An Upgrade-From Albert:
08:53am EST 13-Dec-99 Gerard Klauer Mattison & Co. (Geraghty, J. 212-885-4003)
AMD--Upgrading to BUY; Flash Market Booming; Raising Est.
Advanced Micro Devices (AMD)--Upgrading to BUY; Flash Market Booming; Raising
Est.
John M. Geraghty, CFA Harry H. Sun
(212) 885-4003 (212) 885-4023
jgeraghty@gkm.com hsun@gkm.com
December 13, 1999
____________________________________________________________________________
Price: 29.25 52-Wk Rng:33-14.56 Target Price:$45 S&P 500: 1417
____________________________________________________________________________
Shrs Out/Mk Cap: 147Mil/$4.3Bil 5 YR Est Growth Rate: 15%
Div/Yield: N/A L-T Debt/Cap: 42%
Avg Daily Vol: 2.8 Mil
____________________________________________________________________________
FY Ends -- EPS --
Dec. Curr Prior P/E
98A $(0.68) N.M.
99E (2.60) (2.65) N.M.
00E 1.37 0.80 21.4x
____________________________________________________________________________
Qtrly -- 1Q -- -- 2Q -- -- 3Q -- -- 4Q --
EPS Curr Prior Curr Prior Curr Prior Curr Prior
98A $(0.39) $(0.45) $0.01 $0.15
99E (0.81)A (1.08)A (0.72)A 0.01 (0.04)
00E 0.12 0.09 0.17 0.15 0.37 0.21 0.71 0.34
____________________________________________________________________________
o Upgrading to BUY and raising estimates to ($2.60) and $1.37 from ($2.65)
and $0.80 for 1999 and 2000, respectively. We are raising our unit forecast
for the K7 for both 4Q99 and 2000. We are estimating 900K units in 4Q and
total revenues of roughly $865 million. For 2000, we are forecasting 8.0
million units, up from 7.3 million previously. We believe AMD has the
capacity to produce 10 million units in 2000, if the ramp of its 0.18 micron
process proceeds satisfactorily at its Dresden facility next year. Should
AMD add one or two significant PC OEM customers to its current list, the
chances of exceeding our forecast would be high. In addition, we are also
looking for a continued strong PC market, which should help both Intel and
AMD.
o The flash market is booming, and flash revenues for AMD could increase
87% in 2000, in our view. Driven by strong consumer end markets, including
set-top boxes, cellular handsets and other portable electronics, the flash
market has been one of the strongest semiconductor markets in the current
cyclical upturn. The SIA forecast is for a 63% sales increase to $4.3
billion in 1999, followed by increases of 36% and 19% in the next two years.
We believe this forecast may be conservative since business is strong enough
that most manufacturers are sold out for most of 2000.
As a leader in the flash market, AMD continues to ramp capacity.
Including FASL, its joint venture with Fujitsu, AMD should ramp unit
capacity by 70% in 2000 and over 100% in 2001. However, despite AMD's
current production ramp, demand still outpaces supply.
o Microprocessor (MPU) business appears promising as Athlon (K7) sales are
strong. On November 11, the company announced it could ship 800K Athlon (K7)
chips in the quarter while producing about 1 million units. As a result,
total revenues would surpass $800 million in the quarter. We believe that
Althon (K7) sales are stronger than expected - helped by supply constraints
on high-end Pentium III processors - and should be closer to 1 million units.
Also, we believe the Athlon processor is competitive with Intel's high-end
Pentium III processors in terms of performance. AMD could reclaim the lead
in the MHz race with its upcoming 750 MHz Athlon chip, although INTC may have
developed 800 MHz products recently. In addition, Intel will have more
production available in the first quarter of next year from its 0.18 micron
Coppermine process as it solves some production constraints. Consequently,
the race could get tougher in 2000.
While manufacturing has always been a problem with AMD, we believe its
0.18 micron production ramp is on track. Currently, all wafer starts for
the Athlon processor are at 0.18 micron. Fab 25 should be fully converted
to 0.18 micron in 2000 while Fab 30 (Dresden) should ramp in 2001. Last
quarter, unit shipments of the Athlon were hampered by the Taiwanese
earthquake as AMD was unable to get chipsets from VIA and other Taiwanese
producers. We believe these supply issues have been resolved.
