|
Apr 25, 2007 at 02:15 PM |
With quarterly conference calls coming thick and fast, it has already become very apparent that the recent wave of market forecast reductions for the industry are now impacting the equipment suppliers. Novellus, Lam, Cymer and Semitool have all stated that they experienced
varying levels of tool order push-outs in the first quarter, with much
of these by at least two quarters!
The fear here is that the longer the push-out period, the greater the chance that these convert to order cancellations.
The push-outs are not centred on memory manufacturers - they also come from the logic side. It would seem that in memory there is a capacity stall brewing due in part to decisions yet to be taken on what products that new capacity will produce in the second half of the year.
With memory players shifting away from NAND last year as ASPs declined by 60 percent with the uptake in DRAM, the result has been a price plunge in DRAM of over 50 percent in the first quarter of 2007.
The dilemma for the memory makers is that neither of the two products looks attractive now. So which one is ‘better' to bet on in the second half?
Though conversion from NAND to DRAM - and vice versa - has now become a viable and successful shift, it doesn't come without its own problems.
A key problem is related to lithography, which just happens bring with it the most expensive tools in a fab, and one with some of the longest lead-times.
As NAND has been scaling more aggressively than DRAM, it requires immersion litho tools for critical layers as well as more emphasis on i-Line tools for non-critical. On the other hand, DRAM needs more KrF and ArF than i-Line.
Compared to last year, the capital spending in the NAND market is expected to decline this year and next, but we are also in a market with NAND that is still young and growing, which proves very difficult to project demand.
We are therefore in a difficult position concerning product and equipment selection that has caused the push-outs to occur.
On the logic side we see Intel and AMD tweak overall CapEx for 2007 downwards, while also seeing asset-lite strategies at TI and ST Microelectronics bite as they cut historical CapEx budgets. These are budgets that foundries have yet to spend due to lower utilization rates and new capacity streams already set in motion.
The fear here is that foundries have followed a clear path in recent cycles of waiting until utilization rates are near 100% before spending on new capacity buys. This could lead to a big hole developing in equipment sales from now until the fourth quarter!
Although those equipment suppliers mentioned above have all reiterated that the push-outs are due to timing issues rather than cut-backs, that may well not be the case for the next quarter commentaries!
|