Sonki..
<<Chips Worth Chewing On
By Ian Mount
FEELING CHIPPER?
IN THE HIP-HOP world, they call them "player-haters." Elsewhere, they're known as "naysayers," "doubting Thomases," even "nattering nabobs of negativism." You know who they are, those doubters who point to whoever's on top of the heap and say that they don't deserve to be there, it's all luck, they're about to fall. Sometimes they're right, you know, and sometimes they're very, very wrong.
In today's screen, we take a look at a sector that's been riding high and has been hearing some of that talk ? semiconductors, or chips ? and try to determine how much of it might be true. In 1999 so far, the Philadelphia Stock Exchange Semiconductor Index is up over 82%; since the beginning of October alone, it's up 28.1%. Giddy projections from the Semiconductor Industry Association (SIA) of year-over-year growth of 21% and 20% during the next two years ? from $144 billion in 1999 to $174 billion in next year and $209 billion in 2001 ? haven't hurt at all, especially when compared to the dog days of 1996 and 1997.
But lately there have been some nervous rumblings from those who watch the industry, perhaps indicative of a feeling that the industry has come so far that no longer can an investor simply bet on any player; it is time, according to such worries, to look more closely and pick more carefully. After Advanced Micro Devices (AMD), a maker of high-speed microprocessors that compete with products from Intel (INTC), announced on Thursday that production of its high-end Athlon chips was going better than expected and that it might actually break even in the fourth quarter (it was expected to lose 36 cents a share), Merrill Lynch analyst Joseph Osha, a Wall Street Journal All-Star, lowered his rating and price target on Intel. This sent shares of Intel down 4% on Friday.
"What?" chip investors wondered. "Won't this tide lift all boats?"
Then on Sunday, a column in the New York Times, "The Chip Stampede May Be Hitting a Wall," suggested that chips might have gone as far as they could go and were set for a fall. Chip stocks were so overvalued as to be insupportable, it argued.
So are these people Chicken Littles, or are they just prudent? Well, as you might imagine, that depends on whom you listen to.
One idea that seems to be a consensus among those who watch the industry is that the most interesting stocks to watch ? and perhaps buy ? are those that are angled toward communication devices and Internet infrastructure. For one thing, there is amazing growth in cell-phone and Internet usage. And for another, chips used in these applications have yet to become so everyday as to be commoditized, and therefore still show healthy product margins. Microprocessors and memory chips are another story.
"The companies that we've been pounding the table on are in the communications IC [integrated circuit] sector," says Merrill Lynch analyst Mark Lipacis. "You listen to the companies that buy their products: Cisco Systems (CSCO), Lucent Technologies (LU) and Nortel Networks (NT). Nortel's optics business grew 100% year-over-year; Lucent's optics business grew 60% year-over-year; Cisco's service provider business grew 80% year-over-year.... The buildout of the Internet infrastructure is not going to slow down anytime in the near future."
Using this theory, Lipacis points to PMC-Sierra (PMCS), Broadcom (BRCM) and Applied Micro Circuits (AMCC) as big beneficiaries.
Kevin Landis, fund manager of Firsthand Technology Value (TVFQX), has a similar take on the Internet-infrastructure industry. More specifically, he looks for companies that make chips used in local and wide-area networks (or LANs and WANs) such as PMC-Sierra and Transwitch (TXCC). (He is not currently buying PMC, however, because he says it constitutes a "huge" position in the funds he manages.)
He is also a big fan of chips used in the cell-phone world, though he offers a long-term caveat. "Wireless is really in the midst of a boom right now," he says. "Long term, you worry that cell phones have a lot of price pressure and the components get commoditized ? but for now, it's good times for everybody."
