HP Labs says memristor is for real, and a semiconductor consortium is diving into 400-mm wafers. Here are the takeaways from Future Horizons forum.
For over 20 years, Future Horizons, a UK-based market research firm run by Malcolm Penn has held a live forum that brings global electronics industry executives together to discuss issues affecting the industry. This year it was held in Seville from October 5th to 7th and I was lucky enough to attend.
Here are my high-level takeaways and learning from the event:
Stan Williams of HP Labs showed that the fabled memristor is real and could dramatically change the way we think about mass storage if these new flash memories are as fast and cheap as HP thinks. For me this was the biggest revelation from the event. Pleaseread this story from Peter Clarke of EE Times, which puts it in perspective.
Hossein Yassaie of Imagination Technologies talked about developing strategy at the same time that we heard the sad news of Steve Job's death. The key learning is that companies need a long-term vision of where the business should go and then single-mindedly execute on that vision. If there are too many strategy changes, it calls the long-term vision into question.
3-D packaging was discussed by Mentor Graphics and MonolithicIC. It seems like it will help with density and speed issues in the future but probably only in leading-edge products for the next few years.
J.J. Yamaguchi of the Japanese Semiconductor Industry Association (and Renesas)talked about how the Japanese semi industry is trying to get back to growth and leadership by focusing on some key strengths, but it wasn't clear to me that they have an answer to the Taiwanese, Korean, and US competition.
Energy, environment, and health were the three major semiconductor growth drivers touched on by ST, IMEC, and Freescale, but there are massive barriers to adoption for smart grids, intelligent highways, and techno-medicine that need to be addressed.
The mythical Russian semiconductor industry may finally emerge if the $1 billion investment in Plastic Logic (displays) and Crocus (flash memory) happens this year or early next. There have been many false starts in Russian semiconductor development but this time the money is there thanks to oil and gas revenues so everything depends now upon political will.
There was much talk of the giant step from 300-mm to 450-mm wafers and the impact it will have on the industry. A new consortium of companies, including Intel, TSMC, and IBM, plans to produce wafers. The consortium is working on the technology issues and will invest in upstate New York to make this happen.
Malcolm Penn gave us his forecast for this year's business and it's roughly 6% growth, with roughly 1% to 3% next year. The semiconductor market is closely tied to the vagaries of the global economy and the specter of over and under supply of wafers that invariably leads to wild swings in orders and supply. The biggest issue the industry faces, however, is that the average selling price of semiconductors continues to fall over time. The electronics industry captures less and less of the value of the end product, so how will it get the value back?
My answer to the value question (and was the theme of my session) is that the industry needs to transition into software and services for delivering system success and get paid accordingly. This is easy to say but incredibly hard to do because we will need a major shift in the business model that brings new semiconductor products to market.
David Blaza is senior vice president of UBM Electronics (the company that publishes EE Times and EDN). David has over 20 years of sales, marketing, and publishing experience in the technology sector working for companies as diverse as IBM, Motorola, Mars Electronics, CMP and now United Business Media. He is a graduate of the University of Bradford, England (BS, Materials Science) and the University of Stirling, Scotland (MS in Economics & Technology).
David, great points all around. Software and services, sure, but let's not surrender hardware and silicon design and manufacturing. Software may make platforms more flexible, but hardware will always be optimized, always make the world go around.
(Even one as crazy as ours!!)
CEVA's Eran Briman examines the explosion in vision processing and why it requires a rethink of mobile and embedded processing architectures if the future 'connected vision' paradigm is to be fully realized.