Deep within IBM's Almaden Research Center, a stunning hillside campus tucked away in a nature preserve south of Silicon Valley, Dr. Donald M. Eigler presides over a three-room nook that bristles with tools to pick up, move and corral individual atoms.
Like Eigler, 48, a physicist who wears his hair in a ponytail -- and pads around Almaden with his dog named Argon, after the chemical element -- the equipment in the lab is both sophisticated and approachable. The million-dollar atom movers operate in vacuums at temperatures so low that the atoms are nearly rigid. The machines produce sharp pictures of individual atoms, yet the controls in an adjacent room are an everyday keyboard and mouse.
That allows Eigler to tell visitors, "You can't leave without moving an atom." Position the cursor over what appears on the computer screen to be a small, side-lit mound on a flat plane -- it's actually a cobalt atom on top of a fingernail-size wafer of crystallized copper. Click and drag, then let go of the mouse. That's all there is to it.
PHOTO: NY TIMES
The International Business Machines Corp grabs every chance it gets to provide customers, business partners and journalists a moment to horse around in these nanoscale dimensions. It sends an unmistakable message about IBM's research capabilities.
Even so, work like Eigler's, which probes the secrets of the universe but has not -- so far at least -- resulted in products for IBM is not what research experts find most fascinating about the 3,200-person global empire of IBM Research. Instead, they are struck by the way IBM's laboratories stand out for a more practical reason: among the estimated 35,000 corporate research labs in the US, IBM's are the gold standard for turning research into corporate profits.
Patents
PHOTO: NY TIMES
Last year alone, IBM was awarded 2,922 patents in the US -- an average of more than 11 each working day, and 43 percent ahead of NEC of Japan, in second place. It was the eighth consecutive year in which IBM received the largest share of patents in the US (The national total last year was 176,087).
Of course, not all of these patents ranked with revolutionary IBM inventions of the past -- like the hard-disk drive and the scanning tunneling microscope, the first microscope capable of forming images of individual atoms. But they cumulatively fattened a huge portfolio of patents that generated US$1.7 billion in licensing fees alone in 2000, not to mention what the underlying innovations did for IBM's own product lines.
"If you had to benchmark your laboratory against only one other these days, it would be IBM," said John Seely Brown until a year ago the director of the Xerox Palo Alto Research Center, the widely admired laboratory that invented the laser printer, Ethernet networks and the point-and-click method of using computers.
Paul M. Horn, the senior vice president who runs all of IBM Research from its flagship laboratory in Yorktown Heights, New York, figures that he has his charges in the right frame of mind to preserve that reputation. "Even the guys doing exploratory stuff know it's all about the success of the parent," he said.
A decade ago, IBM seemed an unlikely role model. In the early 1990s, IBM Research was buffeted by layoffs and other belt-tightening measures, including the end of most research on new classes of ceramic superconductors, a field of physics in which researchers at the IBM lab in Zurich won a Nobel prize in 1987.
New direction
The arrival in 1993 of Louis V. Gerstner Jr. as IBM's chairman and chief executive coincided with speculation on Wall Street that the company might be headed for bankruptcy or breakup, which would have led to the dismemberment of the labs.
During much of the 1990s, the most notable thing about IBM's research seemed to be how successful competitors were in exploiting it. IBM pioneered high-speed chip designs based on limited logic pathways, for example, but Hewlett-Packard and Sun Microsystems used them to grab the lead in the multibillion-dollar market for work stations and server computers that use such chips. IBM also pioneered disk storage systems for large computers, but EMC grabbed leadership of that market. And Oracle seized the commercial advantage created by IBM's research on relational databases, which store data in tables for easy access and manipulation.
But with Gerstner's decision to keep the company in one piece and to emphasize innovation, IBM Research helped drive the company's rebound. A string of valuable inventions helped IBM to become the world's biggest supplier of custom microchips, to upgrade its hardware and to become a leading provider of Internet products and services.
Perhaps most important, IBM Research has been reinventing its own management systems and tools, to encourage researchers to focus on work that can translate into products -- and profits.
