Science & Technology
IS BIG BLUE STILL BIG ON RESEARCH? YOU BET
Things looked dark for IBM's vaunted Research Div. in early 1993. It had risen from a converted fraternity house on New York's Upper West Side to become a research powerhouse with back-to-back Nobel prizes in physics. But its budget was being slashed, and there was talk that it would be dismantled along with the rest of the company. Besides, the new chief executive of IBM was a cigarette guy. What use would the former head of RJR Nabisco Inc. have for scientists and computer geeks who wouldn't know a sales call from a dividend?
Just over a year later, it appears many of those fears were exaggerated. Chairman and CEO Louis V. Gerstner Jr. is turning out to be one of the most aggressive exploiters of IBM Research since Thomas J. Watson Jr. founded the division in 1956. Gerstner's predecessor, John F. Akers, was moving toward turning IBM into a brace of Baby Blues. By contrast, Gerstner is intent on using the technological strength of a unified IBM. And he's leaning on IBM Research to help. So instead of being relegated to the role of visionaries or troubleshooters, researchers are being brought into everything from product planning to corporate strategy-setting.
The changes at IBM Research aren't all good news, to be sure. Like AT&t Bell Laboratories and other industrial labs, the division is becoming less of a national and global scientific resource as it focuses more on the needs of the company that pays its bills. Universities and government labs aren't picking up the slack. "If IBM is getting out, who's going to do it?" asks Linton G. Salmon, manager of microelectromechanical systems at the National Science Foundation.
Still, after an awesome $13 billion in losses in 1992 and 1993, IBM and its shareholders were in no mood for altruism. The impressive thing is that the labs have remained as intact as they have. True, Gerstner has made further cuts in the division's budget--to $500 million this year from $550 million in 1993 (chart). But proportionally, the Research Div. has been cut back less than the overall budget for research, development, and engineering. And it's one of the few parts of IBM that has been spared layoffs. "We've never had one, and it's my intention that we'll never have one," says James C. McGroddy, IBM's research director. "As far as I can tell, I have Lou's support."
What's more, half a billion dollars is a substantial sum for science. The division continues to maintain 2,600 people in Westchester County, N.Y., San Jose, Calif., and Zurich. Notes Jeffrey M. Jaffe, the Research Div.'s vice-president for systems and software products: "We're not exactly suffering. Other research people I talk to would kill to get our budget."
Fresh thinking. Gerstner demonstrated his commitment to the division early on, when he made the Thomas J. Watson Research Center in Yorktown Heights, N.Y., his first official visit from the Armonk headquarters. He since has turned to researchers to fill top advisory posts: Former Vice-President for Physical Sciences and Technology Dean E. Eastman, for instance, is leading a project under chief strategist James A. Cannavino to re-engineer IBM's hardware-development process. And researchers were gladdened when Gerstner appeared in the latest annual report with an odd-looking laptop computer prototype, code-named Leapfrog, that came out of the division.
Of course, McGroddy knows that the only way to retain Gerstner's support is to prove that the division is vital to IBM's future. To stimulate fresh thinking, he has installed a brand-new team of four vice-presidents since last year. He also is bolstering joint programs with operating divisions and encouraging his researchers to collaborate directly with outside companies, such as Siemens and Toshiba in memory chips, Analog Devices in chip materials, and Seiko Telecommunication Systems in traffic data terminals. The interchange helps IBMers grasp real-world problems: "It's been like engineers dealing with engineers," says Michael C. Park, the Seiko unit's vice-president for business and network development.
Most controversially, McGroddy is devoting more resources to research in software, which is increasingly important to IBM, while deemphasizing research in the physical sciences. Some of IBM's top physicists have departed, while others have taken emeritus status, including Benoit Mandelbrot, a pioneer in chaos and fractal theory.
McGroddy has little patience for complaints that physics is being given short shrift. He says that the technologies that set IBM apart in the marketplace are more and more related to software and system design, not such base-layer hardware as superfast semiconductors. The division has slashed or abandoned work in gallium-arsenide chip materials, bipolar transistors for mainframe computers, and the costly thermal-conduction modules that keep bipolar chips from overheating. Meanwhile, it's pumping money into such software-related fields as optimization, a fertile branch of math that can help IBM customers ranging from airlines to supermarket chains.
