Latest SmartMoney recognizes DSP technology. No mention of SSPI, but the big players.
>>>>>
March 18, 1999
Dow Jones Newswires
SmartMoney: Future Stock: Y2K And E-Commerce Are Yesterday's News; Here's What's Happening On The True Cutting Edge Of Tech - And How Investors Can
Dow Jones Newswires
No further information is available at this time.
This story appears in the April issue of SmartMoney magazine.
By Jersey Gilbert with Tiernan Ray
Read the headlines or watch the daily market coverage and you get the impression that savvy tech investing is all about catching the latest wave, riding it for all it's worth and then bailing out while the other guys get soaked. Netscape. Jump into the browser pioneer when it goes public. Sell out four months later, pocketing a gain of 194 percent. Keane. Ride the Y2K panic from March 1997 to March 1998. Exit with a 241 percent return. K-tel. Buy shares in the wannabe Web retailer last fall. Watch them soar 521 percent in five weeks. There's only one problem with chasing faddish, hot-and- cold stocks like these: Even if you manage to time your investments perfectly and sell out before it's too late -- a huge "if" -- you're not making nearly as much money as you could be. The fact is, within the tech industries there are trends and there are trends. And the long-term ones are almost always more profitable. Sure, that quick 194 percent return on Netscape looks pretty dazzling on the surface. But if you'd invested in, say, Cisco Systems five years ago and left it there, you'd be up 1,285 percent. Or how about Sun Microsystems: up 1,587 percent. America Online: 8,722 percent. Dell: an outrageous 14,445 percent.
These are the true tech superstars, and all of them have one thing in common. Early on, they came up with fundamentally different ways of doing business while everyone else was looking the other direction. Even AOL and Dell, which owe some of their recent gains to the Internet mania, have a history of innovation. AOL stuck with its idea of building a subscriber service when the "experts" were convinced that it would never work, and Dell perfected direct sales of PCs when everyone else figured you had to use a reseller.
Right now the tech world is on the verge of perfecting some huge new developments. That makes this a great time to become familiar with what's happening at the cutting edge -- and to get to know the companies spearheading change. "The Internet has shown us that we are on the verge of a tectonic shift," says Bill Keithler, manager of the top- performing Invesco Technology fund. More than just how we use our PCs, it's changing how we interact with the rest of the world.
But tech-stock prices are so high today that "you can't afford to pick companies that won't be sustained leaders," says Chip Morris, manager of the T. Rowe Price Science & Technology fund. Though Internet retailers and access companies are generating the biggest excitement in the market, "80 percent of these young, untested companies will not survive," he predicts. As the Internet grows, you'll be much better off investing in companies that supply the technology -- not the ones that need the technology.
To identify the most promising developments -- and the companies most likely to benefit from them -- we spoke to dozens of money managers, venture capitalists, analysts, engineers and company executives. Five key areas emerged: wireless transmissions, fiber optics, network access, semiconductor-chip design and semiconductor manufacturing.
If you are looking for just one company to play these changes, Lucent Technologies (LU) would be the logical choice. Its telecom-equipment business incorporates much of the advanced technology we outline below. And it has aggressively acquired other firms like the networking company Ascend Communications to augment its already impressive technological inheritance from the old Bell Labs.
The Murray Hill, N.J., firm is a sprawling, $136 billion telecom giant, however. No one product line, no matter how much new technology is incorporated, makes that large an impact on Lucent's bottom line. Conservative, blue-chip investors like that; more aggressive investors don't.
Wherever you invest in technology, computer chips are the most important agent of change, as has often been the case over the past 30 years. Analyst Drew Peck of SG Cowen breaks it down to a simple matter of logistics. "The effect of jamming more and more circuitry into smaller and smaller chip space results in a breathtaking rate of technological change," he says. In the world of silicon, smaller is better because it allows more information to move faster with less power.
New telecommunications uses are welcome news for the semiconductor industry, which is just emerging from a troubled three-year bout of factory overcapacity, chronic inventory surpluses and deepening doubts about just how much longer PC buyers will support 15 to 20 percent annual sales-growth rates. Memory-chip makers still face a long road back to financial health.
