Nextlink is the kind of competitive communications provider that is infuriatingly unpredictable, even though its strategic plans probably have been in place for years. With every move the company makes, observers utter a collective "A-ha!" and kick themselves for not calling it earlier. In retrospect, it usually makes perfect sense.

That seems to be the trend for companies that are part of the communications collective of Craig McCaw, and this one is at the center of the hive. Nextlink Communications began life six years ago as a competitive access provider. It grew up, fed by McCaw's investment, to become an inter- and intracity fiber network - the kind of valuable, high-capacity backbone that can be used in several ways. Then the question became what Nextlink would do with its growing fiber foundation: Would it become a wholesaler of optical capacity, or would it build a data competitive local exchange carrier presence using an access technology such as DSL?

It did neither. Whether by design or by fortunate timing, the Nextlink fiber project collided with McCaw's seemingly innate understanding of wireless maturity and a government-issued spectrum opportunity, and a broadband wireless access opportunity was born.

That happened early last year, shortly after Nextlink emerged from the local multipoint distribution service (LMDS) auctions with a fair amount of spectrum. At that point, it still was unclear what company, if any, was going to dominate the LMDS realm - especially given the established presence of companies such as Teligent and Winstar Communications in other portions of the spectrum. The winner of the LMDS auction was an unknown entity called WNP Communications.

It wasn't until Nextlink acquired all of WNP's licenses and became the largest LMDS spectrum-holder in the country that everyone said "A-ha!" - again.

"There are a lot of parallels with what McCaw's been doing in this business and what he did in the cellular business," says Bob Egan, networking research director at GartnerGroup. "By bankrolling that much spectrum, he's made an investment in pure bandwidth."

There have been several more calculated surprises since the WNP acquisition: Last fall Dan Akerson, former chairman and CEO of Nextel Communications, became Nextlink's new chairman and CEO in what can only be described as a McCaw-esque management move. That appointment was followed by a string of other management shifts and departures as Akerson assembled his new team. And earlier this year, Nextlink moved to acquire Concentric Network and, with it, the potential for a cache of Web hosting and IP service capabilities.

With Nextlink loaded with fiber, wireless and Internet technology and a capable executive suite, most of the exaltations of surprise and triumph these days are coming from Wall Street. The challenge for Nextlink now is to make good on its technological and economic promise in a manner that keeps the fickle investment community - and its mid-sized business customer targets - appeased.

Foundation first

Nextlink Chief Technology Officer Doug Carter represents the technological brains behind the operation. Like others in the company's executive ranks, he is relatively new to the company, having joined it in January 1998. But he is not new to McCaw enterprises, and because of that, he is not new to radio technology.

Carter's first radio/McCaw exposure was in 1984, when he was CTO of a pre-cellular radio company that was purchased by McCaw. He held several engineering and operations positions at McCaw Cellular Communications. When McCaw Cellular was acquired by AT&T in 1994, Carter stayed and became senior vice president of network operations for AT&T Wireless Services. He returned to the McCaw nest in 1998 to work simultaneously for Teledesic and Nextlink before devoting his time entirely to Nextlink.

That depth of wireless technology experience is necessary for what Carter is charged with accomplishing at Nextlink: achieving fiber-like reliability in the access network using radio technology over spectrum never used for such purposes, while also bolstering a nationwide fiber network to support and supplement it. That strategy, Carter says, was in place before Akerson was brought in to run the company.

"There really hasn't been a change in the technology direction since Dan came in," Carter says. "The foundation began in early 1998. The well-defined plan of how the pieces come together began in early 1999. Since Dan joined, the biggest change has been the intensity of focus."

That intensity includes the goals of extending local facilities to 60 cities this year using fiber and broadband wireless and continuing to build a 16,000-mile fiber optic network through a venture called Internext. Internext is fueled by a dark fiber cost-sharing agreement with Level 3 Communications, under which the Nextlink fiber network gets 24 fibers, one 1.25-in. empty conduit and rights to 25% of the fiber in Level 3's conduits after the first five. For Nextlink's business formula to be successful, the fiber and wireless elements must work seamlessly in tandem.

"We have a unique set of assets in being the largest LMDS spectrum holder in the country and undergoing a national fiber buildout of the largest cities," Akerson says. "We will be able to get best-in-class last-mile economics because of the combination of those two technologies. If you want to make yourself more attractive and more available to a larger segment of the market, you use different technologies to serve them."

Those assets are ones that other integrated communication provider hopefuls don't possess, Egan says. "It's all about owning your own interfaces and being able to provide bandwidth on your own network."

