Electronics Weekly Electronics Design & Components Tech News
No government could have strategised the success of Arm Nvidia or x86.
Arm was a fortuitous mix of a small, low power core, a nascent mobile phone industry and Nokia’s determination to keep x86 out of the mobile computing ecosystem.
Nvidia was a case of a gaming chip being taken up as a Bitcoin mining chip then an AI training chip. Even Jensen would never dare to claim it was all part of a ‘strategy’.
x86 and Intel achieved success because IBM chose the architecture from among several others – with chance again delivering success. Even Intel didn’t see the IBM design win as their most significant that year.
Another example of an accidental empire was Flash memory which was invented by a Toshiba researcher but Toshiba had no intention of commercialising it until Intel looked at it and thought it a good way to go.
Micron’s founding was the result of Inmos poaching the Mostek DRAM design team, followed by Mostek suing to get them back. One of the team had a college mate whose Dad in Idaho had got rich selling to MacDonald’s a process for freeze-drying potato chips and Dad set them up with a fab in Boise.
Semiconductor successes have usually had another ingredient which can’t be found by strategising – outstanding leaders. Bob Noyce, Gordon Moore, Andy Grove, Pat Haggerty, Mark Shepherd, Charlie Sporck, Jerry Sanders, Jack Gifford, Ray Stata, Morris Chang, Ward Parkinson, Jen Hsun Huang, Tsugio Makimoto, Tsuyoshi Kawanishi, Rob Proebsting, Paul Schroeder, Ming-Kai Tsai, Robin Saxby, George Perlegos, C.C.Wei – all bold, free-thinking, original leaders.
All government can do is create the conditions in schools, universities, research institutes and startups where semiconductor development is encouraged by grants, tax-breaks, loans, access to capital etc.
This is a lot cheaper than subsidising foreign firms to build fabs in a country wanting a ‘strategy’ while these fabs can be abandoned when markets turn – as the UK found after subsidising foreign-owned memory fabs in Scotland, the Northern England and Wales.
We also learned that subsidising foreign fabs didn’t do much help the development of a local UK chip industry.
A ‘Semiconductor Strategy’ for an industry which relies on a succession of brilliant ideas, luck, chance, continuous scientific breakthroughs, outstanding leaders and genius insights is a curiously inappropriate approach.
I totally agree with David. May creative anarchy in the industry long continue!
In addition governments do not understand the sheer diversity in the semiconductor industry and the global interdependence amongst the supply chain. Having a world beating processor or AI chip is meaningless if the ancillary chips and devices are not available and no one country will ever be self sufficient in semiconductors and able to supply everything, whatever politicians might think.
Totally agree.
The one thing governments can do is lay the foundations of an ecosystem. Taiwan did this with ITRI, the UK with Inmos and Flanders with Imec. Give them enough money, then stand back, don’t interfere and watch them hopefully trying to change the world.
I second that.
Mike Bryant,
the main problem with that is that governments are TERRIBLE at standing back and not interfering! The concept is great but implementation is traditionally poor in my experience.
Tell me about it !
At Future Horizons we obviously look at, and work for, companies and governments across the world, and sometimes you see decisions having already been made for political rather than any commercial reason, but they need a report to back it up. We just have to refuse to bid on such contracts.