Big Blue's Global Lab

When Brazilian President Luiz Inácio Lula da Silva met with IBM (IBM) Chief Executive Samuel J. Palmisano in Brasilia on Aug. 5, the former firebrand labor leader and the captain of industry had plenty to talk about. Brazil is planning to invest $22 billion in science and technology innovation in 2010, and it's pressing companies to contribute billions more to the effort. IBM, meanwhile, is prowling the world to set up what it calls "collaboratories," which match up its researchers with experts from governments, universities, and companies. As the two leaders discussed their ambitions, the meeting stretched from its scheduled 30 minutes to nearly an hour. "I hope to see you many more times," da Silva told Palmisano in the end, with a twinkle in his eye, "because it would mean you're announcing many investments in my country."

Any deal between IBM and Brazil is months away, but the company's new research approach is taking off now. IBM is trying to convince countries and companies that it can help them improve their ability to innovate at an important moment for the global economy. "Investments in innovation are critical, especially in a downturn," says Palmisano. "They can help Brazil and other countries, including the U.S., realize an economic expansion."

IBM has hammered out six deals for collaboratories in short order—in Saudi Arabia, Switzerland, China, Ireland, Taiwan, and India. Four more are in the works. John E. Kelly III, director of IBM Research, says there's enough demand for 100 more tieups. "The world is our lab now," says Kelly. "I figure I can have a much larger impact on the company and our research if I operate this way."

The strategy marks a sharp break with how corporations have historically conducted basic research. For decades companies such as IBM, AT&T (T), and Xerox (XRX) treated the work done in their labs as top secret. Research facilities were fortress-like, with special passes for the most sensitive areas. In recent years, companies such as Hewlett-Packard and Intel have begun tapping talent from outside for essential bits of science and technology—a concept called open innovation. Now IBM is moving a giant step further by making collaboration with outsiders an essential piece of its research strategy. The depth of that collaboration, the number of partners, the staff involved, and its global reach set IBM apart. "To move in this direction you have to be willing to not just take risks but be open to accepting ideas from around the world," says Soumitra Dutta, professor of business and technology at Europe's INSEAD.

There's no guarantee IBM can make this approach work on a grand scale. Many of the more limited joint projects have run into problems as companies tussled over expenses and intellectual property rights. But if the approach succeeds for IBM, other companies may follow its lead. "This is a great way to diversify your research portfolio, leverage what you have already, and get new knowledge and inventions," says Karim R. Lakhani, an assistant professor at the Harvard Business School.

Such collaborations are controversial, though. What's good for IBM and its corporate brethren isn't necessarily good for America. Critics say U.S. competitiveness is weakened when flagship corporations do crucial research projects overseas. "I'm alarmed," says Clyde V. Prestowitz Jr., president of the think tank Economic Strategy Institute. He says it's sensible for American multinationals to seek these tieups but urges the U.S. government to intercede and offer incentives to keep the most important research here. "We need a change in mentality. The attitude of Washington has been aggressively hands-off," he says.

STRETCHING DOLLARSThe U.S. may have to act quickly to forestall a rising tide of offshore research. After aggressively moving manufacturing and software programming to other countries, American corporations are setting up research labs overseas as well. But building brick-and-mortar labs and hiring large staffs is expensive, so some, like IBM, are beginning to establish closer relationships with foreign governments and universities. For instance, computer maker Hewlett-Packard (HPQ) formed a joint lab with Tsinghua University in Beijing. Chip giant Intel has set up joint research centers in China and Germany. "We'll have more and more global research," says Andrew A. Chien, vice-president for future technologies research at Intel. "We have an imperative to reach out and tap that power."

The attraction for IBM is clear. The collaborative strategy snags more research with roughly the same amount of IBM money. Performing research with a variety of partners in many locations also exposes IBM to science challenges and ideas that it might not otherwise encounter.

For years, IBM's crown jewel has been its 3,000-scientist research department—with labs in New York, California, Texas, Massachusetts, China, India, Israel, Japan, and Switzerland. Although revenues have declined by double digits this year, the budget for research is holding steady. The payoff: Thanks in large part to the labs, IBM has remained on the cutting edge of supercomputing, chip manufacturing, and datacenter management.

With the collaboratories, IBM hopes to make research an even bigger contributor. Each joint venture is expected to be staffed with 10 to 100 scientists targeting technologies that can deliver results in a relatively short time. The strategy is synchronized with what IBM calls its "Smarter Planet" push, where it offers technologies and services to improve transportation, electrical grids, and other systems.

IBM and other companies pursuing similar strategies face complex hurdles, however. It's not easy to set up and run this kind of lab. First, IBM has to sift through hundreds of potential partnerships to pick the ones that make the best fit. Then the company and its partners have to negotiate contracts that spell out responsibilities and protect each side's interests. One major potential hangup is dealing with intellectual-property ownership. In a typical collaborative research agreement, IBM wants to co-own the intellectual property or have exclusive rights to it, but that's not always acceptable to universities. One potential IBM project in Eastern Europe fell apart last year because the university wanted to control both the intellectual property and research agenda. "That was a showstopper," says Kelly.

The man behind IBM's new research strategy is a tall, sandy-haired 55-year-old with R&D in his DNA. His father worked as a technician at General Electric's (GE) lab in Niskayuna, N.Y., and Kelly would visit regularly as a boy, watching his father work with vacuum tubes and other technologies. Kelly got a PhD in materials engineering at Rensselaer Polytechnic Institute. At IBM, he ran one major research department, in between stints in the chip division.