Compensation for the research unit's managers is partly linked to the performance of IBM businesses they work with, and to the grades they receive from operating managers for the quality of their relationship. It is a two-way street: operating managers are also assessed by the research unit.
And regardless of the assessment, research managers are penalized financially when the business unit whose relationship they manage fails to meet IBM's performance goals. "We're trying to create a no-excuses culture," Horn said.
Spending
It is a well-financed culture. In its annual report, IBM said it spent US$5.2 billion on research and development last year. Although accounting for such spending is fraught with wrinkles that make comparisons difficult, that probably put IBM in third place nationally, behind Ford Motor and General Motors. Among the other top spenders are Microsoft, the Bell Laboratories of Lucent Technologies, Motorola and Pfizer.
Outside the drug industry, though, pure research accounts for a small part of overall research and development budgets. Most of the spending described as research and development actually covers work by engineers, programmers and technicians who are improving existing products. IBM's central research budget, which excludes such development work, is about US$600 million this year. That is more than one-third larger than that of international technology leaders like Siemens and NEC, and more than twice that of General Electric and Bell Laboratories, the two corporate research giants whose achievements have been most often compared with those of IBM, according to Robert Buderi, a research expert who has visited all of the companies.
The budget for IBM Research, which rose about 10 percent last year, has been growing faster than spending on development and the company's overall revenue, which grew just 1 percent, to US$88.4 billion, in 2000.
IBM's profit from licensing its inventions has been growing even faster. The US$1.7 billion it received last year was more than triple the approximately US$500 million brought in by Texas Instruments, the information technology company best known for aggressive licensing before IBM began pursuing a similar strategy in the 1990s.
"We cross-license everyone in information technology," said Horn, referring to the common practice of trading intellectual property rights. "But with every major company, the money always flows toward us."
IBM's research budget reflects its mission. While about 30 percent of the US$600 million budget at IBM. Research is a basic corporate grant to support long-range work like Eigler's, the remainder comes directly from other units within the company or from corporate matching grants linked to those units' investments. That pushes the researchers to ensure that their work relates to specific products or problems facing the business units.
"They've done an excellent job of creating a financial model where their labor is directly recognized," said Dennis Roberson, chief technology officer at Motorola, when asked what he admired most about IBM Research. Roberson's resume includes 17 years at IBM, along with time at NCR, AT&T and Digital Equipment.
A small part of IBM's work focuses on distant horizons that might never be reached, like the possibility of designing computers from arrays of individual atoms. Such work was highlighted last month, when IBM said it had built a logic circuit from tiny tubes of carbon molecules, or nanotubes.
Much more effort goes into broad programs -- like the US$100 million initiative called Blue Gene to build a radically new form of supercomputer -- that span many areas of hardware and software research. The broadest project, and perhaps IBM's highest priority, is work on autonomous computing, a field encompassing the challenge of creating networks of computers and other devices that can organize and fix themselves without human intervention.
Applied science
There are also many shorter-term projects, like research on the variables in the structure of auctions. Much of this type of work is aimed at rapidly improving IBM products and services in new businesses like electronic commerce.
Sometimes, IBM Researchers pursue serious questions in quirky ways, like building electronic jewelry to force themselves to figure out better ways of interacting with digital devices. "If you have an earring cell phone, you can't dial a number," said Daniel Russell, who leads a group focused on how people use computers. IBM has no plans to sell digital earrings, but they help it to study voice-recognition technology.
There is no simple method for wringing profits out of research, but a supporting corporate culture is clearly a prerequisite. Dr. Eigler provides one window into how IBM's has evolved. He is one of just 25 IBM. Research scientists who have been named as fellows, an honor that brings financial bonuses and more research freedom than other scientists in the group are given. But for all the respect he has garnered, he confesses envy for colleagues like Stuart Parkin, whose research into the magnetic properties of very thin layers of metals has led directly to advances in data storage that generate billions of dollars annually in sales.
Eigler said that as a graduate student two decades ago at the University of California at San Diego, and in his early years working at Bell Laboratories, his goal was to do great science. "Now, in thinking about my career, I'm hoping my work moves into product development," he said. "I've seen others get really turned on by it, and I think I'd react that way, too."
