Some "expert" was recently pontificating on about how the
only thing which really counts in the semi business is billings. (Editor's note: Duh!) The gist of
his argument was to ignore bookings and rest assured the semi business is strong because y-o-y billings are still up. Besides
marveling for the umpteenth time at how people can get paid for stating the obvious, I didn't give too much thought to this
"news."
Of course, billings are paramount. Nobody gets paid
for bookings. (OK, senior execs holding scads of free stock and/or stock options can reap big bucks out in
cyberspace based just on bookings.) The majority of the Silicon Valley economy, located down here on planet
Earth, however, doesn't get paid until real money changes hands. That's why you need billings.
On the other hand, you need bookings to in order to generate
billings. (Duh!, again) Since capital equipment doesn't book a large percentage of its numbers in turns business,
a drop in SEMI's bookings figures to be a pretty accurate near-term predictor of the direction of future billings.
Thinking back, I suppose this "expert" was setting the stage for
the news out of SEMI last week in which their January, 2005 book-to-bill (B:B) fell to 0.80 from December's 0.95.
Digging into the numbers, the real news was a -$225M drop in reported billings.
For those who think -$225M (-18%), is just a unique fluke, SEMI
experienced worse drops, in consecutive months, at the tail end of the Y2K boom(**):
-$518M (-22%): Dec'00 -Jan'01
-$244M (-13%): Jan'01 -Feb'01
-$414M (-26%): Feb'01 -Mar'01
Are we now getting a Y2K-like drop-off, only this time without
the preceding boom?
January 2005's $1.011B was the smallest booking number since November,
2003's $923M. SEMI shows November, 2003 was the last of 30 consecutive sub-$1B booking months. Alert
readers may also notice November, 2003 was roughly the start of Silicon Valley's latest housing boom.
January's 0.80 B:B was the lowest since September, 2002.
(Does that date sound familiar?)
Keep in mind, SEMI only releases three-month
averages. To move a three-month average -$225M in a single month, you need a drop of roughly -$675M! (Let's just
assume the number is honest and really stayed over $1B by 1.1%.)
As dramatic as the implications for such a steep drop might seem
for Silicon Valley, I did not hear ANY local report of SEMI's revelation last week. I was in the car
what seemed like all day Thursday and Friday last week with radio tuned to the news stations and I heard nothing.
Come on, is it only "news" when SEMI's numbers are good?
About the same time I found this news, I also heard apologists
speculating SEMI's B:B would "rebound" to 0.90 in February. Is it just me, or does nobody really understand book-to-bill?
Any sub-1.00 number is still a shrinking business! Semi's five consecutive months of sub-1.00 B:B (February
would make six) tell us there is a problem. Now, if somebody in the news would just report it.
We continue to experience semiconductor executives coming out
and re-assuring us the drops are just temporary. They promise the semi business will come roaring back in the second half
of 2005. If there is anything but "optimism" for why such a rebound will occur, they are not sharing it with the
public.
In case anyone thinks SEMI's B:B has no direct connection
to everyday life in Silicon Valley, I also found a blurb explaining, "VLSI's estimate of worldwide front-end capacity utilization slipped from 77.3% in December to 75.6% in January.
The firm predicted a slight improvement to 76.3% in February."(*)
What happened to all the fabs being full? Isn't any utilization rate
in the 70's a big problem? Won't fab depreciation start to clobber earnings in such a utilization environment?
Are the SEMI numbers down because utilization rates are dropping?
Keep in mind, Applied Materials downplayed their recent revenue
decline on the grounds there are 40 new fabs under construction this year. (I might be convinced this is bullish for AMAT if such news wasn't already
priced in.)
I see the real news as suffering a 75%-range fab
utilization rate prior to the ramp-up of 40(!) new fabs. And isn't it odd to notice, even with these 40
new fabs being built, SEMI's numbers went down?
Utilization rates in the 70's mean wafer prices are dropping,
Falling wafer prices means lower IC prices. With widespread reports of demand weakness, and therefore price weakness
today, where will IC prices be in a few months when some of these 40 new fabs start cranking out parts?
The cherry on the sundae was a couple Wall Street "analysts" recently
claiming 4Q04 inventory fell ~38% to "only" $1B. If true, this would be a sign IC production is closing
in on a balance with demand and prices, and fab utilization rates, will soon rebound. At least two major Wall
Street firms upgraded semi stocks in the last few weeks, seemingly based on such "news." Sorry, but Intel alone
is sitting on $2.5B+ worth of inventory. Untold quantities of finished goods are stacking up in distribution channels
worldwide. Anyone claiming there is only $1B of worldwide inventory is inventing reality.
Conclusion: No
matter who else ignores it, a -$675M billings drop from SEMI is big news. Recent semi market weakness is not a
"buying opportunity" based on any metric available to the public. Unless huge IC buying increases materialize
immediately, there is no reason for SOX components (index up 50+ points since January 24), nor local housing (median
up ~20% January y-o-y), to be seen as long-term "buys" at this point. Plan your finances accordingly.