In addition, as Intel is having some supply constraints on its Coppermine
(0.18 micron) Pentium III chips, pricing has not been as aggressive as
usual. We believe that continued success of the Athlon chip would
decrease the likelihood of an all-out "price war" from Intel. To date,
Intel has dominated the market at the high end with the major PC
manufacturers. AMD was utilized almost as a foil to extract better
pricing from Intel. In addition, Intel's market segmentation strategy
enabled it to price aggressively at the low end (Celeron) while
maintaining better margins at the high end simply because AMD did not have
a competitive high end chip. Now, with Athlon ramping, that strategy will
be tougher to implement.
Finally, AMD now possesses one crucial factor that has been lacking in the
past, i.e., production. Thus far, it has met all its production goals for
the Athlon chip. This is important to PC OEMs since product delays can
prove to be both embarrassing and expensive. Lately, Intel has been
unable to satisfy the upside requirements from those same OEMs while
delivering normal production runs. This may present AMD with an opening
at two major OEMs who have not used AMD to any great extent, Gateway and
especially Dell. There has been speculation about Gateway being a
possible customer for some time, and we believe the chances of GTW
becoming an Athlon customer are growing. Dell has always been a complete
Intel OEM, but even it may feel some competitive pressure if the Athlon
chip is the fastest on the market and Dell does not have a competitive
product from Intel, at least in the short term. We believe this is a long
shot, but as Athlon deliveries and performance increase, the pressure is
certainly growing.
o Valuation is compelling based on flash business alone. We believe AMD's
flash business could generate $1.4 billion in revenues in 2000 and earn
roughly $1.80 per share. Other flash manufacturers like Silicon Storage
Technology and Atmel trade at about 30x their calendar 2000 consensus EPS
forecasts. Applying a 30x target P/E multiple to AMD's potential flash
earnings, we arrive at a potential target price of roughly $55. If we then
apply a 20% discount to account for past difficulties in AMD's microprocessor
business, we arrive at our target price of $45. Yet, AMD's microprocessor
position appears to be the strongest it has been since the early 1990s, when
it made significant amounts of money in the segment. Our current forecast is
for a profitable MPU business in 4Q00, but higher volume than we are now
anticipating could obviously yield better results.
Additionally, AMD currently trades at about the lowest price-to-sales
ratio in our universe at 1.0x our 2000 sales forecast. For comparison,
Intel trades at 7.3x estimated 2000 sales while National Semiconductor and
LSI Logic trade at about 4x forecast calendar 2000 sales. We are not
predicting a dramatic increase in this ratio, but would point out that at
$45, AMD would trade at about 1.8x estimated 2000 sales.
o The risks for AMD continue to involve its manufacturing ability and the
difficulties of competing with Intel. Historically, AMD has had
opportunities to gain share in the MPU market, but has been hampered by
production difficulties as well as the fierce competitive response by Intel.
Intel has two items in great abundance: money and engineers. It has been
able to finance competitive responses to AMD because of its financial
strength and has been able to overcome short-term manufacturing and design
problems with its deep bench of engineering talent. Those factors are not
changing. Rather, we think the overall semiconductor market will remain
quite strong and that AMD will be able to deliver product to OEMs rather than
suffer production difficulties as it has in the past.
The other positive for AMD, the flash segment, is sufficiently powerful
for us to have revisited AMD as an investment idea. We believe that
investors underestimate how well AMD is doing in this segment. In the
final analysis, however, investors will pay the most attention to the
microprocessor segment. Here, AMD has a window of opportunity. This
time, we think it may be able to do something with it.
INVESTMENT CONCLUSION
Upgrading to BUY rating. We consider AMD a play on the booming flash memory
market. We believe flash sales could reach $1.4 billion (34% of total sales)
for AMD in 2000 and generate about $1.80 per share in earnings. In addition,
the outlook for AMD's microprocessor business is improving. AMD has proven
near-term ability to ramp the Athlon, and all wafer starts have been at 0.18-
micron production technology since the end of October. Should this
production ramp continue, and we are assuming that it does, AMD could
surprise investors with better EPS results than we are estimating, just as it
did in the fourth quarter that is now ending. We consider the valuation on
AMD to be compelling at 1.0x our 2000 sales forecast, despite its recent
appreciation. Our 12-month price target is $45. As always, the risk will be
the microprocessor segment. |
