Takashi Yunogami, a director of the Fine Processing Institute who was also invited to the Forum as a speaker, said that the culprit is clear. “The pursuit of excessive quality with excessive technology killed Japan.” Yunogami’s pet peeve is that Japan’s DRAM manufacturers who miraculously met mainframe computer companies’ stringent request to deliver DRAM with a 25-year guarantee remained oblivious to profound changes happening in the market place -- a shift to PCs from mainframe computers -- and kept supplying the same 25-year guaranteed DRAMs to PC companies. In contrast, Samsung came up with a lower-cost DRAM with a three-year guarantee “good enough” for PCs. That allowed the Korean giant to eat Japan’s lunch, Yunogami explained.
An executive from MegaChips, a fabless company based in Osaka, said that the Japanese semiconductor industry has been ailing because they have been too focused on chips they design and chips they make. Their failure to look beyond semiconductors is the biggest problem. The survival of the Japanese electronics industry is dependent on whether “we can offer vertically integrated services, which includes everything from algorithms, LSIs and applications,” he noted.
MegaChips has vowed to subscribe to MediaTek’s playbook -- keeping an eye on a fast-changing market with a relentless focus on the development of turnkey solutions. However, he added, “Our customers won’t be typical CE companies. Rather, we hope to go after companies who build housing and facilities/equipment inside the home, for example.” MegaChips hopes to work with customers who neither know anything about semiconductors nor understand the business-changing potential of chips. “Our plan is to offer them integrated turnkey solutions -- everything from chips to apps -- that they can embrace and run with.”
Forum members also discussed pros and cons of the rampant merger trend among Japan’s electronics companies -- often engineered by bureaucrats at Japan’s Ministry of Economy, Trade and Industry (METI) -- as Japan’s industry policy.
Speaking of a new SoC company as a result of the merger of Fujitsu’s Semiconductor group and Panasonic’s SoC team, a Renesas engineer said that he sees neither synergy between the two nor the prospect of the merged company to grow. “There is no imperative to put the two together,” he said.
Many participants at the Forum predicted that the new Fujitsu/Panasonic SoC comapny won’t succeed, speculating that the only purpose of the JV is for Fujitsu to get funding from the state-backed Development Bank of Japan.
None of Japan’s merged ventures -- starting from Elpida to Renesas Technology and Renesas Electronics -- have been successful, partly due to the human factor, said Yunogami. No Japanese engineer is eager to work at a JV, he said.
Yunogami cited his own experience at Elpida, where he later discovered he was the only engineer who volunteered to move to Elpida. Everyone else had begged to stay at his respective parent company. Being transferred to a JV is a traumatic experience for Japanese employees because it amounts to a loss of personal identity, which is closely linked to the original company for which they vowed to work lifetime.
Japanese engineers, however, will soon find out that they have no choice but to shed this mindset and embrace a new -- adapt or be unemployed -- reality.
The Renesas engineer at the Forum noted, for example that he’s thinking about negotiating a spin-off business based on the IPs he and his team worked on at Renesas.
VeriSilicon CEO Dai asked him, “Are you willing to become a CEO and run that startup?” Well shy of uttering a resounding affirmation of his incipient entrepreneurism, the Japanese engineer said, “Well, I’ve started thinking about that.”
While heated debates enthused on various topics during the Forum and even after that, the following are five-point “predictions” the Japanese Semiconductor Executive Forum made before the meeting adjourned.
I worked as a Principal Engineer in the late 80's through the early 90's at a fairly large company that rather quickly went from the industry's darling to a hopeless basket case. That was Digital Equipment Corporation (DEC).
I learned and gained experience from looking at projects from a systems perspective. After DEC failed to perform to expectations and then evaporated, I looked at that experience too, from a systems perspective.
It's on a bigger scale but I believe also that the performance of a national industry must also be seen as part of a bigger system. The article here referenced decisions by government agencies including banks.
The article raised organization, vertical as opposed distributed, etc.
I was most interested in reading about how Japanese engineers identify themselves with a company, in which they fully expected to devote their entire careers. So even that, like the rest of the topics covered, strikes me as something that the US and European electronics industries have already gone through.
My own "paradigm shift" when this was going on was that engineers now, in the US, are seen by their fellow engineers much more like free agents. It's your name and your reputation that matters most and that precedes you, not so much the company you happen to work for today.
I remember at the start of my career, that a more seasoned engineer told me, "You don't owe any loyalty to the company." It sounded a bit harsh at the time, but with the big layoffs in the early to mid 1990s, I could see exactly what he meant. You owe loyalty to yourself first and foremost. The company can and will lay you off at the drop of a hat, if the bottom line doesn't add up right.
Probably a dumb thing to say, but was there any discussion about the "overwork" culture in Japan? Having visited there a few times, I could never come to terms that tired Japanese engineers were staying in office, day after day, not going home, rarely spending time with their families - at times simply because their bosses were at office.
On an individual level, can not such pressure sap the mind of creating innovative ideas and the pleasure of working on technology?
While I agree that engineers might and should see themselves as free agents, reality in the semiconductor industry is that they will work for a mid-size to large company. The economics of making chips do not lend themselves to true free agency, i.e., being a consultant for one's own firm. Please let me know how I'm wrong on this b/c I'd love to throw up my own shingle.
Japan has pretty much lost the huge Consumer Electronics market in the US & W. Europe to So. Korea, Taiwan and China. If unlike Europe, Japan does not want to let go of this low margin but high volume segment, then they have to compete with So. Korea in terms of NPI, with China on the basis of price. The article mentioned something about sticking to 20 year reliability even for DRAM for throwaway Consumer products ! Thats a no no ! The Japanese must re-examine all their assumptions and quickly change course accordingly. This means faster development cycles ( more science & simulation and fewer experiments ), more robotics ( even if it causes more unemployment among factory workers - at least there will be more money from exports to retrain them in something else ). Otherwise I am afraid it would be 1945 all over again. I say this as a well wisher who has managed semiconductor projects in Japan.
Chip's comments illustrate a couple points I've been trying to make. One is that the question is outside just engineering decisions, it involves economics and ITS LAWS. And these laws defy national boundaries.
Also his remarks display a certain desperate helplessness. All 'must' do this, that or the other thing. Says who? Nobody's in control. It's Smith's 'Invisible Hand' gone amok.
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.