More Bandwidth Than We Need?
By Cintra Scott, Smart Money.com
March 14, 2001
JUST A FEW months ago, the idea that our appetite for fast Internet connections could be satisfied anytime soon was almost unthinkable. Then again, the idea that Cisco Systems (CSCO), which sells gear to support more than half the world's Internet traffic, could miss earnings expectations and lay off 12% to 17% of its staff was equally unthinkable.
But we're getting used to thinking the unthinkable. While the conventional wisdom had it that the demand for bandwidth — the "pipes" that carry voice, data and Internet traffic among homes and businesses — was exploding, investors have lately begun worrying that there may already be more bandwidth on the market than we'll need for some time to come.
That's a scary thought. It's one thing to say that the stars of Internet infrastructure — the Ciscos, Juniper Networks (JNPR) and Cienas (CIEN), for instance — are facing a bit of a slowdown because the economy's cooling off. After all, a few more Fed interest rate cuts are likely to get the economy perking again sometime in the second half, and these companies would then stop handing out earnings warnings like after-dinner mints. But it's quite another thing to say that there's a fundamental oversupply of their wares, since who can tell when such excesses will be absorbed? Such fears help explain the brutal takedowns of the networking and telecom sectors in recent months — and, with them, the hair-raising decline of the Nasdaq. To the extent that the stocks of the networkers and their ilk were priced with any rationality, they were priced for assumptions of torrid long-term growth.
Susan Kalla of BlueStone Capital Equities has been sounding the alarm about those assumptions since last December. Kalla argues that the current glut of broadband network capacity could take "multiple years" to work itself out. "We believe incremental new supply of fiber-optic capacity has grown at least twice as fast as demand over the past three years," she then wrote on March 5. The result: Kalla thinks networking-equipment makers could see orders drop 50% year over year. Her Underperform ratings on Ciena, Juniper and Sycamore Networks (SCMR) are still firmly in place. Kalla's views gained prominence after she was featured in the March 12 issue of Barron's (which is published by Dow Jones (DJ), a co-owner of SmartMoney).
Kalla reached her conclusions by analyzing companies from their customers' points of view. For her Dec. 12 research note, Kalla surveyed bandwidth brokers — resellers of excess network capacity — and concluded that their customers, fiber-optic carriers like Level 3 Communications (LVLT), Qwest Communications (Q), Broadwing (BRW) and Williams Communications (WCG), had "put the cart before the horse" by ordering up more network capacity than they could sell to customers. As a result, carriers' prices dropped 30% to 40% in 2000, she estimated. And she thinks they'll drop just as much in 2001 and again in 2002, she added in a March 5 note.
True, the brokers Kalla surveyed said that high-bandwidth applications — like music downloads and video-on-demand — could eat up excess capacity in relatively short order. The trouble is, there's still a bottleneck in local access — that is, the connections between a long-haul network (stretching from, say, New York to London) and the consumers who want to use it (in, say, downtown Manhattan or Notting Hill).
"Although estimates on when the local access pipes could be fattened up vary, consensus opinions were about two to four years," Kalla wrote on Dec. 12, after speaking to the bandwidth brokers. In the meantime, she believes, "new applications are blocked since local passages needed to deliver them to the end users are too narrow." Fattening those passages is stalled by "prohibitively expensive" DSL and cable-modem technologies. For instance, DSL lines cost $700 to $800 a line to install, but bring in just $30 a month in revenues on average, making for a multiyear payback even assuming customers don't drop their subscriptions first.
An opposing view came from (surprise!) Cisco Chief Executive John Chambers on Tuesday morning during a three-day telecom conference sponsored by Merrill Lynch. To a rapt audience looking for some insight into the struggling telecom sector, Chambers reiterated his belief that we've only seen the beginning of bandwidth demand.
The funny thing is, while Chambers may sound hopelessly Pollyannaish in a month of stock-market carnage, there's still reason to believe he's right. Wall Street Journal All-Star Kevin Slocum of Wit SoundView argues that an overhang of excess capacity is a necessity, not a liability, in any market for leading-edge technology. His view, in a nutshell: If you build it, they will come.
Consider, Slocum says, the example of Intel (INTC) and the PC revolution. In the 1980s and early 1990s, every time Intel introduced a new, higher-capacity microprocessor — from the 286 to the 386 to the 486 to the Pentium — skeptics said that users couldn't possibly need all their newfound computing power. But as processor power increased, programmers wrote increasingly complicated software that required more processor power.
Bandwidth, Slocum thinks, is in a similar developmental stage. Bandwidth capacity should grow a bit faster than existing traffic in order to allow new uses for it to develop. "There are probably some pretty amazing things companies could do with 10-gigabyte wavelengths," he says.
Sure, Slocum says, there's something of a constriction in the local-access pipes. But he thinks the economics of that business are changing. Consumer demand for high-speed access is strong enough that, his industry contacts say, service providers are about to begin raising their prices for DSL lines. That would shorten the payback time for building out those lines, a prospect that should reopen the capital-markets spigot for local carriers and speed the "fattening" of those pipes.
True, he says, that could take a while. But he doesn't believe that the industry's sudden contraction reflects some vast excess of capacity. He figures that 80% of the sudden slowdown in spending by Cisco's customers, for example, is related to the cooling economy. As interest rates come down, he says, Cisco's growth could snap back every bit as quickly as it shrank.
But even if he's right, the near-term outlook for these stocks is undeniably fuzzy. With the bear market has come a "changing psychology, a slowdown psychology" that has made investors fear the worst about the future, Slocum says. But that, too, is bound to be temporary. If you buy Slocum's argument, the comparison to Intel is once again telling. Sure, an investor could have worried about chip demand waning with the Pentium processor's introduction in 1993, and bailed out of the stock. But that investor would have missed an incredible seven-year run of revenue and profit growth, during which Intel stock appreciated more than 1,000%. |
