Posted 6/5/98
Archived 6/12/98
Jubak's Picks
--------------------------------------------------------------------------------
Check out Jim's Top Stocks
Intel price history
--------------------------------------------------------------------------------
See charts of Intel's price movement for:
1990
1991
1992
1993
1994
1995
1996
1997 Jubak's Journal
Time to cash in your Intel chips?
Pricing pressures from all sides are hurting earnings forecasts at the semiconductor giant. It may be time to get out -- but not just yet.
By Jim Jubak
What to do about Intel (INTC)? Down more than 22% in the past three months, the stock trades $30 a share below its August 1997 high of $100. And more trouble seems to be on the way.
Computer manufacturers are still sitting on huge inventories of unsold machines, nobody's buying computers in Asia, Intel's next-generation Merced chip won't make an appearance until 2000 -- about six months later than expected -- and the Federal Trade Commission is suing the company for trying to lock competitors out of the market. No wonder that, in the past 90 days, analysts have slashed their predictions for next quarter's earnings by more than 20%.
Of course, long-term Intel investors have been here before. In 1990, for example, the stock plummeted by almost 40% from May to October. In the spring of 1994, shares fell by 26% in a little more than a month. Each time, Intel bounced back. For example, in the six months after the 1993 tumble, the stock gained 64%.
Is it different this time? Should Intel shareholders -- and I'm one -- stay the course, or maybe even aggressively buy shares, because the stock will bounce back like it always does? Or is this the time to sell because something fundamental has changed?
I've put the question that way on purpose. The key to figuring out what to do about Intel now is to compare the current crisis to those the company has faced in the past.
Many companies -- and Intel certainly falls into this group -- go through periodic cycles of amazing prosperity and jarring crisis. The crises shouldn't be surprising: They are a predictable product of the company's business strategy. To the extent that Intel's current problems resemble those that the company has faced before, investors shouldn't worry or panic. In fact, if you decide that this is nothing but a replay of those earlier crises, you ought to stock up on Intel.
On the other hand, not every crisis is simply history regurgitated. Markets do change, throwing up new problems that a company may be unprepared to meet. Sometimes the kind of challenge is similar to those the company has successfully met in the past, but the magnitude of the problems confronting it is so different that the company simply doesn't have enough resources to win the battle. If Intel's current crisis is either new in kind or sufficiently greater in magnitude, then investors ought to consider selling their shares.
So let's play historian.
Details
--------------------------------------------------------------------------------
Company Facts
3-yr Chart
Press Releases
Earnings Estimates
Earnings Growth Rates
Profit Margins
Intel's business model has been pretty much the same for the last 10 years or so -- the years when Andy Grove occupied the CEO's chair at the company: Use massive research and development spending to churn out a string of ever-faster and more powerful processors for the personal computer market. Invest the high margins that came from having the chip that consumers wanted back into state-of-the-art manufacturing plants, so that Intel was the low-cost producer of chips. Cut prices for Intel products as fast as necessary so that competitors would never gain enough sales volume to be able to match the company's low manufacturing costs.
The strategy has, by and large, succeeded brilliantly. Intel owns about 90% of the microprocessor market. Its annual sales at $25 billion dwarf the $2.3 billion of the company's closest competitor, Advanced Micro Devices (AMD). Intel's gross margin is 67%. Advanced Micro Devices' gross margin is 47%.
The strategy, however, hasn't always worked. It is, by its very nature, susceptible to periodic crisis. Intel's high margins, its power to keep undercutting competitors' prices, and its ability to keep market share well above 70% all depend on the willingness of consumers to fork out the extra cash to buy the newest and most powerful Intel chip. Whenever the advantages of paying up haven't been overwhelmingly clear to consumers, Intel has faltered.
Let's go back to 1990, for example. Intel's stock traded at about $5 a share (adjusted for subsequent splits) at the beginning of the year. The price bumped between $4 and $6 until May 1992, when it decisively broke out to the upside. Within a year, the stock had doubled.
10-year history
--------------------------------------------------------------------------------
Investor subscribers can see a 10-year summary of Intel's figures in Financial Results.
Scan even a few numbers and it's clear that the price rise was justified. Intel's net profit margin, for example, went from 17% in 1990 and 1991 to 18% in 1992 and then to 26% in 1993. Return on equity rose from 18% in 1990 to 31% in 1993. Sales, which grew by 22% from 1990 to 1991 and again from 1991 to 1992, suddenly popped by 50% from 1992 to 1993.
