The Upstarts Are Rocking Telecom
pathfinder.com
The Internet's getting a makeover, from copper wires to fiber-optic cable. The companies that make the machines that zap the data are the ones that stand to gain most.
By Eric Nee
Back in September, amid great fanfare, Qwest Communications finished laying the last of its 18,500-mile fiber-optic network crisscrossing the continental U.S. Five years ago, when Qwest started the thing, the network was seen as an audacious, spectacular gamble on then-unimagined demand for high-speed communications. But the Internet took off faster than anyone had dreamed. Demand for bandwidth, as in "I want a T3 line for my office next week, not next month," accelerated beyond all expectations. As a result, the bold network that Qwest completed four months ago already seems a bit, well, ho-hum.
You see, Qwest's network doesn't deliver what everyone now wants--superfast Net access for anyone, anywhere, at any time. So now the company is building a second network to offer just that. Like Qwest's first effort, this network will carry all kinds of traffic--phone calls, faxes, data transmissions, and video. And like the first, its cross-country trunk lines will be high-capacity optical fibers. But Q2 (as Qwest insiders call the new network) will be bigger, better, faster: To deliver bandwidth when and where customers want it, Q2 will be broadband end to end. Unlike the first network (let's call it Q1), which relies on slower cables to move data from those trunks to customers' offices, Q2 will deploy fiber everywhere, even running it into individual office buildings. Since Q2 features the kind of speed that customers are asking for right now, Qwest won't take five years to build it: The first pieces should be in place by summer.
While Qwest's stock is doing well (up 490% since the company went public in 1997), it's too early to tell whether Qwest's great gambles will ultimately pay off. But there's no doubt that one group will benefit from all this rush to build: optical equipment providers. They make the gear that routes traffic through the optical fibers of Qwest's network and pushes the signals along to their final destinations. To build Q2, the company will spend hundreds of millions of dollars on the most advanced optical equipment. Much of that money will go to established players like Nortel and Lucent. But a surprising share will go to startups that didn't even exist when Q1 was conceived--companies like Corvis, Siara, Cerent, Sycamore, Optical Networks, and Alidian. "They have a time-to-market advantage over traditional suppliers," explains Vab Goel, vice president of Qwest's newly created Emerging Technologies division.
These days, speed is everything. That is why not just Qwest but also AT&T, MCI WorldCom, and Williams are turning to these newcomers for the latest optical equipment. In the past, network carriers were conservative buyers that made purchases from a short list of large equipment suppliers--Lucent, Alcatel, Siemens, NEC, Nortel. But now demand for the high speed of optical networks is so great that carriers can't wait. Their impatience has fueled an extraordinary run-up in the value of optical-equipment suppliers. One example: Tiny Sycamore Networks, a Chelmsford, Mass., optical-switch company with just $31 million in revenues during its lifetime, boasts a market value of $25 billion. That's nearly as much as the $32 billion value the market has placed on Qwest. Another example: In August, Cisco paid $7.4 billion for two tiny optical startups, Monterey Networks and Cerent. That's an astounding figure, fully one-third of the total amount Cisco has paid for the 48 companies it has purchased over the past seven years.
The basic fact driving the shift to optical equipment is simple: A network that uses photons to deliver data is faster than one that relies on electrons. Photons shot down a glass fiber by a laser travel at much higher speeds than electrons sent down a copper wire. Also, signals can be packed more tightly on an optical fiber than they can on a copper wire. Result: A strand of fiber can carry 320 gigabits of data per second, enough bandwidth to support 5.7 million PCs with 56K modems simultaneously hooked to the Web. And many strands of fiber can be packed into a single cable. Data delivery via photons isn't perfect--yet. With current technology, an optical signal can travel up to 370 miles before its clarity begins to degrade. At that point, the signal needs to go through a device called a regenerator, which turns the signal from photons to electrons, cleans it up, and converts it back to photons before zapping it on for another 370 miles. That may seem a great distance, but photons carrying data from Los Angeles to New York must be regenerated at least seven times. At each regeneration point, a carrier must erect a building to house the regenerators. Then there's the electricity to run the regenerators, and personnel to maintain them and the building--it all adds up.
