An Optical Delusion?

By Karl Taro/Greenfeld

By now the booms and busts of the Internet economy have taken on an almost seasonal aspect. Callow start-ups--initially apples in venture capitalists' eyes--become in a few short months rotten e-commerce or business-to-business fruit. Portals, community sites, e-commerce companies, business-to-business (B2B) verticals--all are Web sectors once hyped as the next big thing; all are Web sectors in which stock prices have gone off the cliff.

Perhaps tired of trying to pick sectors that will be Web winners, some investors and venture capitalists think they have found a way to bet the whole shebang--optical networking. That's the business of sending digital information via light waves rather than electronic signals, and it has emerged as the latest, greatest Internet investing trend. Which means that companies involved in any phase of optical networking have become hot stocks. Component makers, such as JDS Uniphase and Corning, and system designers, such as Juniper and Ciena, are each up more than 50% this year. Last Friday, despite a vicious market sell-off, Corvis, an optical-equipment maker, launched its IPO at $36; it closed at $85. Corvis has no revenue and three corporate customers. High-end router-and-switcher company Avici launched at $31; it hit $97.

The reasoning goes about like this: think of some of this year's biggest business stories--AOL's merging with Time Warner, Vivendi's acquiring Seagram, Napster's hijacking the music industry. They are all, in some sense, about devising more and creative ways to suck up bandwidth. Words, music, video, interactive TV--they're all data that have to go through the electronic plumbing of the network.

"It's the most fundamental level of demand creation," says Vinod Khosla, a partner at venture-capital firm Kleiner, Perkins. "Optical-networking companies are like Levi's. They're supplying jeans and tools to miners during the Gold Rush." The amount of data traversing the Web is doubling every three months, and as these merged-media entities offer fatter and better Web services, bandwidth demand should accelerate again.

What optical-networking companies do is provide the equipment--filters, amplifiers, converters--and systems to companies that are feverishly building out the Internet, such as Level 3, AT&T and Qwest. Although optical fibers that transmit light waves have been around since the 1970s, only in the past few years have companies like JDS Uniphase and Corning figured out how to send prodigious amounts of information through those fibers by dividing light waves into channels and then packing data into each channel. A single channel is like a light bulb going on and off 10 billion times a second, flashing the 0's and 1's of binary computer code down the fiber.

The technology is as tricky as it sounds, forcing component makers like JDS Uniphase and Corning, among others, to acquire those pieces of the tech puzzle they can't assemble in-house, using their richly valued stock as currency. Two weeks ago, industry leader JDS Uniphase bought fellow component maker SDL for $41 billion in stock. And last week Corning began talks to acquire Nortel's optical-components business for a staggering $100 billion. Negotiations broke off late in the week, in part because Nortel's optical assets were rising in value so rapidly.

The buzz and heat in this sector are reminiscent of last spring in the e-commerce space or last winter in the B2B space, when hundreds of start-ups were launched and dozens managed to go public before NASDAQ tanked. But the venture capitalists swear that this time it's different. For one thing, these are real companies making high(est) technology gear for blue-chip customers. More important, some companies have been showing impressive growth and real profits. JDS Uniphase last week announced quarterly net income of $114 million; sales increased 173%, to $524 million. Corning earned $525 million last year on sales of $4.7 billion. (Optical components are 70% of its sales.) Unlike e-commerce or B2B start-ups, where forecast multitrillion-dollar markets have turned out to be largely theoretical, optical networking, according to industry consultant RHK, is a $6.6 billion annual business right now and is growing 75% annually. "There is an urgency because everyone feels as if they are going to miss out on the best optics deal," says Darius Sankey, a partner at Zone Ventures.

With nearly $4 billion in venture capital being invested in networking start-ups this quarter, the concern is that too many companies are being hurled too quickly into the same markets. For instance, at least a dozen optical-switch makers have been launched in the past six months. (Remember the dozen pet-supply e-commerce companies that were started last year?) That doesn't faze Josef Straus, CEO of JDS Uniphase: "Last Supercom [an industry trade show], there were 40 new start-ups in networking. Why did they get funded? Because they had a good business proposition in a great space. To the horizon I can see--four or five years--it's smooth sailing."

The companies point to what they call the "metro" business as a huge untapped market. After the national optical-fiber backbone is in place, the task will be to wire the cities and local networks, and eventually get that fast fiber to your neighborhood, and finally replace the copper wires and coaxial cable that go into your home. The trouble is, the metro market doesn't really exist today--and the technology that will make it possible, if not necessarily profitable, is only now being invented.

Moreover, bandwidth demand may not rise fast enough to meet the rapidly growing supply. According to Corning's manager for optical switches, David Charlton, in a few years a single strand of fiber will be able to handle all the voice traffic in the U.S. Sure, all sorts of new services will be available over the Web, and wireless appliances will be coming online that have to funnel through a land-based optical-fiber network at some point. But what if all this happens in five to seven years instead of two to three? Can somebody say B2B shakeout all over again? Tom Nolle, president of the consulting firm CIMI, warns that there is too much bandwidth capacity already and that once the current wave of network building is finished, there could be bandwidth glut at the backbone level.

If so, that could leave the market for fiber and its components saturated--and some companies' balance sheets stagnating. According to CIMI's calculations, only 13% of the fiber that's in the ground right now is even lit--turned on and transmitting data. The rest has been laid in anticipation of a bandwidth-demand explosion. With that much elasticity in the system--and technology making existing fiber even more efficient--how much more fiber and components do we need? "There will be snags and lulls," admits Corning president and coo John W. Loose. "But as better optical solutions are found, costs go down, and as costs go down, people find new applications that hog more bandwidth, and that drives demand back up."

That sort of virtuous cycle has worked in the microprocessor and computer-memory businesses--where applications quickly used up the available speed and space--but it has yet to be proved in the bandwidth sector. There's no doubt that optical networking is the business of the future. The problem is, if bandwidth-demand growth slows, it always will be.