Businessweek Archives

The Powerpc Chip Is Fast. Its Takeoff Isn't


Information Processing

THE POWERPC CHIP IS FAST. ITS TAKEOFF ISN'T

When IBM, Apple, and Motorola joined forces in 1991 to break Intel's lucrative hold on the microprocessor market, it seemed they could get the job done. Leveraging the design and manufacturing prowess of Motorola and IBM and the volume purchases that Apple and IBM could guarantee, the triumvirate produced a speedy chip called the PowerPC that they could use themselves and sell to lure other PC and software companies from Intel Corp.'s camp.

For a while it looked like they might succeed in record time. At a combined cost of $1 billion, the troika has cranked out three blazingly fast chips in the past year, and right on schedule. Last July, IBM created a division to make PCs based on the chips and in November teased customers with futuristic prototypes. Apple Computer Inc. boldly promised to transfer its entire line to the new devices, a process that started with its Power Mac line in March. And Motorola Inc. launched a massive ad campaign and loosed swarms of PowerPC "evangelists" on would-be customers around the world. "We had to stir the pot and get people excited," says Motorola Corporate Vice-President Thomas A. Beaver. "That mission's accomplished."

KICKING IN. Excited, maybe. Sold--not yet. So far, the PowerPC is barely making a dent in the PC business--and probably won't until popular software abounds. Chip volumes have not yet encroached on Intel's market share. And the conditions for making that happen have not yet materialized. Apple and IBM haven't yet agreed on a common technical design for building desktop computers that could be cranked out in volume. And IBM has yet to unveil its new PCs: After hinting at delivery in June, Big Blue has delayed the introduction to the fall. And Apple hasn't kept its vow to license its Power Mac design, although insiders say deals will be announced by October. So the PowerPC chip has sparked little interest among high-volume PC makers. Says Rod W. Schrock, director of marketing for Compaq Computer Corp.'s PC division: "The migration to PowerPC looks too slow for us to be interested."

Not that the PowerPC chip is bombing. Apple expects to ship 1 million Power Macs by March. And supporters say competition from the PowerPC has helped force Intel to slash Pentium chip prices by 40%. PowerPC is well-suited for a range of promising markets, from engine control to set-top TV boxes to handheld communicators. So customers such as Ford Motor, Scientific-Atlanta, and Canon have established it as a clear--if distant--second to Intel.

In the PC market, however, the alliance is making slower progress. Aside from Apple and IBM, only a handful of second-tier PC makers have committed to create PowerPC desktop systems. Like Compaq, Dell Computer Corp. says no thanks so far. Six members of a 30-company consortium in Taiwan have models ready to go. But all are waiting for customer interest to rise before starting production.

When will that happen? It's a chicken-and-egg dilemma. Customers wait for standards and an abundant supply of hardware and applications. Software and hardware developers, meanwhile, want to be sure there is momentum--volume--behind a technology before adopting it. And one common spec that would help ensure that is missing. Right now, for example, Apple and IBM each use different methods to deal with peripheral devices, such as keyboards and floppy drives. Says Dataquest Inc. analyst Kimball Brown: "I think it [one design] is of huge significance, absolutely essential."

Yet even with so much at stake, IBM and Apple have been unable to forge a common hardware standard. It's a technical problem, according to Apple, that would be a "significant task" to solve. Now, both companies are downplaying the need for a single design. Says James J. Buckley, president of Apple USA: "I don't think it's an absolute factor in making PowerPC a success."

There may be more to this than meets the eye, though. Insiders say talks have stalled because Apple cannot see the advantage in adopting a common design that would require it to change its operating software again--and lose valuable time in the market. "It's not a technical decision," says an IBM executive. "The question is what do they [Apple] want to do." After all, sales are looking good. Apple's third-quarter profits rose to a surprising $138 million, vs. a year-ago loss of $188 million--mostly on the sale of 200,000 higher-margin Power Macs. Those sales still fell short of Apple's goal of 300,000 units but went a long way to helping the bottom line.

For now, Apple isn't planning on one common hardware design, although both parties insist they continue to talk. Apple's focus is on trying to license its much-admired software to companies including IBM, Motorola, Toshiba, and Olivetti--with restrictions. That would boost revenues and let Apple control who sells clones and where. But analysts say Apple continues to haggle over licensing terms. And with the new version of Microsoft Corp.'s Windows operating system--code-named Chicago--expected early next year, "the window may be closing," says the CEO of a major PC maker that negotiated unsuccessfully for licensing rights for more than a year but is losing interest.

ARRESTED. Of course, IBM could generate big chip sales by shipping gobs of its own PowerPC machines. But that seems unlikely any time soon. And what software would they run? Observers say the hardware is ready. But IBM may be waiting until its ambitious Workplace OS software is ready, sometime next year. In theory, Workplace would enable a computer to run programs written for IBM's OS/2 and UNIX operating systems on the same machine. But so far, Workplace only emulates--not directly runs--popular Windows software, a prospect that turns off customers.

All this adds up to a case of arrested development when it comes to software. Some 250 programs written for Power Macs are available, but there's little PC software in the works for IBM-based PowerPC computers. With programmers struggling to keep up with demand for Intel-based software, "the PowerPC is just not relevant to the software market, and it's not something we need to spend money on," says J. Paul Grayson, CEO of Micrografx Inc. in Dallas.

That's no surprise, say the PowerPC's undaunted backers. "This isn't a one-year plan," says Motorola Senior Vice-President Barry Waite, who says the PowerPC can capture at least 20% of a $30 billion chip market by 1998. Things could start to turn around earlier--especially if IBM's new computers have pizzazz. Analysts say IBM plans to announce two desktop PCs and a laptop--possibly running Microsoft's Windows NT operating software and with 30 programs--in November. But unless more software is available soon, the PowerPC chip may turn up faster in cars than in desktop PCs.Peter Burrows in Dallas, with Kathy Rebello in San Francisco and Ira Sager in New York


Steve Ballmer, Power Forward
LIMITED-TIME OFFER SUBSCRIBE NOW
 
blog comments powered by Disqus