(FORTUNE Magazine) – Something is amiss in the land of computers. On March 4, Intel warned that first-quarter results would come in below expectations; Compaq followed two days later with its own confession. Both stocks plunged more than 20%, and the entire personal computer sector seemed to quake. "We are cautious on personal computer stocks," says Frank Walsh, a Putnam technology analyst. J.P. Morgan's Daniel Kunstler is more blunt: "Stay away from PC stocks."

The proximate cause of the industry's distress is bloated inventory, particularly at Compaq. But the PC makers' fundamental problem runs deeper. Simply put, improvements in hardware speed and power have outrun users' needs so much that the sector must now compete more by luring customers on price than by getting them to upgrade to better technology.

Moore's Law, one of the industry's basic precepts, observes that computing power tends to double every 18 months. Until now, that has meant that PC makers could introduce a vastly improved line of products every year and a half for roughly the same price as the old line. But when people stop caring about more sophisticated computers, an ugly corollary kicks in, which Walsh calls Moore's Revenge: If customers don't demand performance, they'll demand lower prices.

As it turns out, many people do not need faster computers right now because most software already works pretty well on current hardware. The next versions of Microsoft's Windows and NT operating systems, for instance, are likely to work just fine with regular Pentium or Pentium Pro chips, instead of the newfangled Pentium IIs. Chip Morris, manager of the T. Rowe Price Science and Technology Fund, explains, "No one is bringing out new applications that suck up computing power, so people have shifted to buying the state-of-the-art at the lowest cost." The average price of a desktop PC has been declining steadily since 1995. And now, since unit growth is expected to taper off this year from around 17% to 15% or even lower, those price declines will cut deeper.

In this environment, most of the PC stocks probably have farther to fall. On the bright side, that means that the next few weeks could give you the chance to get into some of the industry's kingpins at appetizingly low entry points. Here's the outlook for four of those stocks, starting with the most troubled.

KEEP AN EYE ON COMPAQ'S INVENTORY: The most prominent victim so far of Moore's Revenge, Compaq is an object lesson in how not to manage inventory. Compaq apparently hoped to meet its ambitious revenue targets by shoving excess product into its distribution channels. The strategy worked in the fourth quarter. At that time the company was able to meet analysts' expectations, but only by dumping servers and desktops onto the market. (While server and workstation products made up nearly half of Compaq's $25 billion in sales last year, they accounted for about two-thirds of its $2.7 billion in operating profits.)

But the plan backfired this year. Steep price cuts in February from IBM siphoned off some corporate buyers, and Compaq's inventory has backed up to an estimated eight to ten weeks' worth. Now the company can't ship new products until its distributors clear out what they already have.

The logjam couldn't have come at a worse time for the company. For one thing, Compaq is in the midst of overhauling its distribution system, trying to cut the amount of PC inventory held by its wholesalers and resellers to two or three weeks. The company is painfully aware that the value of the components it uses to assemble a PC decrease 1% to 2% a week. Thus, a computer sitting in inventory for eight weeks commands a significantly lower price than one built with the newest components.

Because the PC maker has been dumping so much product out the door, however, its distributors will have to clear out at least $2 billion worth of goods just to reduce inventories by four weeks, estimates Mark Specker, an analyst at SoundView Financial. Compaq has started offering steep discounts and incentives such as free monitors for products currently in the pipeline. Goldman Sachs' Richard Schutte figures that even before this mess, Compaq's operating profit on PCs was only about $100 a box. The company will be lucky to break even this quarter.

On top of that, Compaq has also embarked on a fundamental transformation of the company from computer hardware manufacturer to full-spectrum supplier of computer goods and services worldwide. That is the goal behind its recent $9 billion acquisition of Digital Equipment Corp. "Given the future Compaq chose for itself, the DEC acquisition was the right one," says Specker. But he wonders, "Was it the right future to choose?" The acquisition contributes to CEO Eckhard Pfeiffer's goal of getting big fast, but it will take a considerable amount of work to bring DEC into competitive trim. Not exactly a paragon of efficiency, DEC employs 64% more people than Compaq to produce about half the volume of sales.

Since Compaq's earnings warning, Wall Street has cut the company's 1998 profit estimates in half. The stock trades at 28 times those reduced estimates. That's a high price for a company simultaneously faced with winnowing inventory, revamping its distribution system, integrating DEC, and fighting off a hungry set of competitors. Bottom line: Stay clear of Compaq until there's more evidence that the company has a handle on its problems.

WATCH HP'S PROFIT MARGINS: Hewlett-Packard has a more manageable three to five weeks of inventory, but it is unlikely to escape the pressure caused by Compaq's troubles. HP--along with Compaq, IBM, and other PC makers who sell wares through resellers--faces a prisoner's dilemma. If the competitors could cooperate to bring inventories in the reseller channel down to, say, two weeks, all of them would benefit. But sticking together will be difficult. If HP cuts inventories to two weeks, that just opens up more room on distributors' shelves for Compaq's wares. Compaq might then slash prices to gain share (or to clear inventory), and HP might have to match. Deutsche Morgan Grenfell analyst Michael Kwatinetz concludes that Compaq's "aggressive actions could lead to an irrational pricing environment, pressuring margins for everyone." In other words, if the skirmish turns into a price war, HP's PC business may see little profit for some time.