Two startups have made their mark by developing technology to attack this costly weakness in fiber-optic networks. Qtera and Corvis have designed amplifiers that shoot photons up to 2,500 miles before they need regenerating. The breakthrough doesn't mean carriers will rip the existing amplifiers and regenerators out of their networks; it simply means that they are unlikely to buy many new ones. Nortel, a leading supplier of amplifiers and regenerators, whose optical business accounts for some $5 billion a year, was developing a similar product. But company execs saw that Qtera's was going to be ready first. So, says Anil Khatod, vice president of Nortel's Optical Networks division, "time to market being as critical as it is, we bought the company." Nortel paid $3.2 billion for Qtera, whose product should ship this summer--under the Nortel brand.
While the new boxes from Qtera and Corvis will cut costs and speed up the network backbone, they won't do anything for another congested part of the network, the metropolitan market. In the typical U.S. city, a disparate collection of switches, cables, wires, and boxes sits between homes or businesses and the network backbone. Data that have traveled over the backbone slow down in this bewildering mesh, like a Formula One racer who suddenly finds that his new path is a clutch of dirt roads.
Cutting through the metropolitan mess is the technology task several optical startups have tackled--including Siara, Optical Networks, and Alidian. Today carriers often need separate boxes to handle voice traffic, Internet data traffic over copper wires, and traffic delivered via fiber optics. The engineers at these startups decided to fix that by building one box to handle all the different traffic. (Even though optical is hurrying to the forefront of communications, data and voice will travel via wires for years to come.)
Each startup has its own way of handling different data in a single box. But by looking closely at a single company--20-month-old Siara--you get a sense of the strengths necessary for a firm to stand out in optical networking. Siara is designing six custom chips, each to handle tasks that now require several chips or even several machines; one handles optical-networking jobs that once required a separate so-called Sonet box, while others handle high-speed solutions such as ATM and frame relay. Siara's chips go into a single box, replacing perhaps ten boxes that were previously required. Besides consolidating the existing traffic technologies, Siara's boxes also feature cutting-edge optical technology--so carriers can upgrade their system while bringing on the best new stuff as well.
As they race to build breakthrough products, optical networking companies attract some of the hottest engineers. Siara CEO Vivek Ragavan, for example, got his project rolling by hiring the principal engineers from the team that designed Advanced Micro Devices' powerful K6 microprocessor. Luring such stars from an established outfit like AMD takes stock options and money--Siara has it in the form of $30 million in backing from the likes of Kleiner Perkins. VCs can't get enough of optical-networking-equipment makers these days, and their support, in the form of both money and the relationships they foster among startups (and between startups and established companies) can be crucial.
Siara is as "pre-" as you can get--pre-product, pre-IPO, pre-revenues. Nevertheless, its technology is so promising and its engineers are so skilled that Redback Networks, a three-year-old company whose stock has risen 1,443% since it went public last May, agreed in November to buy Siara for $4.3 billion. That's less than Siara might have commanded had it gone public. But company execs could not afford to wait. "Joining with Redback accelerates our sales cycle by one year," says Ragavan. Remember: Speed is everything in this market. Redback came to the bargaining table with 120 sales and support people and 120 customers. Siara had few of either. Ragavan adds: "Together we'll build an independent company that can take on the established players."
With Siara, Redback just might do that. Certainly it will play a role in supplying Qwest's new network. The Q2 network will be a blend of products from suppliers like Lucent, Nortel, Cisco, and Juniper, and newcomers like Corvis, Qtera, Redback, and Sycamore. "We are not going to be shy about teaming up with these emerging companies," says Qwest's Goel. His job is to cultivate close relationships with promising startups so that Qwest can be the first to deploy new technology. Qwest engineers help the startups design their products, test them in Qwest's labs, and deploy them on the network. In some partnerships, such as with Qtera and Corvis, Qwest takes warrants allowing it to buy stock in the startup. "You have to spend a lot of time coaching and guiding [the startups]," admits Goel. "But that's not a drawback. It's been a very successful venture for Qwest."