What happened? By 1990, Intel had lost more than 50% of the 286 processor market to Advanced Micro Devices. Intel had introduced its first versions of the 386 chip in 1985 and the 486 chip in 1989, but they hadn't taken the world by storm. In fact, a disappointingly high percentage of consumers seemed perfectly satisfied with the speed and power of the 286 chip.
The introduction of Microsoft's (MSFT) Windows 3.1 in 1992 (following the introduction of the less compelling 3.0 version in 1990) changed that. That operating system, and the wave of new applications software that flowed in its wake, gave consumers a reason to buy speed and power. I still remember the conversations in my office then about who was getting the latest 486 machine, and the schemes I launched to see if I could put one on my desk. I even remember how envious -- and defensive -- I felt whenever I talked to a friend who had just upgraded his or her home computer. (It didn't hurt Intel's bottom line that a combination of legal hardball by Intel and engineering problems at Advanced Micro Devices delayed that company's 486 clone until 1993.)
Intel just delayed the introduction of the Merced chip until 2000. And since Merced is a brand-new architecture, market watchers don't expect the chip to sell in significant volume until 2001 at the earliest.
Something as ephemeral as consumer buzz is important to investors in Intel because of the way it affects the company's road map. Intel very carefully plots not just the dates when it will introduce a new chip, but how many dollars the new chip will sell for at introduction and how fast it will cut the prices of each new chip. For example, the Pentium II 450-MHz chip is expected to hit the market this summer at a price of approximately $780. By that point, the slightly slower 400-MHz version of this chip will be marked down to $590 or so (from about $830 in April). And the 300-MHz version, once the company's speedster, will sell for just $210.
It's pretty easy to understand the effect on competitors. When Advanced Micro Devices introduced its K6 233-MHz chip in March 1997, it was briefly the fastest chip in the world and sold for $469. Today, Intel's version of that chip goes for just about $200. Advanced Micro chips typically sell for about 30% less than Intel versions, so today the K6 233-MHz chip goes for something like $140. That's a huge $300-per-chip drop in a little more than a year.
Intel depends on adding high-margin products at the top end of its line to make up for this kind of price-cutting at the lower end. All will be well with Intel's bottom line if it can get consumers to buy Pentium II 450-MHz chips for $780 instead of Pentium II 400-MHz chips at $590. But Intel runs into real problems when a significant number of consumers decide that there is no especially compelling reason to pay up for more speed and power.
That's where Intel is today -- much as it was in 1990. Windows 98, which goes to market this June, is a relatively minor product upgrade -- certainly not enough to make anyone rush out to buy a new computer. Microsoft's next big product upgrade, the Windows NT 5.0 operating system, will stimulate big demand in the corporate market -- but that product won't hit stores until (probably) mid-1999. (Microsoft owns Investor, but I don't know anything more about the company's plans than is available from public sources.) And NT really won't jump-start the consumer market.
The Merced chip will enable Intel to gobble up the high-end, high-profit workstation market, but the company just delayed the introduction of that product until 2000. And since Merced is a brand-new architecture, market watchers don't expect the chip to sell in significant volume until 2001 at the earliest.
Earnings estimates
--------------------------------------------------------------------------------
Investor subscribers can go to the Analyst Info area to see current Earnings Estimates or the Consensus EPS Trend for Intel.
So it looks like Intel is in for an extended slowdown. Gross margins are creeping downward to 50%. Earnings will be lower for every quarter of 1998 than they were in 1997. I estimate 1998 earnings at $3.15 a share -- that's roughly a 20% decline from 1997's $3.88 a share. Given the continuing turmoil in Asia and the inventory still clogging the distribution pipeline for personal computers, that number could even be a little high. Thomas Kurlak, who covers Intel for Merrill Lynch and who has called the stock's weakness very accurately this year, has cut his 1998 estimate to $2.95.
So in round numbers, if Intel were to trade on 12-month trailing earnings come December and the stock commanded a price-to-earnings ratio of 20 -- about right, given all the uncertainty -- then shares would sell for about $60. That's about a 15% decline from current levels. Painful, but maybe worth sitting through if the company is going to bounce back in 1999.
But will it? Market analysts are starting to talk about a decline in the growth rate of personal-computer sales from 15% in 1998 to 13% in 1999. That worries me, of course. But I worry even more because of two crucial ways in which this crisis is different from past challenges to Intel.
First, I think the microprocessor industry is facing the same kind of massive overcapacity that has crushed prices for memory chips in the last couple of years. Advanced Micro Devices is adding capacity to its existing Austin, Texas, factory and has just broken ground for a new plant in Germany. When they're both running at full capacity they will, by themselves, be capable of producing two-thirds of all the chips needed by the entire personal-computer market.
