From a friend I received this interesting and quite bearish report.
Note the yield and unit count numbers buried within, interesting for future extrapolations. From TECH INVESTOR:
Semiconductor Market Focus
No Patience
Monday, August 24, 1998
In the past three weeks we have seen Wall Street move
from optimism to pessimism about the
device-manufacturing business. Early in August,
seasonal ordering stimulated a bit of a relief rally.
Christmas comes every year, so this should not be a
surprise.
A few bulls have pinned their
hopes on the recent firming of
memory prices. The move in
memory prices is the result of
leaner inventories in the reseller
channel and a transition to
64-megabit Dynamic RAM
production. In the 64-Mb arena,
we are expecting prices to
continue dropping in the coming
months as numerous production
lines mature with the product. Die
sizes are shrinking and yields are
getting very high. This drives costs down, and prices
will soon follow. We have heard the largest and most
active joint venture in the United States is seeing yields
of up to 70 percent on 64-Mb DRAMs before laser
repair. After laser repair, yields rocket to over 90
percent.
Folks, this is not a trailing edge process. These DRAMs
are being made with a 0.18-micron process, generating
550 64-Mb dies per wafer. Soon, the joint venture will
move to 0.15-micron and then 256-Mb DRAM
production. Don't be misled by the recent pop in prices,
the capacity glut is still with us. In Japan, it appears
64-Mb inventories are already accumulating.
The grope for the elusive bottom continues. We really
hate being bearish, but the signs are not good. As we
said in our last article, "Shareholders will likely tire
waiting for the next upturn and [semiconductor stocks]
could make new lows sometime this fall." Looking at
Friday's action it becomes obvious that Wall Street is
getting tired of waiting for the upturn.
Many investors will say that because we have been in a
bear market for semiconductor and
semiconductor-equipment issues we have to be close to
the bottom. It is our belief that a pronounced period of
low shipment levels will coincide with lower stock
prices. Unfortunately, visibility is not going to emerge
until a number of the financial dislocations pass. The
big-picture view is also cloudy.
"The economic fundamentals that led to the correction
are still very much in place. Indeed, they are getting
worse," said Bruce Steinberg, chief economist at Merrill
Lynch. He expects the global slowdown to last well into
next year. This will take time -- probably a long time.
Infrastructure's "Fab Rat" has worked closely with
many Japanese semiconductor and
semiconductor-equipment companies over the past 30
years. Some of his recent comments fall in line with Mr.
Steinberg's analysis, "It took many years for Asia to get
to the present condition and it will take many years for
them to get out. Just because the collapse was rapid
does not mean the recovery will occur at a similar
pace," he said.
On the heels of a bullish forecast released a few weeks
ago, the Fab Rat said, "After looking at the Global
Sales Report it is apparent to me that all of the upside
forecasts are dependent on a recovery in Asia.
Therefore, any realistic forecast has to address the
issues in Asia. I don't see forecasting in a vacuum. It
would be unrealistic to say PCs are going to do fine
unless Asia is also doing fine." Of course, there are a
few who are doing well in the PC space -- how about
Dell Computers [DELL]?
We have maintained that a recovery in the
semiconductor business will evolve in the form of a U
shape. The trough of the U will be agonizingly painful.
Many companies will go out of business. Many
companies will be acquired by larger players. Sales will
appear in spasms. Visibility will be low. The trough
could easily last in to the very late months of 1999.
Allow us to elaborate on what we are hearing about the
equipment business. Contacts in the field continue to
suggest conditions are deteriorating much more than
current earnings reports depict. In addition to the weak
level of orders and shipments, companies are being
forced to assess the impact of the year 2000 problem in
existing manufacturing facilities. The same state of denial
that plagues others with Y2K exposure is alive and
kicking in the chip-manufacturing business.
Interestingly, our research indicates the Y2K problem is
being aggressively addressed by companies that are not
household names. The ones we see looking for Y2K
solutions are run under the umbrella of large
conglomerates -- fabs that build custom parts with less
than state-of-the-art wafer process tools. We have had
conversations about Y2K with some of the largest
companies in the semiconductor business and it is
apparent they have only begun to study the impact.
(Admittedly, Texas Instruments [TXN] appears to have
done a great deal of due diligence.)
We are not going to describe the Y2K impact as
complete doom and gloom, but it is serious enough that
we feel a concentrated effort -- an effort so
concentrated it will continue to sideline new
construction plans -- must transpire over the next 12
months. In upcoming issues of our monthly letter we are
going to try and address many of the Y2K issues faced
by the industry. Yes, it is a long list and it will take a
good part of the year to detail the full story.
This week, wet blankets were tossed on semiconductor
bulls by National Semiconductor [NSM], LSI Logic
[LSI] and Analog Devices [ADI]. There are others who
are struggling with the lack of orders, but these three
were the most visible this week. The report from
Analog Devices highlights the weakness in the automatic
test equipment sector of the semiconductor-equipment
industry. The drop in bookings at Teradyne [TER] and
Credence [CMOS] only highlights a portion of the
weakness in ADI's outlook. The ATE sector is not a
huge market for ADI, so other areas are suffering too.
|
