Here comes the deluge. After a couple of years of talk, a shopping cart full of network computers is now coming to market from the industry giants. But the fortunes of a $120 million company--with the ultra-timely name of Network Computing Devices--may tell just as much about the future of this massive effort to make Scott McNealy's claim, "the network is the computer," actually come true.

Born a decade ago as a "dumb" X terminal supplier, NCD, of Mountain View, Calif., manufactures the network computer that IBM began selling late last year. Now it hopes to grow into a sizable company by supplying network computers to server vendors and integrators. "Network computers are us!" says effusive CEO Robert Gilbertson, who in 1988 was given the Turnaround of the Decade award by CFO magazine for reviving Data Switch Corp., a Connecticut telecom equipment supplier.

But Gilbertson didn't start NCD. The company--and, indeed, the X terminal market--were pioneered by Chief Technical Officer Doug Klein, who in 1988 began using the term "network computing" to describe a Unix environment where the application and data were run on a central server and accessed from simple, low- cost terminals. With the help of co-founders Judy Estrin and Bill Carrico, Klein got the X terminal off the ground. But the business never really soared and trailed off fast after X terminals gained the reputation for being network capacity hogs. NCD foundered after Estrin and Carrico jumped ship in 1994. A series of disappointing-to-disastrous ventures followed. NCD lost $4 million that year and another $5 million in the first half of 1996.

Then it was Gilbertson's turn. He rode into town for his job interviews last spring armed with a blissful ignorance of the finer points of NCD's business. "I didn't know what an X terminal was," he admits. But maybe that was what let him see what no one else seemed to notice--that network computers weren't much different from X terminals. Both allow corporations to replace fully functioning desktop computers with simple, easier-to-administer and cheaper alternatives. "Larry Ellison didn't invent the idea," says Gilbertson, "but he has done us a great service by focusing attention on it again."

Gilbertson didn't keep his revelation to himself. Within a few days of joining NCD, he was able to convince IBM that his company's experience with X terminals made it the perfect partner to build IBM's network computer. NCD started shipping a trickle of units to IBM late last year, but already they've been installed in about 150 customer sites. "NCD was chosen because of their history in NCs," says John Reilly, spokesman for IBM. "We did a review of the industry, and for our current needs, NCD delivered."

IBM is pushing its Network Station hard, and NCD has much to gain from its alliance. Last November, IBM formed a new NC division out of its AS/400 group. The Network Station is optimized for all IBM servers, from PC servers to RS/6000 to AS/400. So there's a potential replacement market of 12 million-plus dumb terminals that IBM has installed in decades past.

Beyond IBM, NCD hopes to provide several server vendors with NCs. The company is already in talks or trials with eight other vendors, although Gilbertson declines to name names at this stage.

He's got a unique proposition. While the machines are designed to run Java- based applications, they work with Windows NT, as well. Even before the advent of the NC, though, NCD was already moving closer to Microsoft. Its installed base of X terminals has almost as many clients running Microsoft Windows applications as Unix applications. And Gilbertson is trying to convince Microsoft that he's a friend. He estimates that Microsoft will gain at least 100,000 extra Windows NT licenses next year through NCD installations. "We want them to understand that we are not a threat, but a complementary channel," says Gilbertson.

Gilbertson's strategy of making his machines fit snugly into any computing environment sets him apart. He refuses to align with what Gartner Group calls the "Vendors Against Microsoft" coalition, or with the NetPC concept developed in response by Microsoft and Intel. Instead, he focuses on giving corporate customers what they say they want: an inexpensive, low-maintenance alternative to PCs.

Analysts are still cautious about NCD's staying power. The company showed a modest profit for fiscal year 1996, but it was still so short of cash in the fourth quarter that instead of paying employees a bonus, Gilbertson sent everyone home on Dec. 20 for a vacation. Plus, the company has an image problem to overcome. "I don't think they've totally shed their X terminal heritage," says Neil McDonald, research director for Gartner. The problem is more than skin- deep. NCD's technology is still too Unix-centric, McDonald believes.

The key to NCD's success will be its ability to ride the NC wave of hype without wiping out. Even company executives seem a little unsure about their prospects. NCD's 1997 sales projections to the investment community range from a modest $135 million, about 12 percent growth, to $250 million, more than 100 percent growth. It all depends on how fast the NC market takes off--or if it ever does.

If IS managers tune out too quickly, or perceive the NC craze to be just one more vendor plot to get them to spend money on unproven technology, then Gilbertson's "Network Computers R Us" push could misfire badly.

But if NCD can deliver on its stated goal--to provide a cheap, low-maintenance terminal that truly gives users even-handed access to any operating environment-- then Gilbertson might just have another Turnaround of the Decade award to look forward to.

Source: PC Week