SEVILLE, Spain – The Japanese semiconductor industry, which has been in decline for more than 20 years, needs to select key technology areas and refocus to rise again, Junshi (JJ) Yamaguchi, former chairman of chip company Renesas Electronics Corp., told the International Electronics Forum, being organized here.
Yamaguchi, who now serves as special advisor to Renesas and as chairman of the Japan Semiconductor Industry Association, indicated that the tragedy of the great Japanese earthquake and tsunami of March 2011 could yet prove to be a turning a point of for the country and its remaining chip manufacturers.
He opened his presentation by painting a picture of the decline of the Japanese chip industry as a percentage of the global output. This went from 51 percent in 1988 to 44 percent in 1994 to 29 percent in 1998 and to just 20 percent in 2010, he said.
Yamaguchi explained the key events that prompted the decline.
In the 1970s Japan had decided to invest in DRAM production. Increasing sales of personal computers in the 1980s drove DRAM consumption resulting in high volumes and revenues for numerous Japanese chip companies.
The turning point was political intervention by the United States that led to the signing in 1988 of U.S.–Japan Semiconductor Agreement in 1988. Under the agreement Japan agreed to not to sell DRAMs at below cost and also to raise the U.S. market share in the Japanese chip market from the prevailing 10 percent to 20 percent. "Limits on Japanese capex and imports helped create South Korea as a force [in DRAMs and semiconductors]," said Yamaguchi.
However, Japan continued to be too conservative Yamaguchi said. Its problems continued when, having been largely forced out of DRAMs it stayed focused on ASIC production for too long. "In the 2000s there was a strong market for SoCs; that is ASSPs rather than ASICs." During this time there was rapid growth in foundry.
Well, just like some of the US execs have been saying, if Japan can bring back value-added manufacturing (that has slowly bled to China), some of the trends can be reversed. It should also focus on R&D-heavy and NRE-focused projects that it can market worldwide.
Certainly mass-marked high volume manufacturing (with low margins) is not going to be one to replicate. I would like to see how the country's leaders and policitians deal with this in the coming years.
I write the following as a Semiconductor technologist who has visited Japan ( 30 times so far ) over the last 20+ years on behalf of major US IDMs and have enjoyed and admired the Japanese for their culture and my colleagues there for their sincerity & work ethic.
Given the political and business realities described above it would be tough for the Japanese to rise again and be competitive w/ Taiwan or So. Korea in Semi manufacturing - at least so far as selling to the US is concerned. The Japanese should instead look for greener pastures like Europe and Russia. Germany beckons as a lucrative market for embedded systems as Infineon continues to fade. Russia can not only provide strengths in Pure Mathematics ( algorithms, architecture ) & Physics ( Solid State, Plasma ) where Japan has been a little weak but also a decent & stable market for the latest Japanese processors and Semi fab tools ( Tokyo Electron etc. ) and operation for the Fab they are aspiring to set up outside Moscow.
In spite of its recovery since WW II, Japan still remains just a vassal state of the US ( having been administered severe " radiation therapy " in 1945 ). Japanese industry really got the short end of the stick and could n't CHEAT their way out as nuclear - armed China has been able to do with impunity thanks to the connivance of Wall St. and US MNCs lured by ever elusive mega-profits in the PRC and in the process given the store away ( e,g. Motorola, GM ).
During the '80s when the US still cared about manufacturing, Reagan was able to force down Japan's throat several anti-competitive measures ( autos, semiconductors ). Heck the whole SemaTech was set up in Austin ( thx to the influence of dad Bush who was the VP then ) to help US semiconductor industry to catch up with the Japanese who had always been blamed by the incompetent slackers here of copying US technology ! Because the problem in US Fabs always was in trying to press-gang EEs who had failed at Design to take up materials processing ( till Craig B. with his PhD in MSE and time at Stanford showed Intel fabs how it MUST be done ). That lesson escaped SemaTech and quite predictably they too failed but in the process allowed droves of eager Taiwanese PhDs to pick up Fab skills at SemaTech. They are the secret to ex - TI Morris Chang's success with TSMC.
In my travels I have also seen Japanese companies who had developed technologies quite independent of US IDMs, being forced to give them away to the Taiwanese and So. Koreans just to keep some US business. These technologies sooner or later find their way into China to who IP remains an alien concept.
It may sound ironic but unlike the still very much Communist PRC, our democratic ally Japan does not have open access to Academia and Labs in the US from which technology and capital has been haemorrhaging out to China at the alarming effective rate of $ 1 trillion a year for the last 15 years !
In the era of dot-com bloom, companies in the Valley has a number of cars in the parking lot 24/7. I didn't believe all employees were forced to work 24/7. Rather, I think they have the passion to stay up all night to finish the job, to make the company thrive. Only if we work together, we will once again thrive. I believe Japan can do is. So can United States of America.
Cultural and languages differences could be a barrier. The responsiveness of chip vendors is depending on the corporate (vendor) direction and the potential volume of sales as always. Money has a lot of weight in this world.
It is very difficult to deal with Japanese manufacturer. For freelancer or medium size organizations, they do not respond well. They do not provide quick response and withhold critical design details. It is very happy exeprience using American organizations. They are transperant and provide very good support. Japanese manufacturer should also change their approach to small and medium scale industry. This may help them lot.
"what can be achieved when people come together on a mission and are prepared to work weekdays and weekends until the goal achieved." I hope this is not the path for Japanese to shine again in the semiconductor industry. Innovation doesn't mean 24-7 working attitude! BTW, the systematic way of doing things and the pursuit of high quality makes Japanese excel in many areas. I still think Japanese can keep on building good stuff in MCU and new energy products. Communication is also a very interesting market that Japanese shouldn't have given up. With the emerging new telecommunication development, Japanese should try harder to catch up.
Hmm, they did loose market share but I don't think they would have done great it they had still retained their DRAM foundries. Thats a mass commodity now and only can generate some profits in voulme shipments and has and will consolidate further unless some other technlogies in the sidelines emerge from their shadows.
The japs have and are doing quite good in chip manufacturing in general and are par none when it comes to optical processing chips. So it is not as bleak as it is painted, sure Sony and Panasonic have kind of lost their sheen but thats just a lull I think which happens in any business. Toshiba seems to be on the way again in LCD domain, so theirs a lot to come and it's pointless just comparing market share.
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.