Hi Geoff Nunn; Regarding NSM as a stock and its history.
National Semi sold about $17.25 per share in 1997.
Back in 1988 they sold $23.70. They are not a growth
company, they are a shrink company. Long term chart:
tscn.com
They are a classic example of a company selling at
a growth stock multiple that has had sales (and
earnings) growth far below that of Ford, GM or
C. (Actually, even TXN has long term per share
growth well under Ford.) My feeling on NSM, is
that it is way overpriced, even at current values.
So much for the stock, and the company's long term
historical growth. Of course when you are an
industry leader its hard to generate high growth,
and that is the position NSM was once in. Ford:
tscn.com
As for the company's products, they are everywhere,
and are of reasonable quality. Do I think they are the
big new leader of the next generation of PCs? Nope.
Of course there's a chance. But as far as posting links,
I can only post links to the companies that send out
press releases. If a company is quietly going about
something, and doesn't want to give its competitors
free advice, they won't publicize what they are doing.
That is why my links are going to typically be to the
weak sisters. Believe me, everybody is using their
lunch napkins to calculate how much of a system
they can get on a chip. They just aren't necessarily
putting out press releases on it.
Now for a comment on integration and what happened
to NSM. When an integration wave sweeps through
an industry, it leaves a lot of devastation in its wake,
and NSM is a great example of this.
NSM had its growth back in the heyday of small scale
integration. They were an industry leader in integrated
circuits back in the early 70s. If you were looking for
rising stars in 1970, NSM would have been your big
growth stock. As an economist, since integration and
technology are outside your expertise, you might have
bought the stock based on its historical trends. NSM
got destroyed when integration devastated the small
scale integration market. Key employees left and
started new companies. Product lines become obsolete,
or sales dropped off as huge numbers of parts were
combined into single chips.
Buying high priced high tech issues without a firm
grip on what technology is all about is a little dangerous,
and NSM is a great example. In addition, the NSM
story is similar to what is just beginning to unfold in
the PC box makers business.
National sold the chips that went into small computers.
CPUs at that time were made from large numbers of
chips. This was the time I began studying digital logic,
the latest way to build a processor back then was
called "bit slice" architecture. You used a whole lot
of chips to make a CPU. It was complicated and
needed a lot of hard working engineers to get one
running. Read "Soul of a New Machine" by Tracy
Kidder to get an idea what goes into a processor
design.
Companies that built small computers using NSM's parts
were largely eliminated when integration reduced the
parts count for a CPU down to one. I can only remember
a few of the names now. Data General, and DEC. But
DEC survived, (until recently). Why were most of the
minicomputer makers unable to survive the transition?
Charts of Data General, Standard Microsystems, DEC, Unisys:
tscn.com
tscn.com
tscn.com
tscn.com
The transition is hardest on the industry leaders whose
revenues were tied into the old technology. What
integration does is allow a much faster, much cheaper
much easier to manufacture product that uses much
less power. Instead of having an evolution in designs,
where each year's product looks quite similar to the
previous year's, integration makes a product that is
revolutionary in terms of price performance.
It is hard to predict when an integration wave will occur.
In fact, integration improvement happen every year, but
result in simply evolutionary improvements. For instance,
the Z86 microcomputer on a chip is a very small 8-bit
CPU with memory and I/O that fits into an 18-pin DIP.
It was the original embedded processor that one found
in most mice, for instance. It is the most common
processor on the planet, with production in the billions.
The cost is currently well under $1 in quantity 10K,
mask programmed.
So for products that could get by with a Z86 CPU,
(not a close relation to the Z-80,) the integration wave
reduced the parts cost to the sub $1 range. The
effect on companies that sold very small processors
was obvious. If the processor was a major part of
their cost, then competition made them reduce their
prices. Margins also decreased, and unless they
could attract much, much, much larger numbers of
customers, their total revenue decreased. Since the
companies had been designed to have a certain
amount of overhead, when the margins on their
products decreased below that overhead they lost
money. A lot of them went bankrupt as management
refused to admit that the marketplace was moving
in a way that left them no possible way of making
as much money as they had before. Rather than
admit that their industry was going to contract, they
stayed in denial and competition eliminated them.
