SAN FRANCISCO--Any time I sit down to listen to an industry forecast, I end up with the same feeling as I get when I walk out of a hyped Hollywood movie: Is that all there is?
The outlook is inevitably up-and-to-the-right, and well-known semiconductor technologies fit neatly into predictable application trends. Thanks for coming. There's coffee and cookies in the back.
But we can't think this way any more. We're at a tipping point in electronics history that's going to require a fundamentally different way of approaching electronics design. I'll get to that in a second, but first to the forecast, which is the reason I found myself sitting in a hotel ballroom at Semicon West, indulging in a chocolate chip cookie and considering the numbers and what more there might be to it all.
A robust 2013
At a Monday afternoon session sponsored by SEMI and Gartner, David Christensen, a Gartner senior research analyst, described an improving outlook for 2013, which was already fairly reasonable in Gartner's Q1 forecast. In the second-quarter revision, Christensen sees semiconductor sales rising 6.8 percent in 2013 to $320 billion (in Q1, Gartner pegged the increase at 4.5 percent). Customers have burned off excess inventory, and we're in a good supply-demand situation now, Christensen said.
What's driving it? Memory and mobile applications. As that excess inventory has burned off (an excess due in part to unexpectedly poor PC sales this year), DRAM demand has returned with little to no additional capacity being built. So, in Christensen's eyes, DRAMs function not just as 2013's thermals but 2014's as well (forecast up 6.6 percent).
"The DRAM application mix is also changing rapidly from PCs being the drivers to phones and tablets," he said.
Later, his colleague Bob Johnson took the podium to describe what he considers the drivers behind electronics industry change. Gartner likes to sketch the world in the context of what it calls several "nexus" forces: social, mobile, cloud, and information. These forces (driven by the consumer) are changing the value proposition within electronics systems (from high-cost to low-cost semiconductor technology) and prompting business-model changes in electronics vendors. For example, he cited the move by some big electronics companies to return to the vertically integrated manufacturing model. (I probed another related model trend
last month in this vein as did Frank Schirrmeister
on the topic of IP core licensing).
Johnson's punch line was this: "I believe that the new value proposition for chips is not the traditional faster-cheaper-more-performance (story). It's really how well they aid the nexus and people's participation in it."
That didn't go far enough for me.
A step farther
I think the new value proposition for semiconductors is not so much how they enable these known applications. The new value proposition is that silicon is the first technology in history that's both a building block and a system. It can be either. A light bulb is a light bulb. An alternator is an alternator. They go into systems and don't exist, really, without those systems. They are parts of an ecosystem built with someone else's application in mind.
But silicon is the ecosystem; it can be the system as much as it can be contributor to the system. Today, systems-on-chip really are only missing analog outputs (displays or speakers or whatever) and even those aren't far off.
And unlike the light bulb or alternator examples, in the case of silicon and electronics, the ecosystem is us.
As exciting and mind-blowing (at least to me) this is, here's the challenge:
What will challenge us will be how we think differently about silicon-as-system. The people who make alternators do it well and they give them to people who make really good car engines. Electronics engineers are good at… designing and building electronics.
Johnson hinted at it when he invoked the names of Apple and Samsung. If we can count on the fingers of one hand how many companies are starting to think this way, we have a mighty intellectual and cultural challenge ahead of us.
But it's early days, it's in our hands, and it's food for thought as we ponder Q3 semiconductor demand, the optimal number of chocolate chips in a 3-inch cookie, and the relentless march of innovation.