A message that I sent this morning to friends and family that took my advice to buy AMD over the last year:
AMDers,
Well, the last few months have been hard to ride out, but we are
approaching the light at the end of the tunnel. Several factors are
combining to set AMD up for similar price gains that were experienced
between last October and last May. To understand why this might happen,
let's first look at the factors that caused the price decline:
1. Over the last year, mutual funds had acquired huge amounts of AMD
(while Fidelity was pushing the price to $97 pre split). Selling by
these funds was a major reason for the price decline.
2. AMD was sold off together with all other semiconductor stocks. This
started with a downgrade of the sector by Soloman Smith Barney earlier
this summer, and culminated with Intel's Q3 earnings warning (of course,
to anyone that has listened to anything that I have been saying, Intel's
warning was no surprise).
3. Intel's marketing and "reality obfuscation" machine have convinced
analysts that even though Intel has faultered in execution at every
corner over the last year (while AMD has executed flawlessly), Intel
will be able to once again crush AMD when their Pentium 4 chip is
released.
4. A large part of AMD's business is in flash memory. The fear that a
slowing in cell phone handset unit growth would cause flash memory
prices to tank has also pressured AMD's stock price.
5. Many influential analysts have projected that a softening in
processor demand, combined with an increase in supply from both AMD and
Intel, will lead to a price war that will destroy AMD. Ashok Kumar went
as far as to say that "AMD would become a road-kill"!
Now let's look at why we are set to take off again by addressing the
above points:
1. Today, Oct 31st, is the last day that Mutual funds can sell shares
for tax loss purposes. Much of the recent selling was due to prior
losses begetting further tax loss selling, begetting more losses,
begetting further tax loss selling, begetting.... well you get the
point. Starting tomorrow, funds can no longer record sales for tax loss
purposes (although individual investors have until December 31st). This
should take a large amount of selling pressure off the stock.
2. It looks like we are at, or nearing, the bottom of the tech sell-off.
As AMD was sold off with the market, it will also be bought back when
the sector rebounds -- regardless of fundamentals.
3. The Pentium 4 has been designed for very high clock rate (Mhz sells),
but the very architecture that allows for high clock rates, limits the
performance at a given clock rate relative to AMD's Athlon, or even to
Intel's own Pentium 3. This means that a 1.5 Ghz Athlon will greatly
exceed the performance of a 1.5 Ghz Pentium 4. In fact, benchmark
results that have recently leaked out suggest that a 1.5 Ghz Pentium 4
will have a hard time competing with a 1.2 Ghz Athlon -- and the 1.2 Ghz
Athlon is here now. The release of 1.4 and 1.5 Ghz Pentium 4 processors
has slipped to November 20th (I can almost assure you that it will slip
futher). The Pentium 4 design is dogged by several other factors than
it's underlying architecture. Intel and Rambus made a devils pact last
year that now forces Intel to design its own chipsets for use with the
Pentium 4 to use Rambus memory. It turns out that Rambus memory is much
more expensive than the DDR SDRAM memory that AMD is using with the new
Athlons, and Rambus is also, in fact, slower. Intel is trying to get
around their contractual nightmare by letting VIA of Taiwan make a DDR
chipset for the Pentium 4, but this chipset will not be available until
the middle of next year. Also, there is the problem that the Pentium 4
die size is huge -- more than twice the size of the Pentium III or
Athlon. This results in chip yields that are a third to a quarter as
good as for the Pentium 3. For each wafer that Intel decides to make P4s
with, they produce only a third as many processors as they could have
made making Pentium 3s. The implications of this are large. From Intel's
perspective their margins collapse. From AMD's perspective, a quick ramp
of Pentium 4 will lead to a processor shortage (subtract 3 P3 for the
addition of each P4) that will fall right into AMD's hands. AMD would
easily be able to sell everything that they could produce with little
pricing pressure. The Pentium 4 is further constrained by over heating
issues. Due to these issues, the Pentium 4 does not become a viable
product until Intel migrates their manufacturing process to 0.13u, and
this does not happen until 2H, 2001. In the meantime, AMD is on track to
deliver a 1.5 Ghz next generation Athlons (Mustang) in Q1. These Athlons
will perform better than then current version (at a given clock speed)
and cause Intel serious problems. Intel should be able to release a few
1.7 Ghz Pentium 4s in Q1, but AMD will be shipping 1.5 Ghz Mustangs in
volume that could only be out-performed by a 2 Ghz Pentium 4.
In Q1, we will see the introduction of high speed mobile Athlons. In Q2,
AMD releases their first multiprocessor systems that will blow away
multi processor Intel Xeon systems -- hitting Intel at the heart of
their high margin business. In Q4, AMD will be sampling their next
generation 64-bit Sledgehammer chip, and it is starting to look like it
will be the iceberg that will sink Intel's 64 bit Itanium (Itanic!)
server chip. The Sledgehammer will have dual 2+ Ghz cores on a single
chip. Intel's Itanium is struggling to hit 700 Mhz and is so big that
they could not think of putting more than a single core on a chip. SUN
and several Linux vendors have already pledged operating system support
the the Sledgehammer, and if Bill Gates doesn't want to let Linux take
over, Microsoft support is sure to follow. In one year, it is possible
that AMD will have a performance leading product in every market sector.
So, don't be swayed by idiotic analysts that crawl out form under rocks
and repeat that, "well AMD is doing great, but just wait until Intel
squashes them." If you understand what I have been saying, Intel has
nothing to squash AMD with! AMD is making large amounts of money
selling chips at an average selling price of $90, while Intel is
experiencing negative earnings growth (subtract gains from investments)
selling at an ASP of $190. As AMD enters the high end mobile and server
markets next year, their ASP will rise, and Intel's will fall. The
implications on earnings for both companies is enormous.
4. Flash memory: It is true that the rate of cell phone handset unit
growth is slowing. BUT THAT IS NOT THE POINT. The rate is slowing, but
demand is still increasing. But, the main thing is that next generation
phones will require much more flash in each phone as functionality is
increased. Phones with increased web access features, PDA functionality,
and built-in MP3 players will on average require 3 times as much flash
in each phone next year as compared with current phones (per customer
requirements conveyed to AMD). And phones only represent a third of
AMD's flash business. It is clear that flash demand will remain strong.
But you get these uninformed analysts that claim, "well, AMD's flash
business is going gang busters, but just wait until supply exceeds
demand, prices drop and AMD becomes a road-kill". Even if capacity
starts to catch up with demand in a few years, AMD is insulated from
price drops due to long term supply agreements with major OEMs. They
just signed another huge three year deal with Nortel Networks today. In
the past couple of weeks AMD made strategic alliances with both Texas
Instruments and LSI Logic to combine processors form those companies
with AMD flash for next generation phones and information appliances.
The fact is that AMD's flash business is setup for tremendous growth.
5. From the discussion above, it can be deduced that a price war may
hinder AMD, but it will kill Intel. Intel could play the price war card
against AMD when all they had was the K6. Now that AMD has superior
products, Intel's hands are tied behind their backs.
winmag.com
I hope you all bought more at $18. I was never wrong about AMD. Their
fundamentals are stronger than ever, and the factors that combined to
kill the stock price over the last four months have largely played their
course. The stock continues to represent a wealth generating
opportunity.
Regards,
Pravin. |