But it is the companies that make PLDs, or programmable logic devices, that he saves his highest praise for. According to Landis, these chips, which can be custom-programmed, unlike traditional "function-specific" chips, were originally thought to be too expensive for use in anything more than prototype products, but the price has come down sufficiently that there is no longer much of a "cost penalty" to use them. PLDs are used in network infrastructure, video games and other applications. Companies Landis singles out in this subsector include Xilinx (XLNX) and Altera (ALTR).
Amid all this jubilation, though, there are some sober words. Ashok Kumar, for one, an analyst at U.S. Bancorp Piper Jaffray and former director of marketing for the Pentium processor at Intel, sees current valuations for much of this sector as out of line with reality, and dangerously set for a fall.
"Typically in this sector, investors tend to be manic-depressive. The pendulum swings from one extreme to the other. It really doesn't stop in the middle," says Kumar. "Now, I think what would be rational and healthy for the market would be some valuation compression." That's analyst-speak for a significant drop in stock prices.
And on such stocks as Broadcom, PMC, Xilinx, Altera and Vitesse Semiconductor (VTSS), Kumar takes a less sanguine few than his fellow analysts. He points out that they are trading at historical highs ? and they are indeed stratospheric: Vitesse is trading at just under double its five-year average price-to-earnings ratio, Altera and Xilinx are trading at more than twice their averages and PMC almost triples its own.
The price appreciation is something of a self-fulfilling prophecy, Kumar argues, and will have to come to an end. "You have analysts continuing to ratchet up price targets not on higher earnings but on valuation expansion," Kumar says. "It's only to be expected that valuations would come back to the historical mean. That's an irrefutable law of nature. You always regress back to the mean as opposed to entering a new paradigm."
Kumar's views run counter to those of Merrill's Osha, who raised his price targets on 11 chip stocks Thursday, arguing against selling simply because prices have become so high. "In the past stocks have peaked when the growth environment deteriorates, and not because of valuation," Osha writes in the Thursday report that argues the sector is in the beginning of a three-to-five year growth spurt. "Semiconductor valuations can always be made to look reasonable by rising EPS estimates, especially at the heady top of the cycle. We are a long way from those halcyon times in the current cycle."
Which brings us back to Intel. Of all the chip stocks, Kumar points to Intel as one that is attractively valued, trading at P/E valuation that's one-and-a-half times its long-term growth rate as opposed to the average of three times for S&P 500 stocks. For the semiconductor giant, Kumar says, "There is little risk for value compression."
Doug MacKay, fund manager for Red Oak Technology Select (ROGSX) and a semiconductor bull, voices optimism similar to Kumar's, though for different reasons. Saying he "wouldn't read too much" into Merrill's downgrade or AMD's spurt, MacKay says, "Advanced Micro Devices has proven time and time again that it's not just about having good technology. Many times AMD has had a faster chip [than Intel] for a brief period of time, but they've blown it on the manufacturing side. For one reason or another, they've always stumbled."
MacKay's personal favorites are wireless-communication chipmakers and companies that make flash-memory products used in popular products such as MP3 music players, digital cameras and devices such as the PalmPilot. He points to Xilinx, PMC and Vitesse as big gainers. He also points to analog chipmakers ("Just awesome," he says) such as Maxim Integrated Products (MXIM), Semtech (SMTC) and Triquint Semiconductor (TQNT) (which is up over 600% this year).
It does bear pointing out that these stocks sure aren't cheap, and if Kumar's "law of nature" is indeed irrefutable, they have a long way to go down. But for those who like the risk ? and upside ? of technology stocks, it's not hard to argue that these stocks are here to stay and grow. According to the SIA, the fastest percentage growth from 1999 to 2000 will come in DRAM memory chips (39%), flash memory (36%) and telecom analog chips (30%). On the other end, computer microprocessor-chip growth will be "slow by historical standards," the group says, because of a maturing PC market.
That's why fund managers say it's best to work backward, by looking at wildly successful ? and quick-growing ? end products to see which kinds of chips go into them. Figure that out, and you know which chipmakers are likely to thrive. >> |
