DGIV-THE NEXT BELL BUSTER.....>>>>>9/8/98
Bell Buster
By Toni Mack with Carleen Hawn
AT&T HAS BEEN IN BUSINESS for 113 years and has
$24 billion invested in telecommunications
equipment. Level 3 Communications has been
building its telecom network for all of one month
and has yet to complete a single phone call. When it
does so later this year, it will have something no
other major telco can offer: Every single bit of its
technology will be shining new.
It's one of the paradoxes of the digital age that
established companies can often find themselves at a
big disadvantage against upstarts. Which helps
explain why newly minted companies can take on
powerful players like AT&T, WorldCom and SBC
Corp. The battle is joined over circuit switching
versus packet switching, two methods of connecting
phone callers.
AT&T and all the big phone networks that followed
were built to connect a caller (or bank terminal or
PC) at one spot to another across town or across the
globe. With circuit switching the line is tied up until
someone disconnects. You can see the inefficiency:
It's like tying up one lane of a freeway for only one
motorist.
Packet switching chops the voices or data into small
bits, each with the recipient's address. The packets
then fly over whatever line is available and
reconnect at the other end. Think of it like an
airline counter where a single line of customers
feeds many ticket agent positions. If there's a
bottleneck at one position, or one is closed, the
packets swerve to an unused part of the network.
Today's titans get more than half their revenues
from phone calls hooked up the old AT&T way, via
circuit switching. Omaha-based Level 3 is building
the first nationwide network based solely on packet
switching. Packet switching is a method barely in its
toddlerhood, with bugs still to be worked out, but
even the old-guard companies acknowledge its
technological supremacy. So why don't they just
scrap circuit switching? Because they have too much
money and know-how invested in circuit switching
to just throw it on the scrap heap. It's the old
sunk-cost problem.
Level 3's chief executive, James Crowe, has nothing
invested in circuit switching. Nor does he carry any
mental baggage about telephones. This 49-year-old
engineer cut his teeth on electric power plants and
veered off into the phone business relatively late in
his career. He was working for the giant
Omaha-based construction firm Peter Kiewit Sons'
in 1988 when he spied an opportunity to wrest some
business away from local phone monopolies. Crowe
tackles the Bells with the same zest he applied to the
sport of rock climbing-scaling cliffs and sheer
mountainsides-in his youth. "Monopolies offend
me," he says. "They stifle innovation and lead
inevitably to waste. Introducing markets to
monopolies is a lot of fun."
Determined to exploit the opportunity presented by
the sunk-cost problem that confronts the established
telcos, Crowe has embarked on a grandiose plan to
lay 23,000 network miles of fiber-optic cable, with
tentacles into customers' offices on three continents.
His assets, as he describes them, are "$4 billion and
a blank sheet of paper."
To give you an idea of how ambitious this project
is, MCI has spent 15 years building its 25,000-mile
network. Crowe wants to replicate the feat within
five years.
A handful of outfits, like Qwest Communications
International, ICG Communications and IXC
Communications, are already doing parts of what
Crowe wants to do. But none are hitching their
fortunes so completely to this newfangled packet
technology, nor taking it to Crowe's scale. The old
behemoths like AT&T and its former Bell
subsidiaries are moving into packets, too, but it will
be a long time before their systems are completely
converted to it.
Crowe has a lot riding on a technology that is, after
all, not entirely tested. He clearly enjoys taking this
sort of risk. Crowe relishes the tale of explorer
Ernest Shackleton as told in the adventure classic
Endurance. After his ship was crushed by
Antarctic ice in 1915, Shackleton and two crewmen
reached an icy peak on an unexplored island. Fog
hid the glacier below. There were only two choices:
stay where they were and die or slide down an
unknown and invisible slope to what could be either
salvation or death. The trio slid 2,000 feet,
unharmed, and trekked to a remote whaling station.
Crowe: "They walk into this town, out of nowhere.
The whalers go, 'Where the hell did you come
from?'" What's that got to do with business? Says
Crowe: "The trick is to find out where that edge
is-how big a challenge can you take on without
falling over into the hole." One day Crowe expects
to march into Telephonesville and have the locals
wonder: "Where the hell did you come from?"
Crowe is betting on Moore's Law, which decrees an
exponential decline in the size and cost of a
transistor for the same amount of computing power.
The concept of packet switches goes back to AT&T
research in 1972. But it has come into its own only
recently, as the transistor (another AT&T invention)
got so cheap that it became practical to apply
massive amounts of data processing to a single
phone call or modem connection. It is now quite
economical to chop phone calls into snippets and
reassemble them; it was not economical to do that
26 years ago, when transistors cost 50,000 times
what they do now.
So the old telcos had no choice but to keep investing
in the old technology. Level 3, unborn, had no such
obligation. "We're watching," says Crowe, "a
change in the whole telecommunications
infrastructure that is on the scale of the shift from
mainframe computers to the PC." The same
dynamic is at work, he says, that enabled Microsoft,
Intel, Dell and the like to whip past seemingly
invincible giants like IBM and Digital Equipment.
