Posted 5/21/99
Jubak's Picks
Check out Jim's top stocks
for the next 12 months.
50 Best
See Jim's list of the 50 best
stocks in the world for the
long term.
The important
question, I think, is
what happens when all
those pipelines get into
the home. And the
answer, I think, is
they'll create a huge
new market.
Jubak's Journal
Who wins when life goes digital?
Look down on the server farm for the hottest
prospects. My money's on companies like Oracle,
Pervasive Software, Exodus and Wind River.
By Jim Jubak
I'm as fascinated as the next investor by the war
to decide who will own the communications
pipeline into the home -- what's known as "the
last mile."
AT&T (T) launches a preemptive $60 billion bid
for MediaOne Group (UMG) one week and then
the next week upstart Global Crossing (GBLX)
gobbles up venerable Baby Bell U S West
(USW) -- or maybe it's the other way around --
for $37 billion. And I don't expect to get bored
any time soon. Every communications company
that can raise a dollar is going to leap into the fray
to make sure it owns a piece of the copper wire,
coaxial cable and wireless systems that will
deliver signals into the home. (For more on this
battle, see my May 18 column "Why AT&T
goofed in the long run.")
But frankly, I suspect that understanding the
implications of this struggle over the next three to
five years could well result in bigger profits than
I'll earn by picking the winners of today's
attention-grabbing battles.
The important question, I think, is what happens
when all those pipelines get into the home. And
the answer, I think, is they'll create a huge new
market.
How? The multiple pipelines into the home are
going to be connected to an expanding variety of
general-purpose devices such as PCs, TVs,
telephones and game consoles. They're also
going to connect with an even greater array of
special-purpose appliances such as PalmPilots
and pagers that deliver stock prices, and wireless
phones that serve as Web browsers. Somebody is
going to make big money out of connecting all
these devices.
How I knew the trend was real
That's not a profoundly original thought. I've
been seeing increasing numbers of media stories
over the last year about the emerging
home-networking market. And when Intel
(INTC) paid $2.2 billion to snap up
communications chip maker Level One (LEVL)
this spring, I knew the trend was real. Level
One's chips, added to Intel's own expertise and
the knowledge the company acquired when it
bought networker Shiva in 1998, are ideal for
attacking this market. Note also that Intel
launched its AnyPoint Network for linking
multiple home PCs over existing copper phone
lines this year.
I think that's likely to be a respectable business,
once Intel figures out how to make getting one of
these home networks up and running a whole lot
simpler. But by itself, this isn't the huge market
I'm interested in. I didn't start to think about that
opportunity until Intel dropped the other shoe. In
April, the company announced that it would
create a whole new business by building and
running huge Internet data centers around the
world. By this fall, Intel will have its first server
farm in operation. Inside, 2,000 Intel-based PC
servers will store data from customers such as
Excite (XCIT) and send it out to users on demand
through Excite's Web site.
And there's my new market. I think the
proliferation of communications devices
connected to a multitude of networks and
pipelines will create a massive new market for
databases of all sizes over the next few years. I
want to own shares in some of the companies that
make these databases, that build the hardware that
powers them, and that produce the software that
makes accessing them possible.
Let me quickly explain why I believe in this trend
and then suggest a few companies that should
profit from it.
Multiple devices mean more problems
In a world where I have just one device for
getting e-mail or tracking a portfolio or keeping
an appointment calendar, my networking and
database needs are pretty simple. I dial in through
my PC's modem and download my e-mail from a
central database and server, or I download stock
quotes over a modem from a server and database
at my broker.
Things get more complicated, though, when I
have multiple devices -- sometimes in multiple
locations -- that I can use to perform the same
task. I now want to read my e-mail from my PC
and my wireless phone. I want to make stock
trades from my PC and from my PalmPilot. And
I want one device to know what the other has
done; the portfolio I keep on my PC should
reflect the trade I made on my PalmPilot.
One solution is to let devices talk to each other.
My PalmPilot goes into a cradle that lets it
synchronize with the calendar I keep on my PC.
Take that a step further by linking up multiple
devices and you have a home network.
But I think that's only a partial solution. It limits
me to devices that can physically share the same
network and that can run compatible software.
And it requires me, the individual user, to
maintain the network and its various databases.
The proliferation of
communications
devices connected to a
multitude of networks
and pipelines will
create a massive new
market for databases
of all sizes over the
next few years.
Storage on a central server
There's another approach. Both America Online
(AOL) and Microsoft (MSFT), the parent of
MSN MoneyCentral, acquired Internet-based
calendar companies this spring. Both the acquired
companies -- When Inc. by America Online and
Jump Net by Microsoft -- keep a user's calendar
in a database at one of their server farms. Call up
your calendar on any device connected to the
Internet and, using a standard Web browser,
you'll see the same calendar. Make a change and
it gets stored on the central server for recall by the
same or any other digital device. America Online
or Microsoft performs all maintenance.
I don't know exactly what the future mix is likely
to be between devices connected to each other by
home networks and devices connected to each
other through the Internet, but I do suspect a mix.
