Intel Shifting . . .

Quietly Making Money

In the pricing wars, there was a ceasefire this quarter. Intel and AMD pretty much left each other alone. That combined with a bigger than seasonally usual upsurge in CPU sales gave both of them much better financial results than could have been guessed three months ago.

Intel basically sold 15-20% more processors and got the same amount of money for them. They also did a good job of keeping costs down, so it didn’t cost them any more to make that 15-20% more.

So an extra billion dollars in revenues meant another billion in profits from the processor division.

The CPU-making business is what gets called a high-fixed cost, low variable-cost industry in accounting circles. What that means is CPUs don’t cost a lot to make, but the equipment you need to make them costs an arm and a leg. A fab plant costs billions of dollars whether you make a hundred or a hundred million CPUs with it.

This means that when times are good, extra sales are almost entirely profit. When times are bad, on the other end, you can’t cut expenses like you can in other businesses. You can’t give half the fab back.

This quarter, times were good, so Intel did well. They’ll probably do even better next quarter.

Where The Growth Is

What is probably the most significant trend in the financial numbers is not what Intel sold, but where.

Since the third quarter of 2002, sales in the Americas have gone up less than 4%.

Since the third quarter of 2002, sales to Asia-Pacific have gone up over 31%.

Let’s compare Intel’s figures in the recent past (all figures from the third quarter of that year):

Region	1997	1998	1999	2000	2001	2002	2003
Americas	48%	47%	45%	42%	37%	32%	28%
Asia-Pacific	19%	20%	23%	27%	31%	38%	42%
Europe	24%	26%	26%	22%	25%	23%	21%
Japan	9%	7%	6%	9%	7%	7%	9%

Perhaps this set of figures might be more impressive:

Revenues From	3Q 1997	2Q 2003
Americas	$2,954	$2,168
Asia-Pacific	$1,108	$3,266

In the last six years, Intel’s revenues have dropped almost 30% in the Americas while almost tripling in Asia-Pacific.

This relatively sudden shift towards Asis and away from America has geographical implications: Intel will be much more inclined to put new plant where their customers are.

It’s more than that, though. The numbers for the Americas show that rapid computer sales growth isn’t indefinite; there comes a point of satiation.

The expanding Asia Pacific numbers are hardly surprising, even excluding Japan, that area has over four times the number of people the Americas have. It should be noted, though, that in general, the new growth markets are in general rather poorer than those who bought computers in North America/Europe during the late 1990s, so pricing pressure will if anything intensify in the years ahead.

It should also be noted that most of the population in the region is much poorer, so much poorer that including them in the potential pool of buyers over the next decade is sheer fantasy.

However, even very poor people snap up technology when the price is right. Mobile phones do extraordinarily well in some of the darndest places, but then mobile phones cost a lot less.

What Intel and anybody else in the computer business have to face in the long term is that there aren’t too many easy pickings left. They can ride the Asian tigers for a few years, but after that, further growth (or even just maintaining revenues) can only come from much cheaper computers.

It’s hard to sell a $200 processor in a $150 system. The biggest threat to Intel in the long run (i.e. five-ten years) isn’t AMD. It’s that little fact.

Ed