Froche, Dennis;
here is another view:
December 7, 1998
Are Chip Equipment Stocks Headed for Disaster?
By Tiernan Ray
CHIP EQUIPMENT STOCKS
IS THE MARKET for semiconductor equipment stocks efficient? Recent run-ups of 40% to 50% for stocks such as Lam Research (LRCX) and Novellus (NVLS) may look like a clear sign things have turned around for the companies that sell to Intel and other chipmakers. But the outlook from industry observers is sobering: It suggests these stocks may see some substantial declines in the near future.
Equipment stock investors have lots of news to crow over. First, several analysts upgraded their ratings on chip companies in the middle of last week. Then on Friday, the Semiconductor Industry Association's latest report on chip sales was leaked to Bloomberg news service. Those figures show a higher-than-expected rise in chip sales of 6.3% between September and October. Boosters of chip equipment stocks say these developments are further proof the chip market has rebounded, and that the equipment makers are sure to follow. The appreciation in shares is merely anticipating that rebound, they say.
Analysts say it's just not so. "We're at a point in the [equipment] cycle where we have no visibility as to earnings and estimates," says Merrill Lynch's Mark Fitzgerald. "I'm sitting here and I don't have a high degree of confidence that companies can make their numbers over the next quarter or two." Fitzgerald and other analysts say that partly because of short order lead times, and partly because of the uncertain technology roadmap in 1999, the rise in shares may have priced in hopes that are unsustainable going into 1999.
"Orders booked bottomed at a level we haven't seen since early 1993, and that's grossly insufficient to support profits for all the companies in this industry," says Theodore O'Neill, who tracks equipment companies for Needham & Co. O'Neill maintains that the upturn in equipment orders since this summer is relatively weak compared to previous cycles, and therefore, the upswing will not lift all stocks. Almost 50% of October's equipment orders, he notes, went to Applied Materials (AMAT) and Novellus only. That means that while the industry has indeed hit a bottom, not all companies will benefit from the recovery the way these two will. That makes investing all the more precarious.
Like the tides, most investors assume that when industry fundamentals have bottomed in the equipment sector, it's automatically time to get back in whole hog. But Chris Chaney, who tracks Intel (INTC) for AG Edwards, and who also follows equipment makers Applied Materials and Teradyne (TER), thinks that even though the industry has bottomed on a fundamentals basis, expectations for growth in 1999 are just too high. "We'll see some sequential growth in the equipment market in the first quarter, but 10% growth [in semiconductors] in 1999 is not going to be enough to raise the rate of growth in capital equipment," says Chaney.
Chip equipment sales are expected to rise 23% in 2000, but they will only see growth of about 5%, at a maximum, in 1999, according to Chaney. That's after a year-over-year drop of approximately 29% in 1998, according to the research firm VLSI Research. The financials of most chip equipment companies have yet to recover, and indeed earnings estimates continue to decline for many companies in the sector. The result will be a long slow recovery in 1999 in which investors' patience will be tried, and any big disappointment could spell disaster.
Nonetheless, Merrill's Fitzgerald admits that the market has pretty much left the analysts behind. "It seems every time we lower our numbers a couple points, the stocks go up," he says.
Why should investors be worried? There are at least three broad dangers that could depress stock prices before any recovery in those balance sheets:
1. Capacity Glut A worldwide glut in chip capacity, exacerbated by the Asian financial crisis, has been pushing down prices on chips this year. Despite the fact that prices for DRAM memories have improved in recent months, the world's semiconductor manufacturers still have tremendous capacity left over, and capacity could increase dramatically next year, with or without the demand to sustain it. Firms throughout Asia shut down DRAM factories this summer to throttle back capacity, but many think Japanese and Korean memory makers, desperate for cash, might decide to open up some of those shuttered fabs and starting churning out more memory products once again. "If the Koreans get hold of a ton of cash in '99, it'll just lead to deflation in IC [integrated-circuit] pricing again well into 2000," says Chaney, who's not counting out such a scenario.
