By Olga Kharif Back in 1998, the world's largest chipmaker, Intel (INTC), decided to invest billions in new fabrication plants that could turn out unprecedented numbers of ultrapowerful microprocessors, and do so at unit costs of up to 30% less. Most of Intel's rivals made similar plans, only to backtrack in 2001, during one of the steepest sales slides the chip industry has witnessed. Intel persevered and has spent $25.7 billion on capital improvements since 1998 in an effort to build an insurmountable market advantage before the next recovery arrives.
Intel is almost certainly going to get that advantage. The first of two huge new fabrication plants, a facility of 1 million square feet, revved up in October in Rio Rancho, N.M. A smaller fab will open in 2003, and an additional giant is scheduled for 2004. Furthermore, in the second half of 2003, Intel plans to roll out new process technology at a fab that will make the tiniest microprocessors yet for consumer-electronics products. No one can match Intel's pace of routinely knocking out $2.5 billion factories.
SHORT-TERM RISK? All this activity must have investors and analysts purring, right? Well, not really. While Intel is still focused on its time-tested strategy of planning for the long term, those who follow the stock are fixated more on the here and now. It isn't as though Intel is losing money: In the third quarter, it delivered net income of $686 million (10 cents a share), on $6.5 billion in sales. And at around $20, its stock is up 55% from its 52-week low of $13.22, which it hit on Oct. 8.
Yet Intel shares are now trading at about 31 times the 61 cents a share the outfit is likely to earn in 2003, according to the estimates of analysts polled by First Call. That's just too rich a valuation for analysts such as SG Cowen's Mark Grossman (neither he nor his firm owns the stock or does business with Intel), who recall that, in the late 1980s and early '90s, Intel's revenues grew by 30% annually, while its stock traded at 15 times the next year's earnings.
Moreover, if your focus is on the next quarter instead of next year, the chip king's expansion schedule looks risky. Rio Rancho alone boosts Intel's capacity 15%, estimates George Burns, president of chip-development researcher Strategic Marketing Associates in Santa Cruz, Calif. Meanwhile, the market is so slack that even a recent 27% price cut for Intel's 1.4-gigahertz Pentium 4 processor, which is being supplanted by its new 3-GHz Pentium, isn't spiking sales. Factor in rising depreciation expense that could further squeeze the bottom line, and analysts are wary of the stock in the short term. Merrill Lynch's Joseph Osha, ordinarily an Intel booster, downgraded the stock from neutral to sell on Nov. 14 (Merrill expects to do investment banking for Intel, according to recent disclosures.) And of the 28 other analysts who cover the stock, 15 rate it a hold.
Such attitudes partly reflect the stock's high valuation. But analysts are also worried about what could be a fundamental shift in the economics of the chip business, where rapid declines in chips' prices vs. the performance they deliver has always been more than offset by burgeoning demand for more powerful processors. Some chip executives think that the historical pattern is changing (see BW Online, 11/26/02, "For Chipmakers, Less of Moore?"). Sure enough, Intel's sales of processors for PCs, which account for more than 80% of its revenues, will grow only 8.5% in 2002 and 5% in 2003, estimates Dean McCarron, president of chip-market researcher Mercury Research in Scottsdale, Ariz. In the 1990s, such products enjoyed double-digit growth.
FRESH EFFORTS AND PRODUCTS. Now, to achieve even that 5% expansion expected in 2003, Intel may have to work extra hard, says analyst Eric Ross, with Investec in New York. Its strategies include trying to boost chip sales in India and Russia, where PCs' household and business penetration is a low 10%. Next year, it will unveil its groundbreaking Banias chip, which will streamline wireless access to high-speed Web networks and help lighten laptops. Intel will also release its next-generation Pentium processor, Prescott, which will run at 5 GHz, a plus for high-bandwidth Net access.
Still, it could take until mid-2004 for these products to fully permeate the mainstream market, says Rob Enderle, an analyst with technology consultancy Giga Information Group in Cambridge, Mass. And Prescott should face competition from Intel's main rival, AMD (AMD), which will also introduce a new desktop processor in 2003.
To compensate for the PC slump, Intel is moving into other markets -- a time-consuming task for a $26.5 billion outfit. Intel should increase its market share for flash memory used in cell phones from its current 20% to 80% by yearend, according to its own projections. During this period, the flash-memory market should grow from $6.9 billion to $8.2 billion. Yet because of Intel's aggressive pricing, it loses $200 million per quarter in that business -- and won't predict when it will turn profitable. "They do need to diversify," says Grossman. "But they're sacrificing [their margins and earnings growth] near term." Intel's operating margins have dropped from 39.4% in 1997 to a projected 17.7% this year. Bear Stearns analyst Charles Boucher expects a rebound next year -- but only to 19.6%, and that figure assumes some ramp-up in demand.
Intel execs see the falloff as temporary. "We're always evaluating what's happening in the marketplace, and we're going to make adjustments based on the outlook," says Mike Splinter, executive vice-president and director of sales and marketing for Intel. "We'll grow the company whether the economic recovery happens or not."
CLAMOR FOR SAVINGS. Yet, backseat drivers on Wall Street argue that unless demand improves early next year, Intel should concentrate more on cutting costs, such as its research and development expenses. In 2001, R&D spending was 14.3% of revenues and is up 5.2% this year, to $4 billion. Capital spending is another area where some think cuts should be made. At $4.7 billion this year, it's already 39% below what it was in 2001. Next year, Intel should only spend $3 billion to $3.5 billion on capital projects, suggests Banc of America analyst Marc FitzGerald. He says Intel could achieve that savings by stopping work on a fab in Ireland due to open in 2004. Fitzgerald also suggests that Intel send 5% to 10% of its less-essential manufacturing to outside contractors.
One benefit of such moves, analysts say, would be to lower Intel's depreciation, which, as a percentage of cost of goods sold, has risen from 20% in March, 1996, to 35% this year. FitzGerald thinks it could rise again next year, hurting Intel's profits unless sales pick up dramatically.
The other candidate for the chopping block could be headcount. Intel expects to cut its existing workforce to 79,200 by yearend, down 13% from 91,000 in the first quarter of 2001. "As long as revenue stays flat, you should expect to see us continue these types of programs," Intel CFO Andy Bryant said during an October Webcast. Intel could announce further cuts in April if the market doesn't pick up, believes Michael McConnell, an analyst with Pacific Crest Securities in Portland, Ore.
CAUTION ADVISED. That's the pessimistic outlook, of course. The other possibility is that corporate spending on PCs will stage a faster-than-expected comeback in 2003. In any case, Intel's stock could edge up thanks to its plan, announced on Nov. 14, to repurchase 480 million of its 6.6 billion outstanding shares. Beyond that, it's hard to ignore the fact that Intel's history of not retreating in the face of temporary lulls has generally paid off.
Even so, most analysts remain cautious and recommend avoiding Intel until the chip recovery becomes more pronounced -- or the stock drops. Howard Sutton, the portfolio manager for the $300 million Tera Capital Global Fund in Canada, is going so far as to short Intel shares, which he expects to fall to $10. "Intel has got to have high [production] volumes to drive margins," he says. "Investors should be cautious."
That may strike many investors as true, especially the fainthearted who worry that Intel's tried and true formula for success -- planning and investing for the long term -- has already seen its best days. Kharif covers technology for BusinessWeek Online