Fiber optic cost $70bn more than necessary
Author of new estimate says much will just rot
By Peter J. Howe, Globe Staff, 3/11/2002
Three years ago, while telecommunications mania abounded, Boston consultant Mark Bruneau sounded a wildly unpopular - but, in hindsight, prescient - warning that telecom and Internet companies blanketing the world with high-speed fiber-optic lines were creating a financially disastrous ''fiber glut.''
Now, with Global Crossing bankrupt and several other long-haul optical network operators reeling, Bruneau, the chief executive of Adventis Corp., has followed up with a new study that places one of the first hard numbers on how much money has been wasted on needless telecom network expansions: $70 billion.
By tallying combined 1997-2002 capital spending by Baby Bells, long-distance carriers, and newer entrants to the market and comparing it to what he calls a rational level of spending to meet actual end-user demand for long-haul network capacity, Bruneau came up with a gap of at least $70 billion between what was spent and what was warranted by growth in Internet, data, and phone traffic.
To put the $70 billion in context, the expenditure is four times what Verizon Communications, the country's biggest phone company, spends to upgrade its network each year. The figure is also twice the combined annual revenues of Lucent Technologies and Nortel Networks, North America's biggest manufacturers of telecom and optical gear.
''A lot of this stuff will just rot in the ground,'' Bruneau said. ''Even if there is a healthy growth in demand for bandwidth over the next decade, not disruptive or insane growth but neither anemic growth, you could conclude that [this fiber capacity] isn't going to get lit, isn't going to get used, and is going to be written off.''
Most analysts would now agree. But the notion of a ubiquitous bandwidth glut has come under sharp criticism. Analysts such as Oklahoma-based Telechoice argue that several city-to-city routes in the United States are close to maximizing use of existing fiber. Other studies project robust demand for metro optical systems connecting businesses to the abundant long-haul fiber and to remote data-storage centers.
''We do make a distinction between a fiber glut and a bandwidth glut,'' said Deborah Trevino, spokeswoman for Williams Communications. The Tulsa-based carrier warned last month that it may file for Chapter 11 protection, driving its shares to penny-stock status.
Like many companies, Williams installed huge volumes of dark fiber, as it built out a national network, figuring that it made economic sense to dig up streets and rights-of-way along pipelines and railroads only once and leaving extra glass fibers to be lit up in future years by installing the necessary networking electronics. Telecom engineers typically say that the real expense of network expansion is backhoe costs for digging up streets and rights of way; installing an extra 100 fiber strands when a trench is open is a tiny incremental cost that may make long-term sense.
But Bruneau said his $70 billion overspending estimate includes dark fiber that will probably stay dark forever because ''there just isn't going to be a foreseeable need for a lot of this long-haul bandwidth.'' By the time some of the unlit fiber is needed many years from now, it could be technologically obsolete and have to be replaced, he said.
Ron Vidal, senior vice president of Colorado-based Level 3 Communications, one of the financially stronger optical start-ups, said that late 1990s economics made the investment plausible at the time. ''There was large-scale availability of capital,'' he said. ''It was an environment where good ideas were getting funded at an unprecedented rate, but it didn't mean they were all going to be successful, and many weren't. That's the way markets work.''
Vidal said it is also important to note that ''enormous amounts of capital went in, but it doesn't mean that a lot of capacity came out.'' Besides the spending on dark fiber, vendors such as Cisco Systems and Lucent Technologies had to take back millions of dollars in optical networking gear that start-up companies bought in 1999 and 2000 but never deployed, because the telecom bubble burst.
Vidal said Level 3 believes that long-haul telecom ''is a game of `the last man standing wins,' so we're doing everything we can to make sure we're still standing and positioning ourselves for the eventual recovery.
''At some point, the capacity that's out there is going to get sucked up,'' he said, perhaps in five years. Level 3, which has focused on building ample empty conduits for future fiber installations, rather than installing the fiber immediately, has $1.5 billion in cash on hand and expects to generate positive cash flow this spring, Vidal said.
Bruneau's study of waste in telecom spending followed a mid-1999 report, in which he put himself out on a limb. He projected that capacity in 2001 would be 400 times the 1998 level, but that data and voice traffic might grow at best by a factor of 20. Telecom companies and investors pouring billions into adding ''more and more bandwidth into the backbone are collectively hallucinating,'' he said then. He now projects that capacity has outstripped demand by more like 3 to 1 than 20 to 1.
It took some time for the fundamental truth of Bruneau's report to be recognized. For example, Williams Communications stock came out around $30 a share in an initial public offering and soared past $60 by early 2000. Qwest was at $35 and headed to $60. Level 3 was trading in the $90s, on the way past $120, and WorldCom was solidly in the $60s.
On Friday, Williams closed at 13 cents a share. Level 3 was down to $4.20. Qwest had slid to $9.71, and WorldCom to $9.19. Vendors serving the optical networking market - including Lucent Technologies, Nortel Networks, and Sycamore Networks - have seen comparable share plunges.
And Global Crossing filed the fourth-biggest bankruptcy in US history earlier this winter, after losing $80 billion in shareholder wealth.
Peter J. Howe can be reached by e-mail at howe@globe.com <mailto:howe@globe.com>. |