Unfortunately, while few of the chip designers at the cutting edge of telecommunications are household names, investors have not been shy with their dollars. After a furious rally to end 1998, all the key chip stocks have current price/earnings ratios well above 32-compared with a more common 15 to 30 over the past five years. Even these stocks' biggest supporters on Wall Street were leery of the premiums they commanded in late February.
So why bother reading this now?
Because those P/E ratios could plummet any day now. Remember, tech stocks are extremely volatile. Ten-year data from Frank Russell Co. indicates that tech stocks' price fluctuations are 50 percent greater than the market's, far exceeding those of other growth sectors like drugs or financial services. Over the past four years, there's been a tech correction about every six to 12 months. Prices of individual stocks often trade within a wide, 10 percent range, especially in the weeks surrounding quarterly earnings reports. Even mega stock Lucent dropped 34 percent last October. (If it were to suffer even a 20 percent decline in the next year, back to $82, we'd consider buying.) The high volatility can give you sleepless nights if you are inclined to worry about your portfolio; at the same time, it means there are many more buying opportunities available to a nimble investor. You have to move fast, however. As last October's furious tech recovery demonstrated, the opportunities don't last long. Stocks of the leading firms all doubled within two months. So know what you want to buy . . . and be ready to invest.
The Wireless World
The internet in the palm of your hand. That's what everyone is talking about
these days, since it's about to become a reality. 3Com says its highly popular
Palm Pilot will sprout a wireless Internet connection this year, while Motorola,
Nextel and Netscape are cooperating on a cell phone that should come out later
this year. What most people don't realize, though, is that that's just the
beginning. Interest in faster Internet connections will lead to a plethora of
wireless devices for use on the road, at home and in the office -- all within
the next three years.
AT&T aims to be a big player in the future of wireless. While its recently acquired Tele-Communications Inc. unit has a cable network that covers about half the country, AT&T plans to set up so-called wireless local loop systems in the areas where TCI doesn't reach. In these systems, radios mounted on buildings or placed in your home will transmit voice calls and Internet connections to untethered PCs. Hewlett-Packard and Intel, meanwhile, are laying plans for in-home wireless networks that will let you surf the Web from your easy chair or that hammock out on the porch.
In all of these developments, the goal is to develop a handheld device that can transmit radio waves at very fast speeds -- and with lower power consumption. The Holy Grail: an entire transmitter and receiver on one chip. The radio-on-a-chip is still a few years away. (The toughest part is coping with radio frequencies, which have a maddening tendency to drift.) But RF Micro Devices (RFMD) of Greensboro, N.C., is partway there, having figured out a way to integrate a few key radio functions on one chip. Right now RF is solidly entrenched as the No. 2 provider of power amplifiers for wireless devices, racking up 1998 sales of more than $100 million. Conexant, the renamed Rockwell Semiconductor, is No. 1, but the company is suffering losses in its conventional modem business.
As the unencumbered pure play on radio semiconductors, RF Micro Devices looks like the better bet on a wireless future. It has already overcome one important obstacle: Along with SmartMoney favorite Vitesse Semiconductor ("Ten Stocks for the Year 2000 and Beyond," October 1997), it is one of the few firms to successfully engineer gallium arsenide into its chips. The result is a chip that's up to 10 times faster than ordinary semiconductors, according to the research firm Dataquest.
The company has also been astute in its choice of business partners. IBM is helping it develop an inexpensive substitute for gallium arsenide. And it's one of the only suppliers of a key part to Nokia, which is increasing its share in the fast- growing cell-phone market, points out Mark McKechnie, an analyst at Montgomery Securities.
RF Micro Devices had one bad quarter in 1998, its first year of public trading. It lost 21 cents a share in the March quarter because an outside contractor botched a job, forcing the company to write off some inventory. Nevertheless, investors have been keeping faith in its future potential. The stock more than tripled from October to February and trades at a too-high 53 times next year's earnings. Young companies controlling key technologies certainly deserve a premium, especially when they're turning a profit. But we don't blame you if this stock strikes you as at $69.75.