Plenty of folks are following Crowe to the edge of
the ice. In June, with nary a strand of fiber yet laid,
Crowe went to Wall Street for $1.5 billion-and got
$2 billion, at the time the biggest single junk-debt
issue ever. Level 3 sports a stock market value of
$12 billion-even though the company will be
eating cash for capital spending well into the early
2000s.
Investors are betting the jockey as much as the
horse. Crowe calculates risks carefully before he
takes them. He has a reserved, almost military
bearing. Son of Henry P. Crowe, a much-decorated
World War II marine, the younger Crowe studied
mechanical engineering at Troy, N.Y.'s Rensselaer
Polytechnic Institute. He later earned a Pepperdine
University M.B.A.
Crowe was building electric power plants for
Morrison Knudsen when, denied a promotion in
1986, he left to work for a privately held
competitor, Omaha-based Peter Kiewit Sons'. He
was eager to pool his talents with those of Kiewit's
shrewd chairman, Walter Scott, an old pal of
Warren Buffett, whose modest offices share a rather
nondescript Omaha office building with Kiewit's,
though there is no corporate connection between the
two outfits.
Electricity deregulation was just beginning when
Crowe arrived at Kiewit. He sold Scott on investing
in some plants. They found a better way to play
deregulation later in telecom, and began building
one of the first networks to compete with the
regional Bells, MFS Communications. That was
1988.
After investing a bit under $500 million in the
network, Scott took MFS public in 1993 and
distributed the remaining shares to Kiewit
stockholders, with Crowe in charge, in 1995. Crowe
proceeded to buy Internet backbone Uunet
Technologies and sold the whole package to
WorldCom at the end of 1996 for an astounding
$14.3 billion. Scott and Crowe together pocketed a
total $636 million in WorldCom stock and options.
WorldCom founder and boss Bernard Ebbers gave
Crowe only a small sliver of the combined business
to run. So Crowe rejoined Scott at Kiewit in 1997
and began laying plans for Level 3. In March they
distributed its shares to Kiewit shareholders and
listed them on Nasdaq.
Crowe's all-business demeanor cloaks a wry humor.
Asked why he went to Rensselaer, he quips:
"Because the drinking age in New York State was
l8." But he displayed his other side at Level 3's July
annual meeting in Omaha. An attendee in Level 3
baseball cap and untucked shirt stood to enquire:
"Are you planning to use non-zero dispersion
shifted optical fiber with erbium-doped fiber
amplifiers at 1,550 nanometers?" Crowe showed his
grasp of technology: "Yes-except it will be
enhanced non-zero dispersion shifted fiber." The
crowd applauded.
At Level 3 Crowe is building what he terms a
"continuously upgradable network." This means that
future expansion will be relatively cheap and easy.
Thus in the early stages Level 3 will fill only one
pair of conduits-garden-hose-like fiber
containers-with that enhanced fiber. (The second
of each pair is for reliability, in case the first gets
cut or clogged.) But simply digging the trench to
lay cable is 70% of the installation cost. So while
Level 3's construction workers are down there, they
will lay an extra five to seven pairs of conduits
ready for future stuffing. They will do the same as
they snake conduits through office buildings.
The expansion potential is awesome. The fiber in
just one pair of conduits will be able to transmit 60
trillion bits of information per second-the
equivalent of 750 million simultaneous phone calls.
That's 15 times as much capacity as now serves the
best-wired city in the world, New York. That's
without installing additional fiber in those conduits.
Will the world ever want that much capacity?
Crowe draws on the PC analogy: "The folks who
invent ways to use computing power come up with
ways to use it faster than Intel et al. come up with
faster processors. The same will happen here.
Economics 101 tells you that if you push down the
unit costs of communications every year, demand
will go through the roof."
The roof-shattering increase in demand has only just
started. Innovators like America Online, Netscape
Communications and Amazon.com are the bare
beginnings. Merrill Lynch Vice Chairman John L.
Steffens recently described plans to give clients
24-hour-a-day access to Merrill's research reports
on-line. Others are already working on applications
as seemingly far-fetched as sending a scent attached
to an E-mail. (See sidebar) Doing these things will
require an almost unimaginable increase in
communications capacity.
Plans are plans. Can Crowe execute? So far he looks
darned good. First, he lured former MFS executives
from wealthy retirement or other telecom jobs to
fill 20 of Level 3's 21 top positions. They get fat
stock option deals, but the options don't kick in
unless Level 3 shares outperform the S&P 500
index. Crowe himself has invested $55 million in
Level 3 stock; other top staffers invested their own
cash. "I'm a big believer that if you're going to run
something you ought to own a piece of it," Crowe
says. All those conditional options do motivate
people. Reports Level 3 Chief Operating Officer
Kevin O'Hara, 37: "I got married in July and took
half a day off." "There's electricity in the air here,"
says Senior Vice President Joseph Howell.
Crowe has sewn up rights-of-way along 15,000
miles of U.S. railroad tracks along which to lay his
network and has taken interests in one transpacific
and two transatlantic lines. Craig McCaw is buying
in: In July McCaw and two outfits he has stakes in,
Nextel and Nextlink phone services, agreed to pay
$700 million for 25% of the capacity on Level 3's
long distance U.S. lines. That's about one-third of
the $2.2 billion cost of those lines, for one-fourth of
the capacity. Good math.