The reality, however, is that one of the important components of Nextlink's strategy is still essentially vapor. While the LMDS licenses clearly have promise, the company is facing several hurdles - not the least of which is a nascent market for broadband wireless infrastructure and subscriber equipment. To date, Nextlink has tested gear from SpectraPoint Wireless (a Cisco Systems/Motorola joint venture), Ericsson, Wavtrace and Digital Microwave Corp. and has started deploying the SpectraPoint equipment in Los Angeles and Dallas, Carter says. But the fact that the company has not publicized any large-scale equipment contracts could be an indication that it hasn't made final decisions or that the amount of gear it requires simply isn't available yet.

"That part of the market's so immature - it doesn't surprise me that Nextlink hasn't spent a dime on it," Egan says. "I don't get the sense that the infrastructure providers know how they're going to mature the infrastructure or the terminals."

Air time

Lack of equipment readiness doesn't appear to be keeping Nextlink idle, however. The carrier's broadband wireless strategy is well-developed, and it is taking advantage of more established network formats first while it waits for the potentially more effective ones to come into their own.

The formula behind the broadband wireless access project is relatively simple: A fiber network expansion is fixed, expensive and time-consuming. Wireless is flexible, less expensive and faster.

"Numerically, Nextlink gets to the majority of its customers on fiber and copper," Carter says. "The radio licenses allow us to increase the amount of bandwidth we can get to individual customers in a short period of time."

On top of that, Nextlink is installing what it describes as a "thick" fiber network made of hundreds of optical strands. This will give the company more long- and short-haul capacity without having to rely on technologies such as wave division multiplexing, at least in the foreseeable future. But it also means an access bottleneck.

"We have this enormous capacity in the intercity, but it is not in itself particularly useful," Carter says. "I can get terabits between the 60 points across the country, but what do I do when I get there? This enormous capacity only goes to a small number of points, and it's going to take a long time before we can get fiber to each of those buildings."

Getting everywhere with fiber is certainly Nextlink's ultimate goal in an ideal world, but the math dictates otherwise. That begs the question of whether deploying broadband wireless is a temporary fix. GartnerGroup's Egan believes the technology will have a long-term presence in most networks.

"There are always going to be markets where wireless will grow up to have good performance, and it will always have a better cost structure than fiber," he says. "Fiber is a big investment even if you don't use it. LMDS is an organically grown bandwidth investment, and you can't get the same kind of economics out of fiber."

Nextlink's decision about extending its fiber core using enhanced copper technologies such as DSL or using wireless was a capacity issue, Carter says.

"My principle mission in life is to address that inevitable choke point that Level 3 and other companies are going to slam into," he says. "Because of the noise and the quality of the copper plant, you can depend on a half megabit to about a megabit on the copper."

That kind of capacity and the technological shortfalls of DSL simply won't cut it for the businesses competitive carriers are targeting, Egan says. "The LMDS market is really going to be a boon for small businesses that just can't get the bandwidth they need," he says. "The DSL market is quickly becoming commoditized, and the reach just isn't there."

Nextlink's strategy for unchoking the choke point is straightforward, partly because of the simplicity of broadband wireless. Nextlink's first priority when it enters a city is to connect large buildings to its fiber network. The first step to extending that presence to smaller buildings, Carter says, is to go to the roof with a "radio propagation analysis tool" - a pair of binoculars with a range finder.

"Those medium-sized buildings are just sitting out there," he says. "We can get DS-3 and OC-3 rate point-to-point radios to them quickly, and we have 45 Mb/s and 155 Mb/s services into buildings that previously had a collection of T-1s. Now all of those buildings become candidates for multipoint radios."

Therein lies the issue of technological readiness. In broadband wireless, point-to-point radios - which have been used in networks for decades and have proven reputations for fiber-like reliability - are considered more expensive, dense and inflexible. Point-to-multipoint represents the hipper new technology that is cheaper to deploy and more capable of allocating smaller swaths of bandwidth to more customers - in theory.

"With point-to-point, we concentrate all the power on a laser-thin beam so we can get higher speeds," Carter says. "When we fan the beam out with the multipoint system, we will take the speed to about 10 Mb/s or so. This is important because we can go to smaller buildings and spend less time doing it. In my view, multipoint is an important component, but point-to-point is equally important."

Nextlink's commitment to point-to-point is likely due in part to the fact that point-to-multipoint solutions are underdeveloped, Egan says. "Part of that is because there's no story around point-to-multipoint," he says. "It's easy for them to be confident about point-to-point because the equipment's been around."

On the other hand, the fact that Nextlink can afford to invest in technologies that other carriers might consider too pricey could be attributed to the company's success and ample backing. "We have the luxury of being able to use multiple technologies and multiple tools that are optimized to the task we're trying to achieve," Carter says.

The Internet and beyond

Nextlink's pending $2.9 billion acquisition of Concentric is, in Carter's view, the component of the company's strategy that completes the technological puzzle. To date, that puzzle consists of broadband wireless radio licenses, an expanding high-capacity backbone using dark fiber built for Nextlink by Level 3 and an increasingly national collection of metropolitan fiber.

