Of all of the myths that drove the 1990s technology boom — dot-coms made good investments, the New Economy would never experience a recession, small telecom companies could beat the mighty Bells — the most damaging may have been the fallacy that Internet traffic was doubling every three months.
The belief that Internet traffic could grow so quickly — if true, it would have meant annual growth of more than 1,000% — led more than a dozen companies to build expensive networks as they rushed to claim a piece of the next gold rush. The statistic sprouted up in reports by industry analysts, journalists and even government agencies, which repeated it as if it were the gospel truth. “Internet traffic,” the Commerce Department said in a 1998 report, “doubles every 100 days.”
Except that it didn’t. Analysts now believe that Internet traffic actually grew at closer to 100% a year, a solid growth rate by most standards but one that was not nearly fast enough to use all of the millions of miles of fiber-optic lines that were buried beneath streets and oceans in the late-1990s frenzy. Nationwide, only 2.7% of the installed fiber is actually being used, according to Telegeography Inc. Much of the remaining fiber — called “dark fiber” in industry parlance — may remain dormant forever.
It is quite amazing in retrospect how much money was invested (and has been lost) due to such mistakes predictions. As I was reading some of the older magazines for the latest Tech Talk series, it is now evident that many among us were taken in by such optimistic views of the future. The more outrageous the prediction, the more realistic it seemed!
Another story in the Journal quotes a rueful Dr. Brinkman, a distinguished researcher who spent 35 years at Bell Laboratories, as saying, “Maybe we should have been less smart.” The story adds:
Never before has the efficiency of an industry’s technology gotten so far ahead of demand, creating a glut of capacity that will take years to work off — and crippling dozens of companies in the process.
Scientists perfected once-exotic methods for cheaply sending vast amounts of voice and data, such as Internet traffic, over fiber-optic lines. These advances far exceeded the pace of telephone-industry innovation in the 100 years before it. Prior to 1995, telecom carriers could send the equivalent of 25,000 one-page e-mails per second over one fiber-optic line. Today, they can send 25 million such e-mails over the same fiber strand, a 1,000-fold increase. Yet the cost of making that upgrade rose by just a few times over the 1995 price, and in some instances actually declined.
A similar thing may be happening in the compuer indsutry where microprocessor speeds are moving ahead far rapidly than what we can use. They challenge here is to expand the market (target the underserved people in the emerging markets) rather than overserving the same existing set of users, who have little incentive to upgrade.