No further information is available at this time.
A more reasonable 40 times next year's earnings -- equal to its projected three- to five-year earnings-growth rate -- would give it a price of around $53. If it gets to that level, we'd pounce.
Fiber Fever
Remember how grandly the telephone companies used to talk about their
expensive new fiber-optic networks? How they were going to take us all into the
21st century with their "pin drop" clarity and unbelievable capacity? Well, they
got the clarity part right. But the capacity part hasn't worked out quite like
they'd planned. Put simply, the world's fiber-optic lines are jammed tight.
Demand for capacity on the Internet is doubling every 100 days, according to MCI WorldCom's UUNet division. But that's not the whole story. Even plain old voice traffic is increasing 10 percent a year, as more phone companies are privatized and then lower prices to ward off competition. "They are approaching fiber exhaust almost every year as they struggle to keep up with the demand," says James Jungjohann, telecom analyst at CIBC Oppenheimer.
There are two main ways to fix the problem. You can either try to physically squeeze more fiber-optic cable into the existing rights of way (at a cost of $70,000 per mile), or -- the better option -- you can get the existing lines to carry more data. Here's how: By breaking up the light passing through fiber-optic cable, a technology called WDM, or wavelength division multiplexing (don't worry, there won't be a quiz), can increase the data contained in an optical transmission by at least tenfold. Some engineers are predicting that the ultimate expansion will be on the order of 200 times. That should hold back the tide of "fiber exhaust" for at least a decade.
All the big telecom-equipment manufacturers, like Lucent, Alcatel, Nortel, Ciena and Siem ens, are selling complete WDM systems to telecom providers all over the world. While the systems business has been a moneymaker for most of these firms so far, clearly there's more than enough competition.
On the other hand, there are only a few leading suppliers of some of the key components of any WDM system. Two of those, JDS Fitel (which trades on the Toronto exchange) and Uniphase (UNPH), are planning a merger that could easily make the combined company the dominant player in fiber- optic components. Together, the two companies make most of the key parts of an optical amplifier, a piece of equipment that boosts an optical signal as it travels over long distances. They also make chips that stabilize the light waves as they pass through the fiber. Shareholders are expected to vote on the merger in June. Assuming it goes through, the new JDS Uniphase will have combined sales of $420 million -- a figure that's expected to grow by 50 percent next year.
The key question is this: What will the company earn? Expectations among investors are so high -- Uniphase currently trades at over 70 times 1999 earnings -- that any merger-related rough patches will spell trouble for the share price.
JDS Fitel's biggest shareholder, Furukawa Electric, with 52 percent of shares, supports the deal but will take the opportunity to dump 3.5 million shares, says Jungjohann, whose employer, CIBC Oppenheimer, is the lead investment bank on the merger. Furukawa Electric will no longer be the majority owner in the new firm.
That suggests that the best time to buy the stock would be after the merger and any announced merger-related writeoffs. Uniphase stock currently trades at $87.63, and since it has been designated the continuing company, its share price is the key price to watch. From 1996 through 1998, as the first WDM networks were rolled out, the stock swung between 40 and 75 times earnings, with five big dips to the bottom of the range. It's not unreasonable to assume the stock will fall back to a P/E of 40 on any news that suggests uncertainty about the merger or in a general market correction. Projections for earnings per share without JDS Fitel were $1.77 in the fiscal year that ends June 2000. That makes $71 a target to start buying shares.
Good Signals
At the heart of every major computer system is a semiconductor chip that performs complex logic operations. In a desktop computer, this logic chip is called a microprocessor, and it is typically made by either Intel, Advanced Micro Devices, National Semiconductor or IBM. In a telecommunications system -- say, cell phones and the base stations they are linked with -- the logic chip is replaced with a math chip called a digital signal processor (or DSP).
Microprocessors had their most explosive growth spurt after the PC caught on, in the late '80s and early '90s. Now it looks as if digital signal processing is about to enter a similar period, driven by the demand for cell phones and Internet connections. More than $3 billion worth of these chips were sold last year, but as you can see from the chart at right, their growth rate exceeds and is expected to continue to exceed that of microprocessors.
