Editor's note:  This article was published November 15, 2005 with factual errors which we will underline and then try to correct in this release.

In the investment business, professionals are always looking for an "edge."  Something they know which nobody else knows.  Or at least something they know which is not already priced into the market.

With the SOXX at 466 this week, a lot of positive things are already priced into semiconductor company stock prices.  The growth in Notebook PCs and consumer electronics, seemingly led by the iPOD, are expected to lead to sales and earnings growth at semi companies.

The open question remains will sales grow as fast or faster than steadily increasing global chip capacity?  With a reported forty (40!) new 300mm fabs on top of continuous and repeated die shrinks, not to mention  smaller process geometries, coming on-stream, we have long felt sales growth would be hard-pressed to exceed, or even match, capacity increases in 2005.

Let's be clear, top-line sales growth is happening.  A sales increase in 2005 over 2004 is baked into the cake. It's those pesky profits we're worried about.  Is there an "edge" to be had by looking a little deeper?  Let's see:

SIA reported earlier this month, "Third-quarter sales at $58.7 billion were up by 8.9 percent from the second quarter, when total sales amounted to $55.6 billion."  While SEMI reported today, "Total silicon wafer area shipments were 1,748 million square inches...7 percent above third quarter 2004 shipments." (Actual figure 7.3%.  Visit our web site to review the details.

FYI, September was the highest-grossing SIA month in eight of the last 10 years and a big jump between August and September is normal.  September is commonly viewed as the final build-up to the annual Christmas and Chinese New Year consumer buying sprees.

At the risk of throwing in more "interesting" but only slightly-related factoids, Asia-Pacific in grew 12.1% q-o-q, accounting for 62% of SIA's sales increase.  Since APAC accounts for "only" 47% of the SIA's reported total, all three other regions (Americas, Europe, Japan) lost share q-o-q.

If you're still awake, this diversion got us to checking numbers, so we decided to do some extra ciphering and found SIA's reported 3Q04 actual $55,555,880 grew to 3Q05's actual $58,651,406, which was an increase of $3,097,526.  Yippee, except this represent only 5.6% SIA growth!  Where did the headline 8.9% come from?. Yikes!  All those PhD's, MBA's and assorted geniuses over there and they can't do arithmetic!  Did nobody else actually check the numbers?

Apparently, the geniuses on staff here at View from Silicon Valley aren't so smart either.  SIA's headline attached $55.6B to 2Q05 sales but it was the 3Q04 figure.   We used the right y-o-y numbers but SIA's verbiage explicitly said q-o-q.  SIA's underlying numbers show 2Q05 at $53.9B, from which 3Q05's reported $58.7B is indeed +8.9%.  Mea culpa!

In our defense, SIA routinely uses or omits y-o-y, q-o-q or m-o-m figures, depending on which ones put results in the best light.  In the process of adding and deleting comparisons, they apparently confused themselves --and us! --this month.  It bears repeating, SIA is a trade group and not an impartial arbiter of factual data.

The rest of this article will now report and reflect on the correct figures:

OK, so SIA is correct with 8.9% q-o-q sales growth (8.876% before rounding).  However, their good friends over at SEMI report 8.842% q-o-q wafer area growth.  In other words, q-o-q sales grew at only the same rate as total wafer area.   How much do you suppose semi companies' overall profits will be increasing under these conditions?

So why the kvetching?  Sales growth of 8.9% for a $200B+ industry is still pretty good.  That’s a fresh ~$18B in only 12 months.

Taking this exercise one step further, these figures omit the increase in units per wafer area due to die and process shrinks.  If you assume a measly 10% improvement in units/area (which seems conservative given the velocity of process shrinks accompanying all the new 300mm wafer capacity coming on-line), then 8.842% x 1.1 = 9.726% increase in unit output, which means dollar sales are growing ~10% SLOWER than unit output.

Lest we be accused just data mining with a different slant than SIA, let's also look at y-o-y.  SIA's 3Q05 +5.6% y-o-y dollar growth was in the face of 7.3% y-o-y wafer area growth.  As reported in Rev.0, y-o-y dollars are growing -30% slower than y-o-y wafer area.  (Or -43% when you add in a 10% increase in units/wafer.)

At the risk getting really wonky, SIA YTD'05 sales are up 6.1% while YTD'05 wafer area is down -0.88% (-0.008792).  The good news is this maybe helps explain the strength in semis.  The bad news is 3Q05's 1,748 million square inches is the highest number EVER. (+20% higher than the highest figure in 2000 or 2001).  And this number is prior to several new 300mm fabs due to some on-stream in the coming months.

In addition to this corrected data, and maybe an "edge." our reporting errors serves to remind everyone that SIA's press releases are engineered to put the best possible light on their numbers.  Consistency is left to the reader.

Conclusion:  Do you think reading beneath the headlines might provide an edge?

Coming up next:  More "edge"

* * * * *