"The piece that was missing was the IP services stack," Carter says.

The search for a suitable acquisition target was occurring simultaneously with an organic ISP initiative within Nextlink, which included the hiring of John Curran, former CTO of GTE Internetworking, as vice president of Internet technology.

"He's one of the true Internet people," Carter says. "We brought him on board to help us understand how we should fill in the IP services stack with the objective of being an Internet service provider."

With Curran's expertise, Nextlink identified Concentric as the best acquisition target.

"We didn't want to wait until we had all the physical assets built and then start filling in our Internet strategy," Carter said. "Concentric neatly plugs in the last piece of the puzzle. It's like a socket locking onto a wrench."

For Nextlink, the addition of Concentric will provide applications expertise to a company already rich with network facilities - a potent combination that could ultimately boost the carrier's bottom line. "It's less physical, more virtual - more data and less voice," Akerson says. "It changes the character of the revenue profile."

Nextlink has other technological issues to contend with, including its operations support systems.

"When I first arrived, I spoke openly about some of the systems issues in the company," Akerson says. "They still exist, but we brought in a new chief information officer and we've crafted a blueprint of how we think we can have better than average back-office systems."

Part of that blueprint already is under construction. Nextlink recently announced that it has deployed electronic bonding systems for use with Pacific Bell, BellSouth, U S West, GTE, Bell Atlantic South, Ameritech and Southwestern Bell. The systems' purpose is to facilitate the electronic passage of customer ordering information and move toward fully automated provisioning.

Finally, Nextlink signaled its border-crossing intentions with the acquisition of broadband wireless licenses in Canada in partnership with Wispra Networks, the provisional winner of six fixed broadband wireless licenses in Toronto, Montreal, Vancouver, Ottawa, Edmonton, Calgary and surrounding areas.

Even with all of that, Nextlink is still a company in development. For one, the company has brand recognition in the investment community and within the communications industry at large, but the extent to which the company name and service suite is recognizable to potential customers is arguable. Part of that could be by design.

"I don't want to operate too far above the horizon from a marketing point of view," Akerson says. "I want to get our horses aligned before we establish a brand."

Even if it hasn't achieved nationwide recognition for its brand, Nextlink certainly has demonstrated success in selling services to date. But the model for how those services are delivered - and, indeed, the services themselves - is in the process of changing quite dramatically. For example, Akerson predicts that by the end of 2000, Nextlink's network will carry equal amounts of data and voice traffic. All of that contributes to how Akerson rates Nextlink's short- and long-term priorities - and, ultimately, to the company's success.

"People tend to make this business more complicated than it really is," Akerson says. "We must deliver a quality product and deliver a bill every month. We have to get the Concentric acquisition completed and integrate the two management teams. We must continue to execute the plan and deliver results on shareholder expectations - including the continued buildout of both the wired and wireless networks. And we have to develop a new set of products that are not in the marketplace today."

In addition to all the network advancements Nextlink Communications has made during the past year alone, one of its major strategic acquisitions in the second half of 1999 was not technological but human.

Dan Akerson, former president and chief operating officer of MCI and former chairman and CEO of Nextel Communications, has the management skills necessary to run a big company that has the potential to get much bigger. He also has the reputation of leading Nextel, another McCaw company, out of the technological doldrums and taking it to new financial heights - experience that has direct bearing on his current assignment.

"We're about at the same competitive position as Nextel was in late '96 and early '97," Akerson says. "That's not to say it's bad, but that the company is building the foundation to be a fully integrated communications provider that will be a competitor to AT&T and MCI WorldCom."

Putting Akerson in the top slot apparently was important enough to McCaw and Nextlink's board and shareholders to warrant a $30.9 million fourth quarter charge last year incurred as a result of moving the company's headquarters from Bellevue, Wash., to Northern Virginia, where Akerson himself is based.

But Akerson appears to have connections that could make that kind of expense worthwhile: One of the announcements to come out of Nextlink following his appointment was an $850 million investment in the company by Forstmann Little, the investment firm where Akerson was a general partner between his MCI and Nextel days. Other efforts - including a $1 billion senior secured credit facility that closed last month - have provided Nextlink with funding into late 2001, company reports say.

Nextlink also has continued on a similar financial trajectory in the one quarter since Akerson took the chairman and CEO seat. Revenues for 4Q99 were up 20% during the third quarter, giving Nextlink a 1999 revenue of $274.3 million. And whether it is because of Akerson's presence or not, Wall Street appears to be responding favorably to Nextlink's business and technological progression. The company's stock performance has improved steadily, splitting just before Akerson joined the company and again since.

Industry observers echo that sentiment. "He's been reasonably quiet and methodical, and that gets high points," says Bob Egan, networking research director at GartnerGroup. "Wall Street certainly appreciates it. He's not overestimating what they can deliver, and he's putting together the basics."