Online information source for semiconductor professionals

The latest sales pitch from SEMI to part with $4,000 greenbacks for their ‘World Fab Forecast’ market research data notes that we can all expect a ‘brighter’ 2009. Fab spending is set to recover by 20 percent over a 20 percent decline expected for 2008 compared to 2007.

Other than stating the obvious that there really isn’t anything ‘bright’ to see in these figures, SEMI is assuming that there will be 20 percent negative growth, when it could actually be worse than that!

We have seen several rounds of spending cuts from the majority of memory manufacturers. Foundries had reduced spending plans to levels so low at the beginning of the year that if wafer ASPs didn’t improve, it would be highly likely that final year spending would actually fall short of the original projections.

On the logic front, we have seen AMD reduce spending due to the fact that it hasn’t the money to spend on top of the cost of taking a complete fab offline. Everyone else is spending the least amount possible, and in many cases, are doing okay despite CapEx cuts.

Unless I have missed something in the last 8 months, a slowing world economy, U.S. credit woes and higher inflation due to oil prices rocketing doesn’t indicate that we will see a major round of capital spending come 2009.

According to SEMI’s sales speak, there are 53 fab tool install projects being tracked and up to 21 fab construction projects being watched for 2009.

Mike Splinter at Applied Materials recently noted that his company was tracking 40-odd 300mm projects. Some of these are postponed projects, be they built but not equipped or new construction projects yet to start.

The impression given by SEMI is that we can expect another fab rush as there is a long list of stalled projects in the pipeline, as well as new projects that are sorely needed as fab utilization rates remain very high.

In some respects, this view is totally believable, but the key question is how many of the postponed projects will come to life in 2009, and how many of the new construction projects will actually get the green light?

With the biggest spenders being memory manufacturers, what is happening in these markets that suggests there will be a return to spending in 2009?

NAND flash overcapacity is set to continue, and with only a slow move back towards equilibrium in DRAM by early 2009, cautious spending patterns could be around for some time.

Although the hype is building around SSDs (Solid State Drives), nearly all the market research firms seem adamant that this new market for NAND devices is a late 2009 and beyond market. Most SSD players are positioning themselves with the right product portfolio rather than actually ramping production next year, which requires fab expansions.

According to SEMI, fab capacity in 2008 is expected to be about 16 million wafers (in 200mm equivalents), a 9 percent growth rate over 2007. In 2009, capacity is expected to grow about 10 percent.

That figure does not seem very much to me, and it supports the view that caution will be a theme through the first half of 2009. Visibility overall will therefore be a key and that visibility simply doesn’t exist as yet. Importantly, 10 percent capacity growth may not mean a 20 percent increase in capital spending as there is a lot of capacity sitting there already.

On a more positive note, there should be more spending in 2009 than was seen in 2008. Demand for ICs remains healthy, utilization rates remain high and at some stage new capacity will be required.

However, with the restrained spending patterns seen so far in 2008, I struggle to see at the moment why we should expect anything that much different in 2009.