(FORTUNE Magazine) – Driving into the old Ascend headquarters in Alameda, Calif., you first notice the flags flying in front of the building: One sports Lucent Technologies' stylish, lipstick-red logo; the other, the California State bear. Together the flags offer an apt metaphor for what's going on at Lucent.

More than three years after being spun off from AT&T, the telecom networking equipment company is at a crossroads. CEO Rich McGinn has done a tremendous job melding disparate parts of the old AT&T into a cohesive, powerful unit that is growing at an amazing rate. In late October, Lucent reported $38.3 billion in annual revenue, up 20% over last year, and profits of $3.8 billion, up 46%.

Now McGinn faces a larger challenge: molding the company into a Silicon Valley-savvy conglomerate. The first step was its $24 billion merger with Ascend earlier this year, which is how Lucent came by its informal Valley outpost in Alameda. There has been a flurry of other moves: Lucent opened a Bell Labs research center in Palo Alto; launched a $100 million venture fund with headquarters on Sand Hill Road; bought INS, a company that consults in the design and integration of telecom and datacom networks; and introduced a host of data networking products. All this activity takes Lucent directly into the path of Cisco Systems; it's competing to make networking equipment that supports entire e-industries, rather than remain a manufacturer of old-line telecom switches. "We're reinventing ourselves as a startup," says McGinn. "The fact that we have lots of zeros at the end of our income statement is incidental."

At first glance, the idea of Ma Bell's offspring as a hot Valley startup seems as ungainly as the concept of McGinn in a Village People costume. Lucent's beautifully manicured headquarters in Murray Hill, N.J., looks like an Ivy League campus; Ascend's, linked by outdoor walkways, looks like the high school Brandon and Brenda Walsh attended on Beverly Hills 90210. Lucent execs brag about the number of Nobel Prize laureates (12) who have worked in its research behemoth, Bell Labs; in the Valley few people care whether you invent, buy, or borrow the technology that drives your products. Companies like Cisco and Intel view the Valley as a giant R&D lab. (Indeed, some Lucent execs bristle at the mention of a westward push, pointing out that the company has employees all over the world and that Ascend actually has more staff in the Boston area than in the Valley, thanks to its 1997 acquisition of Cascade.)

Behind the cultural conflict lies a serious change. Lucent is fighting a prevailing--if naive--mindset that it is stuck providing for old voice networks, while Cisco dominates the equipment behind Internet protocol networks. The former is a stodgy, slow-growth business; the latter, an exciting, hypercharged one. Even a bull analyst like Paul Sagawa of Sanford C. Bernstein says, "I always compare Lucent to the story about the blind men and the elephant. One blind man rubs against the elephant's leg and thinks it's a tree. Another one thinks the trunk is a snake. Depending on where you touch Lucent, you get a different story."

He's right. The truth about this company is complicated. But that's less an issue of mixed messages than of the fact that this industry is more nuanced than a simple voice vs. data battle. AT&T and the Baby Bells aren't about to dismantle their voice networks to build data ones. For that reason, the equipment companies most likely to succeed will develop products that link voice networks with the data ones being built by Qwest, Williams, and Level 3. "The real word isn't 'convergence,' but 'integration,' " says Sagawa.

This is where Lucent is making great strides. After shopping around the industry, it nabbed Nexabit, a maker of high-speed routers that direct Internet traffic. Yes, Nexabit is based in Marlborough, Mass., not Palo Alto; but its technology is decidedly Valley--best of breed, say analysts. Some of Lucent's projects from Bell Labs look like winners too. It has generated buzz for a project called Stinger, a switch placed in a telecom central office that enables DSL technology. (DSL is a process of turning copper phone wires into high-speed data conduits.)

But the most impressive product to come out of Lucent in the past year--and one that indicates Lucent may be able to make the transition to a Valley data-centric company--is "softswitch." Placed between a voice and a data network, this switch seamlessly hands telephone calls back and forth. It was developed to bring the reliability and inter-operability of the phone networks to newer Internet Protocol and packet ones.

Though softswitch has already generated $1.7 billion in contracts and is being tested by 40 companies, its biggest validation came this past summer, when Level 3 selected it as the platform for its nationwide network. For Lucent, the press generated by the announcement--mentioning the words "data" and "Lucent" in the same sentence--may have been as important as the revenue.

Now Lucent's challenge is to turn products like softswitch into huge businesses, given that spending on voice networks will dwindle, says Michael Neiberg, managing director of Hambrecht & Quist.

But with each new business it enters, Lucent encounters a host of competitors: Texas Instruments in chips, Ericsson in wireless components, Copper Mountain Networks in DSL equipment, Juniper Networks in high-speed routers, and Nortel and Cisco in just about everything else.

Wall Street doesn't like this kind of uncertainty. The company's stock, now trading at around $62, is down 21% from its high earlier this year. Lucent's accounts receivable were up in the third quarter, ended in June, and inventory turnover slowed--a combination that caused some to speculate that Lucent was trying to push products on customers who didn't want them. Year-end results released in late October were better--but it's an area that Lucent needs to watch vigilantly in order to keep Wall Street happy.

In the end, it's McGinn who may be Lucent's most valuable asset. Tapped as a relative unknown, he has proved to be a talented manager with a keen sense of the industry's future. (He recently realigned Lucent's businesses on the basis of customer markets rather than technology segments.) Lucent should continue to thrive under the guidance of this affable, convivial, intelligent guy. Most important to shareholders, McGinn understands that life doesn't begin and end in Murray Hill, N.J. (At 53, he has a 10-year-old daughter and a new son, born this fall.)

"[Lucent's] culture is much more open, accepting, and fluid than one might think of a company that had a long parentage before it went public," he says. Open? Accepting? Fluid? McGinn's got the lingo down--if his strategy is as on target, Lucent will thrive in the new order of networking.

plus: MA CASE AND HER BOYS | GROOVY NET MUSIC | OUR FIRST "WEB PAGE" | ALSOP'S BAD TRIP