IBM's collaboratory strategy emerged from Kelly's management of the chip unit during a challenging time. Going it alone in chip manufacturing had become prohibitively expensive, so, starting six years ago, Kelly forged joint development deals with eight corporations, both U.S. and foreign, and helped New York State establish a nanotechnology research facility.

When Palmisano handed Kelly the research job two years ago, he urged him to help further IBM's ambitious global expansion plans. Within weeks, the collaboratory concept was born. "I became convinced you can do real radical collaboration if all the stars align," Kelly says.

GRAND AIMSBy radical, he means a large number of large-scale ventures. While most corporate arrangements with university researchers call for the funding to come from the corporation, IBM aims for a minimum of 50% funding by its partners. Also, the goals must be grand—focusing on areas of research IBM considers crucial to its future. For example, IBM is building and jointly operating a new $70 million semiconductor lab for nanotechnology research with ETH Zurich, a state-funded university in Switzerland. IBM hopes the research will help produce the next semiconductor switch—replacing technologies that have held sway for nearly half a century. By participating in this kind of research, countries have the potential to build new industries, and universities can attract the best faculty and students.

Kelly struck his first partnership with the King Abdullah University of Science and Technology in Saudi Arabia. The school, scheduled to open on Sept. 23, is the country's attempt to create a world-class research university from scratch. It's hiring top scholars from all over the world. "Our goal is to kick-start an innovation-based economy," says Ahmad O. Al-Khowaiter, the university's vice-president for economic development. "We need a couple of success stories, and we think this will lead to one."

KAUST agreed to buy an IBM supercomputer, which is an essential tool in the research projects that IBM and the Saudis are targeting. Among other things, the two teams will collaborate on a study of the nearby Red Sea, which they believe will help improve oil and mineral exploration. "[The computer] is a magnet for smart people, and it makes it possible for us to solve big problems," says Majid F. Al-Ghaslan, KAUST's interim chief information officer.

Now, dozens of deals are in the offing around the world. IBM uses a complex algorithm to identify the most promising situations. The factors considered include everything from available talent to government stability and corruption.

Kelly got an update on the latest possibilities during a daylong meeting of lab directors and department heads on July 15. About 30 people gathered in a conference room and listened to Colin Parris, head of the collaboratories development team, present the update. He showed a map of the globe with dots all over it—except for sub-Saharan Africa and South America. Kelly sat near the front of the room with his legs crossed and his chin resting on his hand. "Look, guys. South America. Nothing yet," he said, gesturing at the map. "You've got to get started."

There's certainly plenty of action elsewhere. A dozen potential deals are under discussion in China alone, and one partnership is already off the ground. IBM began working last October with China Telecom (CHA), the government-controlled communications giant, to apply its data analysis technologies to China Telecom's huge database of subscriber and service information. As a test project, China Telecom wanted to be able to understand customer desires so it could craft improved packages of mobile, landline, and broadband services. Using IBM algorithms, researchers are chomping through billions of service records looking for patterns.

Niu Gang, deputy director of China Telecom Technology Research Institute in Shanghai, says the challenges his company faces are similar to those IBM confronted in the 1980s. "[IBM] successfully transformed itself," he says. "We hope technology innovation will help transform us into a brand new company."

For IBM, the alliance offers a pathway into an important market. China Telecom's service data gives IBM scientists a rare opportunity to try out technologies on massive amounts of real-life data and fashion algorithms optimized for the telecom industry. IBM is already using some findings as it works with other telecoms.

IBM looks for a large payoff from each collaboratory. In the case of the Zurich project, the company did not want to bear the expense of building a new clean room for advanced chip research on its own. A mutual friend of Matthias Kaiserswerth, head of IBM's Zurich lab, and Peter Chen, chief of research at ETH, had learned they were both interested in building new facilities. He introduced them at a gathering at the tony Grand Hotel Quellenhof in Bad Ragaz, Switzerland, in 2007. The two hit it off immediately. "I had never done anything like this before," says Kaiserswerth. "It comes down to whether you trust people and have congruent interests."

Researchers at IBM and professors at the university conducted a series of meetings where they hashed out priorities. They found a 70% overlap—plenty to make an alliance worthwhile. For instance, they targeted an area of physics called spintronics that could lead to the new generation of chips.

Often, IBM gets projects started by dispatching a single scientist with a big idea. Henry Chang, a 22-year veteran of IBM's labs, returned to his native Taiwan last year and helped a professor at National Taiwan University write a proposal for how the country could shift from an electronics manufacturing economy to one with more high-value tech services. The model he used: IBM's own transformation from a computer company to a services and software giant. "The government liked the idea. They wanted to have a conversation about making the transition," says Chang. On Aug. 12, IBM signed a deal with Taiwan's government to improve the national health-care system.

Such projects could fuel concerns about American competitiveness, but two-thirds of IBM's scientists work in the U.S. and there will be collaboratories here, too. IBM is working with Virginia Tech and Arlington County, Va., to develop a lab focused on advanced systems for crisis management.

IBM still faces plenty of challenges in getting its collaboratory strategy to work. But Palmisano put Kelly on notice in a one-on-one meeting in July that he wants him to think even bigger. Kelly isn't exactly sure what that will mean, but he's mulling it over. "The biggest challenge is to not overextend ourselves," he says. "I want to be very aggressive, but I don't want to trip up."

Business Exchange: Read, save, and add content on BW's new Web 2.0 topic networkOpen Innovation OriginsHenry Chesbrough wrote a pioneering book on the concept of "open innovation" in 2003. He examined early experiments at Intel and other companies and argued that the traditional model of doing research in-house is giving way to an approach that integrates more ideas from the outside.For an overview of the book and Chesbrough's latest work as executive director at the Center for Open Innovation, go to

Before it's here, it's on the Bloomberg Terminal.