Details
--------------------------------------------------------------------------------
Company Facts
1-yr Chart
Earnings Estimates
Company Facts
1-yr Chart
Earnings Estimates
Intel has been on a building binge too. Capacity at the company is expected to grow by more than 50% by 2000. Add in National Semiconductor's (NSM) new plant in Portland, Maine, and existing production lines that Integrated Device Technology (IDTI) is converting to microprocessor manufacturing, and the industry clearly faces a glut of supply at a time when demand is less than robust.
How this excess capacity will play out depends very much on the second trend that makes the current crisis different. The unexpected demand for computers selling below $1,000 has rocked the entire industry this year -- and no company more than Intel.
Sub-$1,000 machines now account for 40% of retail sales of personal computers. (The percentage sold to corporate accounts and by direct vendors is far lower.) Intel, which spent most of the year hoping that the sub-$1,000 market would go away, was late to market with a chip for this segment and won't actually have a competitive product in release until September. That chip, the Celeron 333-MHz with an on-die memory cache, is now expected to sell for about $210. At that price, Intel isn't making nearly the profit that it does on an $800 chip.
Chip manufacturing is a very odd business. The effort to produce a very high-speed chip produces a few of the desired products -- and lots of somewhat slower chips where something in the production process has gone just slightly awry. All that excess capacity, therefore, won't produce a glut of chips at the highest speeds, but a massive number of chips destined for the low and middle range of the market. Expect Intel to accelerate its price cuts in those market segments from even its current very aggressive rates.
The semiconductor industry clearly faces a glut of supply at a time when demand is less than robust.
That's obviously not going to have a positive impact on Intel's earnings. (It will, of course, have an even less desirable effect on earnings at Intel's competitors such as Advanced Micro and National Semiconductor.)
The majority of Wall Street analysts haven't surrendered yet, but I think the consensus will gradually factor in that mid-range glut and price-cutting. Already the consensus is moving toward the view that 1999 earnings are likely to be flat compared with those reported in 1998. Right now, many analysts are still talking about a second-half recovery for Intel. But at some point I think the market will realize that we're looking at the second half of 1999.
If I'm right, and we're looking at two lousy years for Intel instead of one, this clearly isn't the time to buy. Look at the analyst opinions: 11 out of 33 analysts have "strong buys" out on this stock. We're nowhere near a market capitulation. Right now, it's just too easy to pass off weakness in Intel as the annual summer sell-off in technology. The crunch will come in October if the September numbers fail to show a recovery in sales and earnings.
If you hold Intel now, you have, I think, two options. You can sell because you don't want to sit on dead money until the stock begins to move sometime in 1999 on news about Windows NT and the buzz on Merced. A year of zero return is a long time, and if you have a solid alternative investment and the discipline to get back into Intel in 1999, I think this is a wise strategy.
A year of zero return is a long time, and if you have a solid alternative investment and the discipline to get back into Intel in 1999, I think this is a wise strategy.
Or you can stay put and ride it out. That's not a completely crazy option. Timing a stock's bottom and top is difficult, and it's probably impossible to hit the absolute high or low. I certainly can't tell you how far ahead of any renewal of Intel's earnings growth the stock might start to move. The market always anticipates good news and that's especially true with stocks that are as widely held and deeply loved as Intel is. Investors really want to believe in Intel, and that both puts a floor under the shares and makes it likely that the price will go up on the slightest whiff of a turn in the company's fortunes.
Me? Well, remember that Jubak's Picks is run on a 12-month time horizon. Waiting for a recovery that's 18 months away just doesn't fit that portfolio. I'm going to start looking for an exit point for the Intel shares I now hold in Jubak's Picks. I might have to be patient and wait for a September rally, but selling into the current technology sell-off isn't my idea of good timing. I'll let you know in this column when I find a better moment.
I'll also let you know when I get back into the stock. I remain convinced that an investor should own Intel for the long haul. (So I'm keeping it in the 50 Best Stocks in the World portfolio that I will update next in August.) None of the company's competitors are strong enough to mount a meaningful attack on Intel's control of the middle and upper end of the microprocessor market. It's clear that Intel-based servers will continue to gain market share among network and database customers even without Merced. And when that chip eventually appears, it will gain a dominant position in the highest-margin end of the market.
Just remember that time is money. Every day you wait for an eventual turnaround cuts into your return. I don't think there's anything wrong with getting in early on a stock. But there is such a thing as too early.
|
