Amdahl Escapes `Death Valley,' But Now What?Robert D. Hof
It looked like the end of the world. In September, 1993, recalls E. Joseph Zemke, chief executive of mainframe computer maker Amdahl Corp., sales simply dried up. Already buffeted by smaller, cheaper machines, Amdahl had nothing on its order books. "It was like Death Valley," Zemke recalls. "I still don't know what happened." But he knew what to do: hunker down, close factory lines, and slash the workforce for the third time in nine months.
The bitter medicine worked. The Sunnyvale (Calif.) company surprised Wall Street on Apr. 25 with a $7.1 million first-quarter profit--the first since late 1992. But Amdahl's far from out of the woods. Almost all of its $1.7 billion in revenues came from sales of IBM-compatible mainframes and related gear, a market that declined 21% last year (chart) because of drastic price cuts of 35% a year since 1990.
Amdahl's survival plan will require a major shift in its approach--and a bit of luck. Instead of being an adept follower of market leader IBM, Amdahl must forge its own path--for the first time in its 20-year history. Its plan: take its own mainframes, combine them with the industry's hottest hardware and software, and sell the package to large corporate customers. To do that, says Gartner Group analyst Charles C. Burns, "the whole company needs to dramatically change its stripes."
On the farm. To Zemke's credit, Amdahl has gotten off to an aggressive start. During the past 10 months, it has forged three major alliances--with Electronic Data Systems, Sun Microsystems, and nCube, a maker of "massively parallel" computers that crunch databases at blinding speeds. With EDS, Amdahl has a joint venture to market Amdahl's Huron software-development system, while the deals with nCube and Sun call for Amdahl to resell those manufacturers' machines to Amdahl customers.
In addition, Amdahl is now farming out mainframe work. In order to move quickly to lower-cost processors using the CMOS (complementary metal-oxide semiconductor) technology that is used in personal computer chips, Amdahl is relying heavily on Fujitsu Ltd. The Japanese company, which owns 45% of Amdahl's stock and which has supplied much of its technology, now will take over the development and possibly even manufacturing of the new machines.
But ultimately, reselling other companies' machines simply won't bring in veryattractive profits to Amdahl. Analysts estimate that the company's sales of Sun machines, for instance, carry gross margins of only about 20%, below already compressed 28% mainframe margins (chart). Zemke says that it's worth the price: This way, Amdahl machines still anchor the customer's data center.
That will allow Amdahl to play its aces in the hole: service and consulting, which consistently top even IBM's in market researchers' surveys. Amdahl's maintenance, support, and consulting services--already making up 28% of sales--grew 11% last quarter, partially countering the drop in hardware. And nonhardware gross margins are at a hearty 45%, some 17 points higher than hardware.
But the long-term picture is less clear. Robert Djurdjevic, president of Annex Research, thinks that Amdahl should focus its attention more on specific markets. "They're throwing a dart here and a dart there," says Djurdjevic. "What's missing is a blueprint."
High risk. Moreover, Amdahl's record of developing and marketing new technology is decidedly mixed. The company dropped a long-awaited Unix mainframe proj-ect last year because it would have cost too much to build. So customers may well be wary of fresh technology forays, such as the nCube deal. To buy that setup, says John Wood, vice-president of computer and network services at Royal Bank of Canada, would be "a pretty high-risk thing to do." Meanwhile, the competition is intensifying. On Apr. 27, Hitachi Ltd.--the third player in IBM-compatible mainframes--announced plans to use IBM's chips instead of creating its own.
Still, Amdahl has some advantages. Its reduced size means that the company can move more quickly than rivals suchas IBM. The question is where to move. "They're going to have to demonstrate that," says Stan Johnson, information services manager with the Port of Los Angeles, "or they're going to become a very minor player." Or find themselves in Death Valley once again.