One of the most appealing things about digital signal processors is the way they can be combined with other chips that read and interpret things happening in the real world -- like light waves, radio waves and temperature. For instance, the combination of analog and DSP technology overcomes radio signals' nagging tendency to drift. (The analog chip takes in the radio signal, and the DSP cleans it up and eliminates the "noise.") These dual chips are called mixed signal processors.
You have used a mixed signal processor if you've recently switched from an analog cell phone (the original kind) to a newer digital wireless phone. That's just one of the first high-volume applications of DSPs. Just about every type of consumer or business electronics product is being converted -- or someone is planning to convert it -- to digital processing. The television industry is already starting to roll out various versions of digital TV. Modems and disk drives are converting to digital processing. There are even plans in the works for digital radio transmissions, which will give you CD-quality music on your radio. Less sexy, but potentially more revolutionary, is the fact that wireless connectivity of all kinds will be possible. That is, you will be able to connect your PC with your laptop or your phone with your printer or your television's remote with your toaster, for that matter.
Four companies -- Analog Devices, Lucent Technologies, Motorola and Texas Instruments -- dominate the market for these chips. Of the four, only Texas Instruments (TXN) has completely reorganized itself to make DSPs its core business. Two years ago it shed laptops and defense electronics. Last year it sold its memory-chip division.
The results so far have been impressive. The company, looking to concentrate on a product with 20 percent or better margins, has replaced two businesses with margins under 10 percent and one business that was losing money. It has captured 45 percent of the pure DSP market and a big share of the mixed-signal market as well. And that is key, says Shekhar Wadekar, a Wall Street Journal All-Star Analyst at Raymond James. "Once you are the preferred vendor, it's hard for someone else to get into your market. Incumbency is a powerful thing in this business." The stock has quadrupled since the middle of 1996. "They're going where the chip market is headed," says Wadekar.
The downside: Texas Instruments is no longer an undiscovered turnaround story. Its current P/E, based on anticipated 1999 profits, is 32, higher than Intel's 28. That is because the Street expects TI's earnings to double over the next two years, while Intel -- which gets virtually none of its earnings from DSPs -- should expand earnings only by half. As the pure play in faster-growing DSP technology, TI probably deserves a slight premium, yet both stocks are going for way above their historical norms.
TI currently trades for $93. A pullback to $80 would give it a P/E of 21 times the projected earnings for 2000. That's consistent with the typical valuations Intel got from the market in the '80s and early '90s, when it was the dominant chipmaker in the fastest-growing segment of the business.
Faster Networks
The first thing you notice about the telecom revolution is that equipment
sales are exploding. The chart below is but one example. Demand for ports -- the
number of lines a network can accommodate -- is projected to grow by an
impressive 106 percent over the next four years.
What the chart doesn't show is the demand for speed. Network builders' biggest goal right now is making information move faster. That's because services of the future -- such as video on demand, real-time videoconferencing and digital audio transmissions -- depend on superhigh-speed networks. The key component? You guessed it. Semiconductor chips -- in this case silicon that moves data along the network and directs it to its proper destination. In recent years several firms have made solid profits providing low-priced chips to network equipment makers such as Cisco and Lucent.
As the demand for speed increases, the core semiconductors become more complex. That in turn increases the competitive advantage of getting to the superspeeds first. Right now PMC-Sierra (PMCS) is the clear silicon market-share leader in the type of network favored by phone companies. (It's called ATM; phone companies like it because it is as reliable as plain old voice calls.) The Burnaby, British Columbia, company also has partnered with Cisco to write standards for sending data along fiber-optic cables. It's a leader in corporate networks too.
Unfortunately, Wall Street is well aware that PMC has the key chips for all the different networks. Attracted by that, and by its operating margins of 25 percent, investors have nearly tripled its share price since last October. Now it sports a current P/E over 50. Yet the stock has had a P/E under 35 for most of its history. With analysts projecting per-share earnings of around $1.85 next year, you can expect that the stock -- now at $73.88 -- might trade for less than $65 sometime in the not-too-distant future. Grab it if it does.
