Even though it employs 250,000 people directly and has the highest level of R&D as a percentage of sales of any major US sector, the US semiconductor industry appears to believe that further support is needed.
According to a press release announcing the formation of a Semiconductor Working Group as part of the President’s Council of Advisors on Science and Technology, ‘the time is right for a fresh look at the policy issues shaping innovation and global competition in the semiconductor industry’.
The group will identify the core challenges facing the semiconductor industry, then deliver a set of recommendations for the Federal government, industry and academia.
Interestingly, the working group – co chaired by ex Intel boss Paul Otellini – believes the cost of designing and making chips is stifling innovation and says the industry may soon be unable to keep up with Moore’s Law.
Intel is certainly feeling the pace; its legendary ‘tick-tock’ process development cycle – representing a microarchitecture update and process shrink – has slowed significantly. Yet TSMC and Samsung, for example, are both readying 7nm processes and beyond.
It’s been a long time since device manufacturers, with the exception of Intel, developed their own process technologies. A decade or more ago, before it got out of electronics, IBM was a major driver of process technology, supported by a range of companies.
The partners shared the cost of developing a process which they could then use to make differentiated products. On this side of the ‘pond’, the ill fated Crolles2 Alliance had a similar outlook.
Since then, GlobalFoundries has picked up the baton, leading advanced technology development in New York State with the support of the State goverment. Would Federal money speed progress?
But maybe the semiconductor industry is changing. Not so long ago, you knew exactly what Silicon Valley was up to – leading edge semiconductors. Today, it seems to be more interested in anything else.
Maybe it is an investment related issue. During the dot com boom, VCs came to blows as they literally fought to invest in start ups. Today, the legendary Silicon Valley VCs must be thinking twice about whether to put their money into semiconductors – those developing leading edge chips will need to spend more than $200million just to get the part to market. But what role could the Federal government play there?
What will the group recommend? Who knows? But it hints that public money should be put towards creating technologies that explore new computer architectures and enable new applications.
R&D was once done in house. Now, with their focus firmly on short term ‘D’, chip developers rely on universities to undertake the longer term ‘R’ function – and they appear to be well funded.
We await the group’s findings with interest and hope it’s not a recommendation to ‘circle the wagons’.
