Wintel: The High Growth in
Tech Lies Elsewhere
As more of us hook into the Net, demand for special
computer chips soars, and businesses you haven't heard
of are already profiting.
Richard A. Shaffer
Of course Microsoft and Intel are monopolies.
Together the two companies are the force that
IBM used to be, setting the computer industry's
direction and pace. But they aren't nearly as
dominant as IBM was in its day, and as was true
back when IBM was the target of trustbusters,
small companies continue to spring up and
flourish in the shadow of the giants.
Not that long ago, IBM was twice as large as all
the rest of the American computer industry
combined. Today Intel's $11 billion in annual
revenues from the U.S. market comprise only a
quarter of all U.S. semiconductor sales, and
Microsoft's domestic revenues of $8.9 billion are
less than a tenth of all U.S. sales of packaged
software. Meanwhile, consider just a short list of
smaller software companies that have prospered
in recent years--Citrix Systems, Legato Systems,
Network Associates, Siebel Systems, Veritas
Software, and Visio. Add in the burgeoning
semiconductor companies--such as Advanced
Technology Materials, C-Cube Microsystems, MRV
Communications, and Qlogic. My point isn't that
the government is somehow wrongheaded in
going after Microsoft and Intel, merely that
booming younger companies in high technology
aren't hard to find if you know where to look.
There will be many more of them in the years
ahead, whether the government wins or loses its
antitrust cases. One reason is that an
unprecedented amount of money is pouring into
new companies. Last year, according to research
by my company, venture capitalists put $5.8
billion into fledgling technology firms--a fivefold
increase in five years. Corporate capital is also
plentiful. Intel, for example, manages a $500
million fund for new companies, which is larger
than most venture capital partnerships. The
coffers of the public markets are also wide open.
To date this year, digital technology companies
have raised $4.1 billion through initial public
offerings. That's up from $2.5 billion in the
comparable period last year and $1.6 billion five
years ago. New telecommunications companies
are finding it easy these days to borrow hundreds
of millions of dollars apiece.
What's more, the computer and software
industries are changing in ways that render Intel
and Microsoft less important. For example, as
more of us use our personal computers to browse
the Internet, the quality of our overall experience
depends less on the speed of the
microprocessor--a market Intel dominates--than
on the speed of the communications network, a
market in which Intel is barely a factor. Because
the Internet is indifferent to the processor and
the operating system connected to it,
compatibility with personal-computing industry
standards, i.e. Intel and Microsoft standards, is
less vital.
So, where to look for high growth? Let me
suggest some answers by flying at high altitude
over one corner of the landscape, computer
chips--specifically those made by a host of
smaller manufacturers with nothing like the brand
presence of Intel.
Rising traffic on today's information
superhighway--the Internet--brings corresponding
increases on the feeder routes--local-area
networks. That's good not merely for vendors of
networking equipment but also for the
integrated-circuit companies that supply essential
parts. The smaller publicly held suppliers that
impress me are Level One Communications, MMC
Networks, and PMC-Sierra.
Level One, in Sacramento, makes chips that allow
computers to exchange data according to the
so-called Ethernet standard, which governs most
corporate networks. Those local-area networks
are getting faster, with many moving from zipping
data around at ten megabits per second to a new
standard of 100 megabits per second. Level
One's chips, which go into network equipment
that directs this speedy traffic, are in great
demand. Revenues and profits have risen
steadily for the past six quarters, and Level One's
profits in the March quarter were twice those in
the corresponding quarter last year.
In the past two years, PMC-Sierra, in Burnaby,
British Columbia, has reinvented itself,
abandoning a dwindling modem business to focus
on networking circuitry. Good move. The company
now dominates the market for integrated circuits
that enable high-speed communications
according to the so-called ATM and SONET
standards, and it's working closely with such key
equipment vendors as Lucent Technologies.
Although revenues and profits have been flat
recently, the company consistently earns more
than 25 cents on every dollar of sales, more than
double its former margins and five times the
semiconductor-industry average.
MMC Networks, in Sunnyvale, Calif., has been
public only since last October. Since the beginning
of last year, revenues have been rising over 25%
from quarter to quarter, and the profit margin is
going up too. The company makes networking
chips that are easily customized. Cisco Systems is
a customer. Significantly, MMC's technology gives
data-communications networks the reliability and
predictability of service needed to compete with
telecommunications networks. (I am an investor
in Cisco and MMC.)
There are a number of private chipmakers to
watch in this field, including Advanced
Communications Devices of Fremont, Calif.;
I-Cube of Campbell, Calif.; Jato Technologies of
Austin, Texas; Maker Communications of
Framingham, Mass.; and XaQti of San Jose.
Increasing demand for Net access, of course,
brings business to the many service companies
trying to connect homes and corporations to it,
such as the telephone companies and those
operating cable TV networks. Whether these
companies link up via cable modems, digital
subscriber lines, or television satellites,
integrated circuits manage the connection. The
standout is Broadcom of Irvine, Calif., which
enjoyed a blowout initial public offering in April.
Revenues soared in the past two years, as
Broadcom became the technology and
shipment-volume leader in digital circuits for
cable set-top boxes, fast local-area networks,
and satellite broadcasting. On the private side,
pay attention to Centillium Technology of
Fremont, Calif.; Libit Signal Processing of Tel
Aviv; and Ultracom Communications of Campbell,
Calif.
To keep pace with the demand for faster
communications, chipmakers are turning to
materials other than silicon. Vitesse
Semiconductor of Camarillo, Calif., for example,
uses gallium arsenide for integrated circuits that
enable high-speed networks of optical fibers to
connect with the slower telephone lines that most
businesses use for data traffic. The company is
growing 150% a year, and in its most recent
quarter, operating income almost doubled. As
data traffic surpasses voice on the world's
communications networks, only optical technology
will be able to provide the network backbones with
enough capacity. Gallium arsenide should play an
increasingly central role in linking the optical
backbones to the electronic networks we already
have.
Regrettably, I'm not the first to notice these
trends, which is why the public companies I like
are pricey and the private ones probably will be,
too, when their turns come. But all are well
aligned with what I see as the future of
computing and communications--and persuasive
evidence that in this forest the tallest trees aren't
blocking all of the sun. |
