Why I Wouldn't Touch a Chip Stock
By Donald Luskin
May 24, 2002
ALL THROUGH the great bull market of the 1990s, technology visionaries and New Economy gurus sang the praises of Moore's Law — the prediction by Intel (INTC) founder Gordon Moore that the number of transistors on a silicon chip would double every 18 months. Moore's Law was like a fountain of youth for the semiconductor business, or perhaps a perpetual-motion machine, driving each successive generation of chips to greater power, greater speed, smaller size and cheaper price, and an infinitude of new uses. It was a technology investor's dream come true.
But soon investors may realize that Moore's Law is a bit like buying stocks on margin — it's great when times are good, but it's deadly when times are bad. Moore's Law is turning out to be another force that has a dark side. It's a force that will make it extremely difficult for the semiconductor industry to pull out of today's tech recession.
To find out why, let's go right to the source of the force, and read what Gordon Moore himself said about it 37 years ago. He first wrote about the principles that would someday be called Moore's Law in 1965, just six years after the first simple planar transistor was invented. In an article for Electronics magazine called "Cramming More Components Onto Integrated Circuits," Moore wrote, "The complexity for minimum component costs has increased at a rate of roughly a factor of two per year (see graph below). Certainly over the short term this rate can be expected to continue, if not increase. Over the longer term, the rate of increase is a little bit more uncertain, although there is no reason not to believe it will remain nearly constant for at least 10 years."
And so it did. Moore revisited this prediction a decade later, in 1975, and revised downward the forecasted rate of doubling from every year to every 18 months. In that form Moore's Law has held remarkably constant ever since, right up through the very latest generation of Intel's Pentium 4 processors.
The dark side of Moore's Law starts to become apparent when you realize that it's much more than a prediction about engineering, although it is commonly misunderstood as only that. Intel's own Web site incorrectly summarizes Moore's Law as stating that "the number of transistors per integrated circuit would double every couple of years." No, Moore's Law is a prediction about economics, linking engineering advances to the cost of implementing those advances. In the same article Moore wrote: "For simple circuits, the cost per component is nearly inversely proportional to the number of components, the result of the equivalent piece of semiconductor in the equivalent package containing more components."
A simple way to phrase Moore's Law to capture its economic dimension would be to say that the cost-effectiveness of integrated circuits will double every 18 months. In other words, in 18 months you'll get twice the power for the same price, or the same power for half the price, or somewhere in between.
Moore's Law — understood as economics, not engineering — dooms the semiconductor industry to an incurable addiction to growth. Merely keeping revenues unchanged requires getting customers to want twice the power every 18 months, so that they'll keep paying the same old price. Or if they want the same old power, there have to be twice as many customers — because they'll only be paying half the price. That's why Intel's revenue growth has imploded, even as it ships record volume. In this deep tech recession, Intel just can't keep up with the law named after its own founder.
In a rapidly growing, innovation-rich economic environment, keeping up is no problem. New high-end applications are developed and demanded — creating the incentive for customers to upgrade, paying the same price for twice the power. Entirely new customer categories come in at the high end, too, spontaneously created by applications that were unfeasible at lower power. Think how many people bought PCs in order to be able to use word-processing software, then desktop publishing, then the Internet. But...now what? Same thing with markets for servers, and with routers and switches...now what?
And at the low-cost end, semiconductors can find their way into an infinite number of uses. Whole new industries can spring into existence when semiconductors hit the right price-point — cell phones, smart cards, PDAs, digital cameras, MP3 players. But...now what?
Moore's Law means the semiconductor industry has to run just to stand still. And it means that with each new day of this tech recession on which the industry didn't run fast enough, the volume hurdles to get back to peak earnings just get higher and higher.
So why, then, do semiconductor stocks — companies like Intel, Advanced Micro Devices (AMD), Broadcom (BRCM), Xilinx (XLNX) and Micron Technology (MU) — sport an average forward price/earnings multiple of 45.1, far above the Nasdaq's 33.7 and the S&P 500's 19.5? Perhaps die-hard tech-stock investors who grew rich on Moore's Law in the 1990s don't know what they're up against as they confront the dark side of the force. They'll find out soon enough: Racing against Moore's Law, it's going to take more than a little cyclical economic recovery to justify these unrealistic expectations.
It is perhaps revealing that the average forward P/E for semiconductor equipment stocks — companies like Applied Materials (AMAT), KLA-Tencor (KLAC) and Novellus (NVLS) — is an even more stratospheric 71.5. As good as tech-stock investors think it's going to be for the chip makers, apparently they think it will be even better for the arms-merchants who sell the equipment that the chip makers virtually have to buy to wage competitive war, even in a stagnating market. These companies operate on a law all their own — Rock's Law, propounded by Arthur Rock, the legendary venture capitalist who put up the seed money to start Intel: The cost of capital equipment to build semiconductors will double every four years.
I wouldn't want to be a semiconductor company CEO right now, struggling to come out of a devastating recession with a product whose price falls by half every 18 months but that's made with equipment whose price doubles every four years — and with my stock priced at bound-to-disappoint valuations. I wouldn't want to be a semiconductor investor, either.
Donald Luskin is chief investment officer of Trend Macrolytics, an economics consulting firm serving institutional investors |