For DELL, it is not a question of whether or not
the PC marketplace will undergo a final integration
wave. From an engineering standpoint, it is clear.
Integration allows a reduction in costs so massive,
that eventually it wins out over discrete complexity.
It has happened before in the minicomputer market
place, as the processors were put on a chip, and
it will happen again when the entire systems get
put on a chip.
No, there is no question that eventually the majority
of the market place will reduce itself to commodity
level. That is, the majority of people are satisfied with
slightly aged technology in automobiles and TVs, so
will they eventually be satisfied in computers. No,
the real question is when does this happen. I think
the obvious success of the sub $1000 computer
is clear and convincing evidence that the
time is now. And even if people aren't satisfied with
slightly out of date technology over the next few
years, the faster speed and lower power consumption
of highly integrated products will win designs in the
end. It is just a matter of time, but the break is
already in the dam.
People who think customers are going to go back
to paying more for a computer are hiding from the
long term trends. The bell has rung, it now cannot be
unrung. Back in National Semi's day, a low end
computer cost much, much more than now, even
before an inflation adjustment. The price for
low end, general purpose computers continues
to decline. It drops drastically when integration
allows a reduced size and cost, it never goes
back up, and the next drop is in process.
The race is on to make next year's product, and the
number of entrants will be much larger. Prices will
drop much further. Next year's sub $1000 computer
will be much, much more powerful than this year's.
The reason is that integration will continue to march
ahead, under the force of competition.
Dell is in denial to these trends, and doesn't appear
to me to be making a plan of action. This is normal
human behaviour. Humans believe what they want
to believe, in the face of overwhelming evidence.
Nobody wants to believe that the industry they make
their money from is in decline, and Dell is no different.
Moving into higher cost machines is untenable. The
competition there will be tremendous, and those
machines will be greatly reduced in price just a half
decade after the low end drops out. This is the way
of an integration wave. First the low end, then up
through the rest of the product line. And customers
jump into the low end cause they get more bang for
the buck there. A lot more. And an integration
wave moves fast. That high end market will hit
the wave at most 5 years after the low end, as
higher integration levels allow the integration of
ever more complex computers.
When I was in the mini-super computer industry
in 1985 I saw the first wave coming. I got out of
the industry as it crashed. Only a couple not very
profitable companies were left. Then the wave
went on to take down the supercomputer industry.
Where is Cray Computer now?
What's happening now is the second wave. The
first wave put a CPU on a chip. The second wave
puts pretty much the entire system on a chip. Both
waves reduce the price of computers by close to
10x. The second wave went through embedded
processors a long time ago. It just went through
mobile computing, and now it is starting on the
desktop market. You don't have to be an engineer
to see the trend, you only have to read the financial
pages for a couple decades on the computer industry.
These are slow changes, but they are damn quick
if you're holding a stock with a P/E of 50.
There really isn't anything Dell can do. The trends are
in place, and it is clear to me that the total number of dollars
spent world wide on general purpose computers is going
to decrease, (or increase at much slower rates) just like
it did for supercomputers and main frames when the
integration waves ran through those industries. Give
the PC industry 5 years. Total industry wide sales may
be less than what Dell sells today, after accounting
for inflation. I know this seems ludicrous. But look
at that sub $1 computer. It used to cost a lot more.
Sure, as the price dropped more were sold, but as a
commodity. And the guy who kept that market was
the guy who was vertically integrated. And Zilog
ended up with a stock so cheap that they finally got
bought out this past year, I believe.
A final thought. So the third world is going to buy all
these computers from us? Do you think they are going
to buy $2000 computers or sub $1000 computers? Do
you think they would buy $200 computers or $1000
computers? You know how much money the average
citizen of the world makes. Now which computer is
going to sell in volume? The one that costs a year
of wages, or the one that costs a month?
I see Dell in the position of DGN 25 years ago. An industry
leader in an about to decline industry. Sure they will
survive. But not with great stock appreciation. Of course
I am looking at the long term, as the charts I have given
above show. Who knows what happens tomorrow? My guess is
for an up market this summer, with the Dow running to 9600,
but it is just a guess.
-- Carl |
