Has the Net Stopped Growing?

But the question now is, how badly has the industry overshot reality? More specifically, how much more capacity is in networks than consumers will ever need?

A number of telecommunications analysts, perhaps looking to prove just how meticulously they cover the industry, have taken to counting the number of fibers in every optical network that every telecommunications company has laid. Getting even more anal retentive, many have taken to counting how many of these fibers are "lit" - hooked up to equipment and carrying Internet traffic. What they have found is that only a tiny fraction of fibers are lit - as little as 3 percent by some estimates. That seems to indicate a shocking overcapacity. In fact, according to telecommunications research firm Probe Research, perhaps only 14 percent of the fiber laid across the Atlantic may ever be needed. In some markets, the problem is even worse.

But carriers complain that counting fibers oversimplifies the problem. For every dollar spent digging a trench and laying fiber, it will cost several more dollars to pay for the equipment and labor necessary to actually light and run a network. And since the financial markets have largely stopped lending money to telecommunications companies, it's difficult to light all that dark fiber.

"The game is not to see how much fiber you can put in the ground, the game is how well you actually run a network," says Qwest COO Afshen Mohebbi. "I don't care if someone has 99 conduits for fiber in the ground. Maybe you could lease them to the sewer department, but until you have customers paying you to carry traffic through the network, it doesn't mean a thing."

Level 3 estimates it would take $500 billion to light all the fiber that companies have laid. What that means is that while there is little opportunity for anyone to break into the business as an Internet provider for years to come, there will eventually be a shortage unless those networks can be lit.

Also, while carriers have too much fiber on long-haul routes that cross countries and traverse oceans, there is not enough in local networks, the systems that service metropolitan areas. Until that imbalance is solved, consumers will not see much benefit from the existing fiber.

The problem for the carriers (and subsequently, the manufacturers) is that despite years of hype and promise, no one has introduced an Internet service or application that has dramatically increased the need for bandwidth. The closest thing the industry has had is the music-swapping service Napster, which is now dying. According to telecommunications consulting firm Adventis, at the height of Napster's popularity the service was using about 5 percent of the network capacity available in the United States. That percentage has never been approached by any other service.