Some devices in the home will need to share data
and tasks. A home TV and PC might share a
single Internet access connection through a
network that would also enable a user to store
photos on the PC and display them on the TV, or
transfer video from a VCR to a PC for editing.
My guess is that these networks will make a
respectable showing in homes that also double as
offices, where connecting a couple of PCs and a
printer and a scanner is necessary.
I do expect the big numbers to go to the Internet,
however. By putting most of the software and
most of the database on a central server, the
access devices themselves can get by with less
memory and less processing power. And that
means they can be cheaper. Right now, too, it's
much easier and cheaper to access the Internet
than to network a home. (Wireless local networks
will fix the ease of use problem, but at the
moment they're still too expensive.) The early
figures seem to confirm my opinion. Hotmail,
Microsoft's Internet browser-based e-mail
service, now has 40 million customers.
BancBoston Robertson Stephens estimates that
the market for hosting third-party Internet data
will grow by 70% annually over the next three
years, to $7.5 billion. Right now, the total market
is about $500 million.
Details
Company Report
1-yr Chart
Earnings Estimates
Company Report
1-yr Chart
Earnings Estimates
Details
Company Report
1-yr Chart
Earnings Estimates
So what about some potential winners?
I've got two candidates on the database end.
First, Oracle (ORCL). The company
dominates the market for large-scale
databases that run on Unix and Windows
NT, the two major operating systems
used in server farms. Oracle's newest
release of Oracle 8i has been rewritten
specifically for use with the Internet.
Second, Pervasive Software (PVSW).
Pervasive operates at the other end of the
database market, specializing in very
small databases for the new generation of
communication appliances and other
devices. (Even though these devices
might be downloading data over the
Web, they still need a limited ability to
store and manipulate it.) The company's
new product for the mobile devices and
embedded chips comes in three sizes --
an 8K engine for smart cards, a 100K
engine for data storage in devices such as
the PalmPilot and a 350K engine for use
in devices running Windows CE.
Exodus is the best pure play
Among companies making a big business out of
server farms, I'd certainly put IBM (IBM) at the
top of the list. The company's combination of
expertise in big computers and its great service
division make it a leader in this area. Intel is just
getting off the ground here and it will have to
learn the business from scratch. But the
company's deep pockets certainly will come in
handy considering that server farms can cost
$100 million a pop to set up. Exodus
Communications (EXDS) is the best pure play
here. The company has eight data centers up and
running with 10 more scheduled to go operational
by the end of 1999. The company's client list
includes Yahoo! (YHOO), Silicon Graphics
(SGI) and TeleBanc Financial (TBFC).
I think the big hardware winners are likely to be
Sun Microsystems (SUNW) and Intel. Sun and
Intel make the two chips that power most Internet
servers. (Sun, of course, also makes the complete
box and has the biggest share of the Unix server
market.) But both companies are also big players
in the Internet device business, an industry
projected to record unit growth of almost 100% a
year over the next five years. (For comparison,
the PC business is forecast to grow by 13% to
15% a year during that period.) Thanks to Java
and its simpler sibling Jini, Sun is also a major
player in software operating systems for Internet
devices. Intel's PC business overshadows all its
other chip-making activities, but the company is a
major producer of chips that go into appliances
and other embedded uses. And thanks to the
acquisition of the very low power StrongArm
chip as a result of a pre-Compaq deal with Digital
Equipment, Intel is likely to get stronger in this
area.
Jubak's Archives
Recent Jubak articles:
• Why AT&T goofed in the
long run, 5/18/99
• Is At Home the next
America Online?, 5/14/99
• The hidden winners of the
AT&T/MediaOne deal,
5/11/99
More…
The stock and company I like best here, however,
is Wind River Systems (WIND). The company
makes the leading real-time operating system for
appliances. Its customer list includes Qualcomm
(QCOM), Cisco Systems (CSCO),
Hewlett-Packard (HWP) and Intel. Thanks to a
warning in January about earnings for this
quarter, the stock is now cheap, trading at just 21
times earnings for the year that ends in January
2000. I'm adding the stock to Jubak's Picks with
a target price of $30 a share by May 2000.
The other stock from this group that is trading at
an attractive price right now is Oracle. In my next
column I'll take a look at that stock -- along with
SAP (SAP), PeopleSoft (PSFT) and others -- in
the context of the troubled enterprise resource
planning software industry.
Research Wizard
Earnings Estimates
Changes to Jubak's Picks
Buy Wind River Systems (WIND)
I'm adding Wind River Systems (WIND) to
Jubak's Picks with a target price of $30 a share
for May 2000. The company will be a major
beneficiary of the rapid growth in the market for
specialized appliances that can communicate over
the Internet. Wind River's real-time operating
system commands a major share of the market for
software to run the chips in these devices -- and I
think the company's product line and customer
lineup is strong enough to hold off challenges
from Sun Microsystems and Microsoft. The stock
took a beating after the company warned that it
would miss earnings estimates for the quarter that
ended in April, but I expect it to rebound strongly
in the rest of calendar 1999. (The company
announced its fiscal first-quarter earnings this
morning. You can listen to the conference call on
the Wind River Systems Web site.) |