Galaxy study. The search for relevance predates the latest crisis at IBM. Ralph E. Gomory, who ran the division from 1970 to 1986, instituted the first joint programs with the operating divisions, and John A. Armstrong furthered the efforts in his tenure as research chief from 1986 to 1989. The goal of the joint programs was to speed products from lab to market. But forming them required surmounting huge political obstacles. Despairing of change, William T. Siegle quit as director of IBM Microelectronics' Advanced Technology Center in 1990, before its developers were united with the Watson researchers. He became chief scientist at Advanced Micro Devices Inc. "Frankly," says Siegle, "one of the things that motivated me to leave was fear that it couldn't happen." But the union did happen. Today, researchers and developers collaborate so closely in IBM's Semiconductor Research & Development Center in East Fishkill, N.Y., that it's impossible to tell who's who without reading their badges.
The pressure to be "vital to IBM"--a stock phrase--is teaching cloistered scientists what business is all about. Take Roger H. Koch, manager of superconducting science & technology. He was rebuffed internally around 1990 when he broached the idea of developing superconducting sensors for the U.S. Navy. "They said: 'We don't deliver products; we deliver paper,'" recalls Koch. Now, Koch is managing that very job. The rewards, he says, are different: building things rather than satisfying curiosity. Adds Koch: "The people who are willing to change their heroes, they do very well."
Relevance notwithstanding, McGroddy doesn't want his agenda controlled entirely by short-term considerations. History shows why: IBM's mainframe computer executives long opposed building parallel-processing supercomputers that could operate independently of a traditional mainframe, fearing that they would cannibalize mainframe sales. But the Research Div., using its own funds, persisted with the project anyway. So when the mainframe execs finally bowed to the inevitable, the company was ready with products.
Today, with Gerstner's approval, only about half of the Research Div.'s budget is devoted to supporting projects in the operating units. For the rest, the division sets its own priorities. Meanwhile, a slim 3% of the Research Div. budget--down from 4% a few years ago--goes to what is awkwardly called "disconnected" research. While some projects may end up in products some day, others seem driven mainly by curiosity or a quest for prestige. Researchers are studying phenomena ranging from quantum corrals to interacting galaxies to a mathematical model of epilepsy. One researcher recently calculated the mass of the proton and other particles, "providing new evidence for quarks and the theory of quantum chromodynamics," the division's magazine said last year.
New stress. IBM Research isn't in the clear yet. After three years of cuts, morale below the top ranks is weak. Last year, only half the respondents answered favorably to a question about satisfaction with the division. "There's still a lot of nervousness, a lot of skittishness," says Mark F. Bregman, who as vice-president for systems, technology, and science oversees the hardest-hit units.
If McGroddy can keep the money coming in, that anxiety should diminish. But researchers are coping with a new kind of stress: overwork. Wrote one employee in TALKJIM, an electronic forum with McGroddy held earlier this year: "The short-term focus and the emphasis on development has left little time and energy for the pursuit of technical vitality." The pressure for results is even higher at the top. The overbooked Bregman appears wistful that he may never get back to measuring the mass of the neutrino, a favorite project that now is likely to be passed off to a university. Bregman, 36 and married with two children, starts work at 7 a.m. and occasionally stays past 10 p.m. "There's just not slack anymore," he says.
At least, though, it's nice to be considered useful. In the Ivory Tower era, when researchers were prized principally for their I.Q.s, "everyone was jostling to prove who was smartest," says Bregman. "We don't have time for that now." With its increased focus on cost-effective results--rather than endowing the world with scientific discoveries--IBM Research is "becoming a little more like the rest of the [electronics] industry," says AMD's Siegle. For a company facing challenges as daunting as IBM's, that's probably a sensible thing to do.
IBM RESEARCH: GETTING MORE FOR LESS
IBM Research Director James C. McGroddy is seeking more useful results in spite of a $125 million budget cut. Here's how:
CUT PHYSICAL SCIENCES
The Research Div. has cut base technologies--including semiconductor materials, chip packaging, displays, and its Nobel prize-winning physics research--to 25% of its budget, down from 40%.
BEEF UP SOFTWARE
Under McGroddy, software and services research is up to nearly 25% of the Research Div. budget, from around 5% three years ago. Software was previously considered too "applied" for the research staff.
Researchers are leaving the ivory tower to work more closely with IBM operating units, with outside customers, and with partner companies, such as Siemens, Toshiba, Analog Devices, and Seiko.Peter Coy in Yorktown Heights, N.Y.