February 25, 1990

IT’S BEEN a wild week for the computer business in Silicon Valley and beyond. Though John Sculley and his fascinating problems keep Apple Computer near the top of everyone’s “Companies to Watch” list, IBM’s long-awaited explosion into the UNIX work station pool had everyone from Sun Microsystems to Digital Equipment Corp. to Hewlett-Packard counting their fingers and toes.

The capabilities of IBM’s new, long-awaited and blazingly fast line of work stations based on UNIX, an operating system for computers used mostly in science and engineering, were a far cry from a reprise of the underpowered and overpriced RT PC. The RT was IBM’s last, goofy entrance into the reduced instruction set computer (RISC) market in 1986.

As one industry watcher put it, IBM belly-flopped into the pool.

A significant part of the announcement to many RISC work station developers was its low entry price ˆ only $13,000 for a computer capable of calculating at 27.5 MIPS (millions of instructions per second), or $473 per MIPS.

The low price was a surprise to many computer companies. Less than a year ago, for example, Sun Microsystems’ president Scott McNealy was quoted in a New York Times piece anticipating IBM’s upcoming RISC line:
“When you think of IBM, you don’t think of price performance,” he said. “”They’re not going to go in and sell work stations for 750 bucks a MIP (one million instructions per second). They’re stuck with the proprietary overhead of their minicomputer and mainframe business models.”
Ooops.
Despite this slight miscalculation, the Sun/Digital/H-P triumverate, as well as other vendors such as MIPS Computer Systems, already have the hearts, minds and desktops of the scientific and technical community. So it’s clear that IBM’s biggest challenge will come in marketing the product line.

In the personal computing and mainframe worlds, IBM has become the gorilla that can sit anywhere it wants. But the engineers and researchers IBM is trying to woo with the new RS/6000 line are barely susceptible to its bulky persuasion. They want to get their jobs done, whatever they are and no matter how arcane, as fast and cheap as possible.

“IBM certainly has the technology now to compete very effectively. The marketing questions will come down to if they can respond quickly to customer demands,” says Andrew Heller, a former IBM Fellow, now a venture capitalist with Kleiner Perkins Caufield & Byers, who pioneered the RISC work station family at IBM. “They have to target a world comprised of hundreds of niches ˆ that’s UNIX ˆ with a sales force that’s trained to be infinitely broad and only an angstrom (one-billionth of a meter) deep. That’s a hard transition.”

Whether IBM is successful in making that transition remains to be seen. But what strikes me as significant, what’s really the most explosive part of IBM’s announcement, was the chain reaction it caused.

Last week’s trade magazines were full of competitors leaking upcoming product information and lobbing verbal grenades into the gush of positive press about IBM’s new RISC System/6000 product line.

Sun is already claiming it’s going to have a spectacular $5,000 work station out by the end of 1990 that will beat IBM’s price performance. Both Digital and H-P made it clear they’ve got “leapfrog” technologies for shipment this year that will leave IBM in the dust ˆ Digital with a dual-processor, RISC-based work station by mid-year, H-P with a 100 MIPS machine by year-end. All three companies say they’ll upgrade their product lines in the next few months. And IBM itself says it’s already got plans for new products that will just accelerate the leapfrog game.

Although these hardware horse races don’t necessarily spur sales ˆ who wants to buy today’s fastest, cheapest computer when there will be a faster, cheaper one tomorrow? ˆ the pace of these new announcements signal yet another exponential leap forward for computer technology.

Considering that it’s IBM making the waves today, a few companies may get tossed out of the pool in the next couple of years. Common wisdom is that IBM’s now-serious entry into the RISC/UNIX market will shave five or six percentage points off every other vendor’s market share. And if they don’t have that to give up, it could mean trouble for some of the smaller, more innovative ones.

But that’s not a new story, especially not where IBM is concerned. It’s simply interesting, once again, to watch how the drama will unfold this time. As the saying goes, “The more things change, the more they stay the same.