While the effects of a reopening of fabs in Japan and Korea wouldn't be felt until later in 1999, even hearing that capacity is again roaring forward could cause investors to assume the rise in chips will be cut short, and that could depress equipment stocks.
2. Intel's Shifting Business Model Intel is the single-most-conspicuous cause of equipment makers' woes. PC microprocessors represent 40% of all chip sales, and with a commanding share in that market, Intel's take-home also supports much of the equipment market. Applied Materials alone derives 5% to 10% of its sales from Intel. But as pricing in the PC market has eroded, Intel's winning business model is coming under attack. "Intel has historically built at the technological cutting edge, to try and sell top-of-the-line machines, but the company didn't do it in a very cost-effective manner," says Fitzgerald. "With the eroding value of the PC, Intel has had to focus on cost cuts."
The company has been bailing water this year, canceling equipment orders wholesale. If the dollar value of PCs continues to decline in 1999, Intel's spending on equipment will probably decline proportionately. Fitzgerald estimates Intel's capital expenditures could fall as low as $2.8 billion in 1998, down from 4.2 billion this year. "Intel is the big factor that could throw equipment buying into the negative column," in 1998, says Fitzgerald.
3. The Uncertain Technology Timeframe What keeps the equipment industry moving forward is the continuous need to move to the latest and greatest technology. But some bets by equipment players have already flopped, and others may disappoint in '99. Spiffy new 300-millimeter wafers were supposed to be all the rage in 1999, but with a glut in chip capacity, Intel and others are holding off on buying equipment until after the millennium.
Copper-chip technology, which helps chips run faster as they become more complex, is a safer bet than the 300-millimeter-wafer technology, but it's clear that 1999 will be a year of experimenting with these chips, rather than putting them into high-volume production. "There's still some mileage in the existing technology before people have to run to copper," says Martin Reynolds, a chip analyst at Dataquest. Using better design, called "pipelining," Reynolds says Intel and others can add speed to chips before laying out cash to move to copper. That could be bad for Novellus, which has bet heavily on copper.
Likewise, John Schuler, who follows equipment sales for trade group SEMI, says that chipmakers have learned how to use existing lithography technologies to go to smaller chips than previously thought possible. "Technologies such as CMP [chemical mechanical planarization] have made it possible to use the existing lithography to obtain smaller line widths," thereby obviating some cutting edge technology, he says. Carl Johnson, president of the research firm Infrastructure, agrees. "There's going to be no major increase in orders to move to 0.18 micron [more sophisticated chips]," says Johnson, "because the industry is figuring out how to reuse the equipment that's already in place."
So what's the scenario for '99? As some new equipment orders get delayed, look for investor disappointment with the sector as a whole and for greater consolidation in the industry. If you're in a betting mood, look first at Applied and Novellus. If Needham's O'Neill is right, these two companies are stealing share so fast that they may be the only reasonable stocks in the group. O'Neill also likes buying the suppliers to these two, such as Applied Science & Technology (ASTX).
Advanced technology like copper probably won't amount to much in '99, while vendors of older wares, such as Mattson Technology (MTSN) and Gasonics (GSNX), will have trouble staying above water thanks to a lack of global presence, according to Fitzgerald. IPEC (IPEC) and Speedfam (SFAM), two copper-chip companies that agreed to merge two weeks ago, are a typical takeover target, most likely for Novellus or KLA Tencor (KLAC), which are battling Applied's march into copper. Overall, most spending will go to traditional goods that Intel and others will be stocking up on, such as photomask equipment. That should benefit mask-making companies like Etec Systems (ETEC) and Dupont Photomasks (DPMI).
At the end of the day, though, it'll be difficult to separate the wheat from the chaff because, as one analyst points out, "These stocks all trade together. They're so highly correlated that you could just throw darts at the names and construct a portfolio out of three of those dart hits," and do just as well. With that kind of confidence level you may be better off avoiding the group altogether.
|