Of course, there's nothing new about startups' making fabulous new equipment for bigger companies like Qwest. But in the past, as soon as such outfits burst onto the scene with some cool technology, Lucent, Nortel, or, most often, Cisco bought them up. That still happens, as with Cisco's recent purchases of Cerent, whose optical technology is designed for the metropolitan market, and Monterey Networks, which had developed optical technology for the backbone network. For Cisco, acquiring was the quick route to success in the optical market, so it swallowed hard and paid the price. (Nortel, too, is buying in to supplement its line of optical products. Lucent, by contrast, is looking to develop its optical systems in-house.)
But the small fortune Cisco paid for Monterey and Cerent points to a significant change in the industry. In many cases today, startups are able to gain enough size and capital to make a go of it as stand-alone companies. Sycamore, for example, has a market cap of $25 billion and a $400 million four-year contract with a major carrier, Williams Communications. That's an awful lot of company for anyone to try to acquire.
And now private companies are being launched with the kind of venture backing that helps a startup develop into an established player. Corvis just raised more than $200 million from venture investors--even though the private company has yet to ship its first product. With that kind of money, Corvis can rapidly build the kind of sales and support infrastructure, manufacturing capabilities, and engineering prowess of established companies.
Increasingly, startup founders who want to retain at least partial control of their company's destiny have yet another option: merging with another startup. Redback and Siara, for example, have a combined market cap of $13.3 billion. Their execs hope that they've now got a perfect blend: the nimbleness of a small firm and the heft of a company that has the sales and support staff to sell new products quickly in this burgeoning market.
The optical revolution has only begun. Switching from copper wire networks to optical ones will take years, and Qwest and the other carriers will most likely accelerate their purchases of optical equipment as they extend their fibers to every home and business. It should be an extraordinarily profitable time for the equipment makers that catch the light wave.
The Godfather of Broadband
By Eric Nee
As a partner at Kleiner Perkins Caufield & Byers, Vinod Khosla has funded more optical startups than any other venture capitalist. These are not small-time deals. They're home runs. Here's his record for last year: Two of his startups, Cerant and Siara, were sold for $6.9 billion and $4.3 billion, respectively. A third networking firm, Juniper Networks, went public and is now worth roughly $18 billion. Who knows what Khosla will get from his roster of startups that are still private? Among them is Corvis, which one analyst has estimated could fetch more than $10 billion even though it has yet to ship a product.
Khosla's done a whole lot more than provide networking companies with cash. He's helped build the industry. "He's a visionary. He's able to look out a few steps better than anyone I've ever seen," says William Kind, COO of Siara. Khosla, in fact, founded Siara, where he recruited many of the original employees. He is so enamored of what he calls the "optical revolution" that he's even developed a road show, called the Kleiner Perkins Day. CEOs from eight of Khosla's optical startups give a daylong presentation on the future of networks to top executives at such firms as Qwest Communications, SBC Communications, and MCI WorldCom. Khosla is even a board member of Qwest, a highly unusual position for a venture capitalist whose firm didn't even invest in the company.
Khosla may be the godfather of optical now, but for a time in the early 1990s it seemed that the days of creation were over for this co-founder of Sun Microsystems turned venture capitalist. Back then, Khosla basically dropped out of the VC game. "When Microsoft was getting so dominant and it looked like Windows and NT were going to win everything, it became sort of depressing," he admits. In 1992 he moved his family to India, and for the next three years he worked only half-time at Kleiner Perkins. Khosla, who has large photographs of his children around his office, relished the time off. But in 1995, when the World Wide Web began to take off, he rejoined Kleiner Perkins full-time.
Why the focus on network equipment? "We had become religious about Internet protocol [IP] everywhere," says Khosla. "Once we did that, investing in the communications infrastructure made a lot of sense." Khosla immersed himself in the subject to better understand which companies to invest in. "I remember being on vacation in April 1995, lying on the beach in Hawaii, reading about the physics of optical communications." That's not most people's idea of a vacation. But given his results during the past couple of years, and given the potential a company like Corvis has for the year ahead, his beachside research has clearly paid off.
chart:http://www.pathfinder.com/fortune/2000/01/24/net5.html |
