We haven't even gotten through 2010 and the prognosticators in the semiconductor racket are already polishing their hafnium-doped silicon dioxide balls to reckon what the future will hold chip revenues in the next two years.
The Semiconductor Industry Association, which provides a monthly snapshot of global chip sales, put out its forecast late last week, and as has been clear from the slackening growth rates in the past several months, the Great Recession, at least as far as the semi industry goes, is over. The beautiful juxtaposition of constrained supply and rebounded demand has pushed global chip sales back up where they belong.
The SIA, in fact, is now saying that the chip biz will break records this year, with $300.5bn in sales across all types of chips, a 32.8 per cent increase over 2009's sales levels.
"We experienced record sales this year due to strong global demand across a broad range of end markets," explained Brian Toohey, president of the SIA, in a short statement accompanying the projections. "We expect more moderate growth through 2012 as the economy recovers and consumer confidence restores."
The SIA's long-term forecast calls for chip sales to rise by 6 per cent, to $318.7bn, in 2011, and the market is expected to cool off further in 2012, with only 3.4 per cent more growth, to $329.7bn in aggregate revenues. Over the four-year period from 2009 through 2012, the SIA says the compound annual growth rate at works out to 13.4 per cent. (It would have been interesting to see that the growth rate was from 2003 through 2007 inclusive, before the Great Recession crashed chip sales.)
In an email to El Reg, Bill McClean, president of chip market watcher IC Insights, said the SIA needs to get out more and stop being such a bummer.
"Overall, the SIA forecast for 2012 looks like the 'death' of the semiconductor industry, and we couldn't disagree more," writes McClean. "What really strikes me about the SIA forecast is how pessimistic it is for 2012 at only 3.4%. One of the big problems I have with the forecast is the lack of assumptions. What are the unit volume and average selling price assumptions behind this forecast? What are the capital spending and thus capacity forecasts behind these market forecasts? What about worldwide GDP and electronic system sales? All of these assumptions are needed before you can put any stock in any semiconductor market forecast, including ours."
McClean says IC Insights is forecasting global chip revenues will grow by 31 per cent in 2010, almost the same as the SIA, but based on his assumptions about the future, he expects revenues to rise by 11 per cent in 2011 and by 10 per cent in 2012.
(Provided the Mayans aren't right and exact their revenge by having the world end in 2012, right on schedule.)
IC Insights is assuming that two years from now, the unemployment rate in the United States drops by two points to maybe 7.6 per cent, and further says that 2012 being an election year, that generally means the economy does better. (There are exceptions to that rule. Like 2008, for instance.) "The US economy depends more on the mindset of the 90+ per cent of people that have jobs than on a few percentage points of reduced unemployment," McClean says, adding that China, India, Latin America, and other regions of the globe are buying more electronics that are loaded with chips. "This is a wave of tremendous power and it will be gaining momentum over the next few years, not losing momentum."
Chip demand will be high, therefore, and chip fabs have been fairly conservative in their wafer baking capacity, according to McClean.
"With IC unit volumes on pace to follow their historical 8 to 10 per cent growth rate per year, you would need steep declines in semiconductor average selling prices to have only 3 per cent market gains," explains McClean. "With capital spending as a percent of semiconductor sales hovering around the very low figure of 15 to 16 per cent (even after this year's surge in spending), the semiconductor industry is nowhere near the overcapacity levels that would drive prices down this much, and it is unlikely to encounter significant overcapacity over the next few years."
We'll see who is right in a little more than two years. ®