A key chunk of Crowe's strategy fell into place with
the $165 million acquisition in April of little
software startup XCOM Technologies. This deal
helps Crowe deal with the Achilles' heel of
packet-switched networks. Right now,
packet-switched voice calls offered through the likes
of Qwest and IXC are cumbersome and slow. The
caller dials up a local computer, often has to punch
in an I.D. number, and then punches in the number
he wants to reach-and waits while the information
is transmitted and decoded at the other end. That's
okay for penny-pinching technotypes who will trade
the delay and low quality for savings. Not for the
business customers Crowe is courting.
While the future may belong to fancy stuff like
transmitting movies and smells over fiber, the
present is still largely voice traffic. A full 81% of
telecom's $540 billion worldwide annual revenue
comes from plain old voice calls. Crowe can't build
a business without soaking up some of that
voice-call revenue.
XCOM helps here. Its developers (see sidebar) have
concocted a so-called soft switch. Callers can dial as
they ordinarily do, with instantaneous hookup,
because XCOM's software translates the information
in a form that the local networks recognize as a
circuit-switched call.
Crowe plans to have fiber lines laid between 25 U.S.
cities by the first quarter of 2001, with the rest of
the network completed by 2003.
Of course AT&T, MCI, WorldCom, the regional
Bells and others are not asleep at their switches. All
have some packet-switching capacity, and they're
expanding it-as with AT&T's $11 billion purchase
of Teleport and Sprint's $2 billion investment in
packet-switched equipment. With $11 billion in
annual cash flow, AT&T does not lack the
wherewithal. The sunk cost problem again.
But older firms face a hurdle Level 3 doesn't.
Edmund Mullane, who follows telecom for New
Japan Securities, explains: "They have shareholders
to listen to and dividends to pay. None can afford to
go through two years of net losses, which would be
required to upgrade their networks to [the new]
technology. Their shareholders would revolt." The
same sunk cost predicament: You guys invested in
those copper wires, so you better keep getting
revenue from them.
A Level 3 deal will make the most sense to a
business that can evade existing carriers at each end
of a connection-say, on a videoconference between
its headquarters in Chicago and its branch office in
Atlanta. But a lot of traffic will necessarily go out
from Level 3 to the public-switched network-as
when a Merrill Lynch broker calls a customer who
gets her phone service from SBC. Will the other
phone companies cooperate?
"The incumbents' first step is always to refuse to
connect," says Level 3 Chief Operating Officer
O'Hara. WorldCom last year announced that others
would have to meet its demanding specifications to
hook up to its Internet lines. Level 3 hasn't yet
qualified.
Many an otherwise great idea has foundered on bad
marketing. Once the lines are in place, can Level 3
market its services? "Crowe is on target with the
technology," says Jeffrey Kagan of Atlanta-based
Kagan Telecom Associates, "but his major
competitors already have a huge customer base that
they can migrate to their own new networks as they
upgrade. Level 3 has to figure out how to convince
people that its network is worthwhile."
Crowe is hard at work on the problem. While Level
3's network is under construction, he's leasing space
on Frontier Corp.'s long-haul fiber lines; that way
he can start serving customers even before his own
lines are in place. The biggest target: financial
institutions like Fidelity, Citicorp and Goldman,
Sachs, with fast voice and data volumes.
Crowe predicts that he can undercut AT&T and local
phone companies 15% to 20% in delivering local
and long distance service to office buildings. If a
business of 80,000 employees runs up an $8 million
monthly phone and data transport bill, Level 3 will
do the work for something like $6.5 million.
For now at least, Level 3 is interested only in
business customers voracious for data. It does own
48% of RCN Corp., which is slowly building cable
and fiber lines to residential customers in the
Northeast. But this isn't Crowe's main thrust. He's
in a hurry and can't sit around waiting for
regulators to force open the consumer market.
"Someday Congress or the courts will tear down
those barriers," Crowe says. "In the long run, I
can't believe they'll deny residential and
small-business customers the savings the big
customers can get." But he's not holding his
breath-and in the meantime there are enough big
and fast-growing customers out there to keep Level
3 and its competitors busy for years.
If someone like Bernie Ebbers at WorldCom or
Michael Armstrong at AT&T has the courage to tell
shareholders that much of their existing physical
assets are outdated and should be written off and
replaced, Level 3 might have trouble competing
with companies that have huge customer bases.
But Crowe is hewing to Walter Scott's favorite
maxim: "Protect the downside, and the upside will
take care of itself." If he can't pull off the whole
business plan, he'll still have a valuable network that
another, bigger company will want in a world on
the verge of a tremendous explosion in the demand
for communications capacity. "At worst, we'll
become part of someone else's strategy," Crowe
says. "The industry spent 100 years building a
network which, in effect, extends our ears around
the world at a reasonable cost. We're now at the
Model T stage of building a network that can extend
your eyes around the world at a reasonable price."
One way or another, Level 3 is almost guaranteed a
nice chunk of that future.
|