The researchers' acquired taste for applied science has paid off. A recent study for Technology Review magazine by CHI Research, a company in Haddon Heights, New Jersey, that tracks corporate investment in research and development, concluded that IBM's "technological strength" was more than twice that of any other major company. The study was based on the number of patents filed -- IBM is on pace to match last year's record total by the end of this month -- and on how often those patents are cited by other inventors. That is a rough measure of a patent's effect on innovation.
From 1975 to last year, CHI said, some 416 of IBM's patents were among the elite group of the 10,000 most often cited, more than twice as many as Bell Laboratories, at No. 2. A closer look at recent patents produces similar results, according to CHI.
"IBM's patented technology is as good or better in the last five years as it has ever been," said Anthony Breitzman, chief technology officer at CHI.
To some degree, such results reflect the sheer size of IBM's research effort. Among IBM Research employees, in eight major laboratories in the US, Asia, Europe and the Middle East, are 1,400 Ph.D.'s. But, based on how often their patents are cited by inventors, many smaller companies like Micron Technology may be getting more bang for their research buck. Cisco Systems is particularly well-known for a patent portfolio built by acquiring innovative start-ups; patents now owned by Cisco have been cited at a much higher rate than IBM's in recent years.
Rivals
Some large rivals also grumble that IBM's current reputation as king of the research hill stems in part from its flair for public relations. Motorola researchers, for example, were further along than IBM in figuring out how to use copper wire technology to increase the performance of semiconductors when IBM made a splashy claim that it had mastered that technology, according to Roberson, the Motorola research chief. Motorola applies the technology at least as broadly as IBM today, but IBM continues to be credited as the pioneer and leader, he said.
Motorola researchers complain that the same perceptions appear to be forming in the new field of magnetic random access memory research. Both companies are racing to develop high-capacity recording materials that do not lose data when a computer is turned off.
Although it supports research in areas like electronic plastics, IBM has played a negligible role so far in the medical sector and the biotechnology breakthroughs of recent years. Even in information technology, it must often share the spotlight. Just last week, for instance, Motorola announced a new method to combine previously incompatible semiconductor materials that is expected to cut costs and bolster performance for many electronic products.
Here in Silicon Valley, Hewlett-Packard has garnered headlines for ingenious experiments that could become the building blocks for designing computers and memory devices millions of times smaller than today's devices. Such results reflect Hewlett's stepped-up investment in basic research in recent years, a push that has been accompanied by a vigorous effort to file more patents and accelerate product innovation.
Acquisition
Hewlett's proposed acquisition of Compaq Computer, announced last week, would add considerably to Hewlett's product development resources. But Compaq has not been known as a leader in research -- most of its work has been focused narrowly on large-scale computing systems, the Internet and the proliferation of personal electronic devices beyond the PC.
In the world of material sciences, IBM's main rival in many segments remains Bell Laboratories, which was regarded as the clear world leader during the days before the breakup of its original parent, AT&T. Bell, which was split again when AT&T spun off Lucent in 1996 and will lose another group of researchers when Lucent completes its pending divestiture of Agere Systems, is still a pacesetter in fields like optics and communications. And when it comes to growth in research investment, Microsoft, not IBM, is grabbing headlines.
Ever since the founding of central corporate research groups a century ago by pioneers like General Electric and DuPont, companies have struggled to strike the right balance between research that is likely to provide short-term company benefits and pursuing technology that could radically change the competitive landscape down the road. The toughest calls have been whether to support basic research into mysteries of nature, like the search for subatomic particles known as neutrinos, where it takes a vivid imagination to see any connection to the business.
The challenge has only intensified today, as technologies change rapidly and divisions of companies routinely face off against rivals that are also customers, suppliers or partners to other parts of their company. Missed opportunities in research strategy show up with the regularity of fireflies on a hot summer night.
At IBM, the responsibility for heading off such problems rests most heavily on Horn, who began his research career at the company by using statistics to describe what happens on the surfaces of materials as they are altered or bonded to other materials.
