Cary,
I'm missing too much info to do a bottom up estimate of CYMI revenue. For example,
1. # New Wafer starts per week at .25u or smaller feature size.
2. # Litho steps per wafer
3. Current fab capacity usage/growth by year.
Current estimates from First call are:
97: $0.81
98: $1.04
CYMI has a habit of beating estimates by about 50%. On that basis, one could expect about $1.05 for 1997. (Actual for first half is 0.39, Estimate for 2nd half is $0.43. The First Call 1998 estimate would indicate to me that the analysts believe a major bottleneck or process problem will stop the migration to 0.25u dead in its tracks, that R&D will occur, not production.
IMO, the probability of either of those scenarios is less than .15.
The last SIA roadmap that I saw called for:
.25u in 1998
.18u in 2001
.13u in 2004
.10u in 2007
.07u in 2010
I would expect 300mm to become real during 1999-2000.
I believe all of these steps in the roadmap will stimulate demand for CYMI's lasers.
Re Komatsu and Lamda: I believe that CYMI, like Intel, will continue to outpace their laser progress and that CYMI will maintain its near monopoly share of the market for semi eqpmt lasers.
CYMI's business model calls for an 18% net margin. As they achieved 15% last quarter up from under 12% a year ago, I don't see any reason why they shouldn't achieve 16.5% next year and 18% in 1999. So far they have a habit of beating estimates.
In terms of # of years to reach service revenue targets: It's an estimate, I couldn't provide a rational argument for either a faster or slower ramp to 1/3 of gross sales.
Re UTEK's PGILD: Even though the tool's announced, I'm still unsure re its actual status. Is it Beta or available in production volume? I believe it's still very early days. I received some literature from UTEK which indicates it's using a 308nm XeCl laser. I don't know with certainty that it comes from Cymer. I believe that UTEK and CYMI collaborated at LLNL in developing the product but can't find that document neither.
In addition to doping, the tool is also used for "Laser Assisted Self Aligned Silicide Formation". Sounds like it does a very nice job of this process. However, I don't know enough about chipmaking to know how extensive the requirement is for this process; or what the alternative process is; or when the PGILD tools would be required for it.
I would guess that Fabs will migrate first to .25u then to 300mm then to .18u rather than trying to do 2 things at once. The PGILD processes will be essential with .18u and below. Thus, I would expect R&D requirements to begin in 1998, pick up in 99 with bleeding edge production by early 2000.
As a WAG, perhaps a dozen in 1998, 2-3 dozen in 99 and big numbers in 2000 and beyond.
My fondness for this company was phrased best by, I believe, Sheena Easton when she sang, "Nobody does it better, ..." :-) They provide an enabling technology. Once one fab has it in production, all other fabs producing that chip type will be forced to catch up; and the migration path for other types of chips will be facilitated.
Hope my rambling has been of some use,
Ian. |