Luckily for HP, PCs account for less than 20% of its business. Another third comes from printers, where HP controls about half the market, and the rest from software, Unix computers, and scientific -testing equipment. Although Japanese makers of printer hardware, like Epson and Canon, are competing fiercely on price, the real earning power in that business comes from the recurring sales of high-margin supplies such as ink cartridges. As a result, printer price wars could also shave HP's margins, but not as sharply as in PCs.

And interestingly, HP may be one of the few tech companies actually to come out ahead in Asia's crisis. While the company derives 16% of its revenues from the region, Salomon Smith Barney's John Jones estimates that HP manufactures as much as 40% of its products there. In other words, its Asian sales are weakening, but its costs are dropping at the same time.

HP's $63 stock is pretty much fairly valued at 19 times this year's earnings estimates. And if HP stock simply tracks its expected earnings growth rate, that would translate to a 15% annual return. But for those expectations to pan out, HP needs to continue building its $43 billion in revenues without squandering its 34% gross margins on less profitable businesses. Making PCs, for now, is not turning out to be much of an answer. And as J.P. Morgan's Kunstler, who rates the stock a long-term buy, points out, the challenge to margins will get more difficult as the company gets bigger. "The law of large numbers is working against them," he says. Unless the stock dips to the upper 50s, let HP tackle its challenges without you.

GRAB INTEL UNDER $70: Intel CEO Andy Grove does not think that Wall Street gives his company the P/E it deserves. Yet even after its recent 17-point drop, the $77 stock still trades at 24 times this year's estimates. Perhaps Grove is thinking that a P/E in line with Microsoft's 50 would be more appropriate, but then Microsoft did not just announce that revenues this quarter would come in about 10% lower. Wall Street expects this to be the first year since 1989 that Intel's earnings will actually decline.

Intel's bum numbers are a direct result of the inventory backup among PC makers, who are far and away Intel's biggest customers. And with the growing popularity of sub-$1,000 PCs (International Data Corp. predicts that one third of all domestic PCs will be in this category by 2001), hardware manufacturers are finding it difficult to give Intel the $250 per chip it was getting a year ago. Merrill Lynch's Tom Kurlak, the Dark Prince of semiconductor analysts who has damned Intel with his "neutral" rating since last summer, thinks the average price of Intel's microprocessors will decline 28% this quarter compared with last year, while their unit sales will be up only 12%. Says Kurlak: "The PC as an end market is no longer able to drive Intel as a growth stock."

What Intel needs is to enter new markets where it can charge higher prices. Three years ago, Intel was banking on entertainment applications and streaming video over computers to drive demand for its latest chips. Today, it is much more excited about "Intel-based" servers. Kurlak agrees, but he thinks Intel's best ticket into that market is its Merced chip, designed for servers and high-end workstations. Merced is not slated to come out until next year and is unlikely to become a major contributor to Intel's bottom line until after 2000.

Morgan Stanley's Mark Edelstone is not so gloomy. He notes that Intel's stock tends to bottom out in quarters when revenues from the prior-generation microprocessor are surpassed by those from a new one. The Pentium II is already established in servers and workstations, and is finding a place in PCs. "We are rapidly approaching the sweet spot of the product cycle for the Pentium II," he says. In addition, a stripped-down Pentium II dubbed Celeron (sounds like a high-tech vegetable, doesn't it?) will be introduced in April. Edelstone says that could help Intel hold its own in the sub-$1,000 PC market.

So, whom to believe, Kurlak or Edelstone? Kurlak's fundamental analysis has been right on the mark. But as one money manager cautions, "Kurlak would like nothing more than to have the entire world throw in the towel on Intel, so that he can then come in and recommend the stock." Be that as it may, it's fairly certain that Intel's stock will be choppy during the next couple of months. Look for a chance to buy this bellwether below $70.

DELL IS THE CLEAREST BUY: Dell claims that it has been unscathed so far in the price battle that's bleeding Compaq and HP. And indeed, it stands to reason that a company that builds to order and sells directly would be insulated from inventory-related problems. Compared with other PC vendors, who turn their inventories over maybe 10 to 12 times a year (not counting their distributors' own inventory delays), Dell turns its inventory more than 30 times. In addition, Dell mainly serves corporate technology buyers, who are less sensitive to price and more concerned with performance. (That's one reason Dell's average sales price is $500 more than Compaq's.) Even so, says T. Rowe Price's Morris: "It would be fantasy to think that Dell can totally escape the industry's pricing pressures."

Dell has several options to maintain profits when the squeeze hits. For starters, it is aiming to take more custom orders over the Internet, a sales channel that is particularly popular with consumers and small-business owners, two kinds of customers Dell is keen to win over. Internet sales currently run more than $4 million a day, and Dell hopes to increase that figure to half the company's revenues (which were $12 billion last year) over the next few years.

The company is also working hard to sell more servers. Goldman Sachs' Schutte estimates that if servers made up as large a proportion of sales for Dell as they do for Compaq, Dell's overall operating profit per unit would shoot up from $275 to $400, a full 47% higher than Compaq's operating profit. Server sales more than doubled for Dell last year.

All indications are that Dell is doing fine; yet some of the industry's uncertainty has crept into the stock recently, taking the price from its high of $70 to a recent $63. DMG's Kwatinetz thinks Dell can keep earnings chugging along at a 30% annual pace or better through the end of the millennium. However, with the P/E at a lofty 34, the stock is likely to be susceptible to turbulence in the rest of the industry. If you have a chance to buy Dell at a lower price in the next few weeks, do so. Anything in the 50s is a steal.