Horn figures that his biggest challenge is pushing more of IBM's research effort into serving the company's huge global services unit. Global services, which runs computer operations for other companies, designs and installs computer networks and software and provides consulting, recently became IBM's largest single business.
IBM Research has developed specific products for the services businesses, like a software package used by IBM consultants to help clients understand how e-commerce can reshape their businesses. And IBM has been able to market access to its researchers as a special benefit of entering into close relationships with its services unit. Air Canada, for example, recently awarded IBM a US$900 million contract to manage computer operations for seven years, in part because IBM also assigned researchers to help the airline develop new check-in kiosks and an electronic maintenance manual that Air Canada hopes to sell eventually to other airlines.
Efforts to link researchers to business operations are not new or exclusive to IBM. As far back as 1956, the IBM unit that sold computer systems to the federal government set up a jointly financed group with IBM. Research to develop new products. But the company's effort to create and nurture new links has become much more aggressive in the last decade.
Other projects
Besides pegging compensation to cooperation and making research executives responsible for specific relationships, 12 major IBM programs are jointly managed by executives from IBM. Research and other IBM units like the Software Group E-Commerce program.
It also has 16 other research-intensive projects that are not jointly managed but are at least partially outside Horn's domain. For instance, the Life Sciences unit, an effort to make IBM a bigger name in the increasingly computerized world of medicine, has many researchers on its staff but reports to John M. Thompson, vice chairman of IBM, who oversees emerging businesses, rather than Horn.
"IBM is better than any other research organization in the way it keeps evolving its model of how to connect to the company," said Buderi, author of Engines of Tomorrow: How the World's Best Companies Are Using Their Research Labs to Win the Future.
FIVE-YEAR WINDOW? A defense institute CEO said a timeline for a potential Chinese invasion was based on expected ‘tough measures’ when Xi Jinping seeks a new term Most Taiwanese are willing to defend the nation against a Chinese attack, but the majority believe Beijing is unlikely to invade within the next five years, a poll showed yesterday. The poll carried out last month was commissioned by the Institute for National Defense and Security Research, a Taipei-based think tank, and released ahead of Double Ten National Day today, when President William Lai (賴清德) is to deliver a speech. China maintains a near-daily military presence around Taiwan and has held three rounds of war games in the past two years. CIA Director William Burns last year said that Chinese President Xi Jinping
President William Lai (賴清德) yesterday said that China has “no right to represent Taiwan,” but stressed that the nation was willing to work with Beijing on issues of mutual interest. “The Republic of China has already put down roots in Taiwan, Penghu, Kinmen and Matsu,” Lai said in his first Double Ten National Day address outside the Presidential Office Building in Taipei. “And the Republic of China and the People’s Republic of China [PRC] are not subordinate to each other.” “The People’s Republic of China has no right to represent Taiwan,” he said at the event marking the 113th National Day of
SPEECH IMPEDIMENT? The state department said that using routine celebrations or public remarks as a pretext for provocation would undermine peace and stability Beijing’s expected use of President William Lai’s (賴清德) Double Ten National Day speech today as a pretext for provocative measures would undermine peace and stability, the US Department of State said on Tuesday. Taiwanese officials have said that China is likely to launch military drills near Taiwan in response to Lai’s speech as a pretext to pressure the nation to accept its sovereignty claims. A state department spokesperson said it could not speculate on what China would or would not do. “However, it is worth emphasizing that using routine annual celebrations or public remarks as a pretext or excuse for provocative or coercive
CONCERNS: Allowing the government, political parties or the military to own up to 10 percent of a large media firm is a risk Taiwan cannot afford to take, a lawyer said A Chinese Nationalist Party (KMT) legislator has proposed amendments to allow the government, political parties and the military to indirectly invest in broadcast media, prompting concerns of potential political interference. Under Article 1 of the Satellite Broadcasting Act (衛星廣播電視法), the government and political parties — as well as foundations established with their endowments, and those commissioned by them — cannot directly or indirectly invest in satellite broadcasting businesses. A similar regulation is in the Cable Radio and Television Act (有線廣播電視法). “The purpose of banning the government, political parties and the military from investing in the media is to prevent them from interfering
RSS