Japanese industry to survive, the Japanese business culture could not allow that to happen.
I don't think that's the primary reason. 30 years ago, easy for Japanese to dissect a German camera or British TV and reproduce and improve on design for cheaper manufacturing. Not so easy these days with micro-scale integrated products. Can't dissect a Samsung TV or Apple Iphone so easily.
I thought Handle Jones put his finger on a core issue this week at the Common Platform event in Silicon Valley when he said:
Japan is hampered by its slow decision-making process. “They have not adapted to a very fast moving world,” he said.
Everything comes in a big circle, it wants to be an IDM - NTTDocomo again. I am wondering how big NTT docomo's market outside Japan? Japan need to face the reality - world is not buying Japanese electronics for various reasons. NTT docomo will not make money if they need to support empty fabs and so much personnel. Just drag down another company if you can't face the reality.
The biggest obstacle for a company to stay vertically integrated is separating out the various groups into separate profit/loss groups. For simple minded (i.e. short term) investors, this allows them to make short term decisions that are almost always tactical rather than strategic. A protectionist policy (whether it's government or a large company) will increase the liklihood of remaining vertically integrated.. The Japanese govt. is less likely to be focussing on the short term.
There is another angle here. It's less about giving the semi operation a chance to "thrive" on its own. It's about saving the parent company! The difference in Japan is that none of the execs wants to be known as the person who killed semiconductors. It happens in two steps: spin out the operations, then in a couple of years it survives on its own or goes softly into the night - like elpida. Hitachi is far healthier without elpida and renesas.
I agree that the most interesting thing is that Renesas is not involved. It could mean that they have found a buyer for Renesas Mobile who will also take the surplus SOC designers as well. Apple is in sore need of baseband technology to embed on their APs. Its either renesas mobile or st-ericcson
Fascinating article Junko, thanks. Extraordinary how the ebb and flow of fortunes in this business has gone - who could have predicted this at the height of the Japan success in the 80s/90s? Is it too simplistic to ask whether education and the importance of innovation is at the heart of this? The Japanese were so good at improving what already existed but seem to struggle to take it further. Discuss...!
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.