The other company making waves in this sector is Broadcom (BRCM). Like PMC, this Irvine, Calif., manufacturer has several key networking chips. It is the leading supplier of chips for high-speed cable modems and controls key technology for satellite receivers. Sales rose to $203 million last year, from $37 million. Analysts predict Broadcom earnings will gain 47 percent this year and 35 percent in 2000.
Broadcom went public last year, and naturally, its shares soared. But the bubble sprang a leak at the beginning of this year, and its P/E slid from 116 to 75 times next year's earnings. Kevin Landis, manager of the San Jose, Calif.- based Firsthand Technology Leaders fund, echoes the sentiments of many when he says he'd never buy a stock with a forward P/E of more than 50, no matter how much he likes it (and he likes Broadcom). That suggests there would be a much bigger pool of buyers if the stock fell to $39, from its current $60.09, which would put it at 50 times expected 2000 earnings.
Wafer Madness
Wireless networks. Lightning-fast connections. CD-quality audio. They all
depend on the radical new chip designs we've just been describing. But in truth,
those chips are only part of the story. Making them is the other part. "Over the
next 10 years, there will be more changes . . . in semiconductor fabrication
than there have been in the past 40 years," predicts Morgan Stanley's Jay
Deahna, widely regarded as having the best database on capital-equipment
expenditures.
Right now equipment manufacturers like Applied Materials and Novellus are perfecting tools that can be used to produce substantially smaller, faster chips. But the tools that will be most in demand, according to Deahna's research, are what's known as lithography equipment. These machines use sophisticated lasers to lay out the minute electronic features on the surface of a silicon wafer. (The wafer is basically the foundation of a chip.) The ability of these tools to focus light within continually shrinking dimensions is the key to all future progress in digital technology.
Of the major lithography toolmakers, Netherlands-based ASM Lithography (ASML) looks most attractive. One reason: Three separate consortiums are working on the next generation of lithography equipment, and ASML belongs to two of them. Odds are that the company, whose stock has gained 27 percent since we profiled it in our March 1998 feature "Silicon Values," will be in the winning consortium.
Consensus estimates on Wall Street predict ASM Lithography will earn $1.22 a share in 2000, after the semiconductor industry recovers this year and starts buying equipment again. That would be a 352 percent gain over this year's expected earnings of 27 cents a share. Given that, the stock seems reasonably valued at $42.75. Yet this stock has traded below 3.5 times sales twice in the past three years. Deahna figures it's trading at 5.5 times sales now. As we mentioned earlier, if the semiconductor recovery turns out to be slower than expected, that will knock share prices down throughout the industry. There's a chance you'll be able to pick the stock up for $28 or less this year, if you are patient.
Another big winner in the plant of the future will be the suppliers of automation equipment. Consider this irony: Semiconductors are used in the most sophisticated applications of consumer electronics, but much of the manufacturing of chips themselves still involves workers carrying silicon wafers around between manufacturing steps.
The latest advances in semiconductor design -- which require a 300-millimeter wafer -- will bring an end to much of that. For one thing, 300mm wafers and their containers will weigh nearly 20 pounds. For another, a wafer with up to 3,000 chips on the surface will be worth well over $400,000 -- not something you want to drop. "The plant of the future will only need a man and a dog," runs the line at equipment conferences these days. "The man will be there to feed the dog, and the dog will be there to keep the man away from the equipment."
The leading maker of automated wafer-transport systems, with 90 percent of the North American and European markets, is PRI Auto mation (PRIA), an $820 million company based in Billerica, Mass. In recent years, building a new wafer-fabrication plant meant you had to buy about $10 million to $20 million of PRI's equipment. CFO Stephan Allison says it conservatively expects to sell $50 million to $70 million per factory as 300mm wafers become standard.
Right now PRI trades at 4.5 times sales, but it, too, has often traded at three times sales or less. This suggests you should jump into this stock -- now trading at a little under $41 -- at around $27.
|
