LONDON – For now I am seeing 2011 as a "normal" year for the semiconductor industry in which growth is set to be in the 2 to 6 percent range.
There is, if anything, some general macro-economic uncertainty and chip market weakness in the final quarter of 2010 and into the first quarter of 2011, which is about as far as visibility extends. There is also reported DRAM weakness which is expected to persist throughout 2011.
On that basis I am predicting worldwide semiconductor market for the fourth quarter of 2010 at about $75.67 to $75.80 billion which would produce an annual global chip market for 2010 of a about $298.5 billion. This would be up 31.9 percent on the annual sales of $226.31 billion recorded for 2009 by the World Semiconductor Trade Statistics organization.
It would appear that demand in most sectors will remain reasonably healthy and that inventories, although climbing are at reasonable levels. The major issue in contention is therefore whether ASPs will climb as newer ICs are introduced at premium prices, or whether they will continue to languish or fall as they have done for a number of years as the chip industry has fought to digitize consumer electronics.
Collapsing DRAM prices alone could take $5 billion out of the market in 2011 on their own. There is also talk of an oversupply of foundry capacity in 2011.
In general there is likely to be a mix of price robustness in some areas and weakness in others. If 2011 proves to be a typical year I estimate the four quarters' sales would come in thus:
1Q11: $72 billion to $74 billion
2Q11: $74 billion to $76 billion
3Q11: $80 billion to $82 billion
4Q11: $79 billion to $83 billion
This corresponds to a 2011 annual total for worldwide semiconductor sales of between $305 billion and $315 billion, a percentage increase of between 2 and 6 percent, respectively.
Of course, every year has its own particular circumstances. We are yet to find out for certain what those are for 2011. If you think you know please contribute below.
I think your view is pessimistic. DRAM is certainly a concern but I don’t buy into the excess foundry capacity view. My detailed modeling of capacity versus demand for 2011 has utilization falling under 90% in Q1 due to seasonal factors but then running in the mid-nineties the rest of the year. ASP tends to follow utilization and with mid-nineties utilization I expect ASPs overall to be flat (worse case) to up (most likely) next year. My estimate is 10% growth.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.