CEOs are learning how to better tap university R&D. The results could be powerful.
June 1 2005 by Chief Executive
Luis Mejia of Stanford University recalls the fall day nine years ago when Larry Page, then a Ph.D. student, came to Mejia’s office of technology licensing to patent his search engine software, called PageRank.
The patented software had been available for licensing for almost two years, but no company pursued it, says Mejia, who matches Stanford research that could be commercialized with companies seeking ideas. So without any suitors, Page and his partner Sergey Brin founded a small upstart company named Google based on their search engine software, and the company’s IPO last year not only made Stanford $190 million but was also the biggest event in the tech world since the dot-com boom.
|Pointers for CEOs|
The story speaks volumes about the state of university-corporate collaboration when it comes to developing tomorrow’s innovations. For decades, corporate R&D leaders and university researchers have looked at each other across a broad chasm: Companies have wanted access to research they can commercialize quickly, often within a fiscal quarter, and they’ve had scant patience for the academic pursuit of the frontiers of knowledge, or basic research. The universities, meanwhile, have been suspicious of corporate motives, often believing that company researchers only have dollar signs in their eyes. The end result has been that only a tiny fraction of university innovation, an estimated 6 percent, has actually reached the marketplace.
With potentially dozens of undiscovered Googles lurking within the Ivied halls, key decision-makers on both sides are beginning to realize they need to narrow the gulf between them. Universities are worried that federal funding is being reduced. And CEOs of the likes of Intel, HP and IBM believe that unlocking university research is one of the keys to starting a new wave of U.S. innovation.
As a result, a pattern of much more intimate relationships is unfolding between top research universities and major technology companies. “Universities are a tremendous source of research,” says Patrick Scaglia, vice president and a director at the Palo Alto, Calif.-based HP Labs, the central research group at Hewlett-Packard. “But if you look at it as a way to develop a product for the next quarter, it won’t work that way.” Tapping universities for research requires a long-term view, he notes, and often up to a decade or more of diligent R&D with no guaranteed payback and few home runs.
While there is no single formula for working with universities to tap research, HP is among those leading firms that have crafted an approach, using sponsorship of grants, fellowships, research projects and “pie-in-the-sky” theoretical research to tip the research curve.
IBM, one of the top corporate innovators in the country, taps similar sponsorship, but also uses its Ivory Tower connections to capture the best and brightest ideas and students, says Robin Williams, associate director at the San Jose-based IBM Almaden Research Center, one of eight IBM research facilities worldwide.
Intel goes so far as to place one of its four research “lablets” at the University of California Berkeley campus, where students, professors and Intel researchers work side by side on inventions that will make tomorrow better, says Joe Hellerstein, director of Intel Research Berkeley.
The reason these more proactive approaches are needed is that the U.S. is losing its competitive edge worldwide as funding for research decreases and basic research gives way to more practical, project-oriented work, says Johns Hopkins University president William Brody. He notes that U.S. federal research and development spending peaked in 1965, at just under 2 percent of GDP, and has dropped to 0.8 percent today. He adds that Asia now produces three times the number of scientists and engineers as the U.S., and Europe more than twice that number. And he notes that the U.S. edge in high-tech exports is slipping, from 31 percent of global high-tech exports in 1980 to 18 percent today while Asian markets, excluding Japan, have climbed from 7 percent to 25 percent of high-tech exports over the same time frame.
“I’m worried that the U.S. is moving away from risk-taking,” says Brody, recalling that America’s history is built upon discoveries by pioneers Henry Ford, John D. Rockefeller and Bill Gates. “Innovation leads to productivity gains, [which] leads to GDP growth,” he says. The current lack of risk-taking and innovation will “eventually drive America into second-class status,” argues Brody, who is co-chair of the Council on Competitiveness’s National Innovation Initiative.
The harsh warning is echoed by Deborah Wince-Smith, president of the Council on Competitiveness. “There is a direct correlation between productivity and innovation and this is a first-tier economic priority for the country,” she says. Noting a consistent decline in commercial and federal spending for innovative and risk-taking research, she believes that universities can help fill the gap. “Universities are an anchor for innovation and can be regional hot spots with profound impact on the ecosystem of an entrepreneurial culture,” she says.
The National Science Foundation (NSF) reports that R&D spending in the U.S. rose only marginally, or about 1 percent, to $283 billion from 2000 to 2003. Of that number, $54.1 billion, or 19.1 percent, was spent on basic research, the type that involves a high degree of uncertainty in terms of technical success and commercial value, but which can lead to revolutionary breakthroughs. Some 55 percent of all basic research in this country is done by universities and colleges, according to NSF.
Much of that basic research is underwritten by the federal government, which, in 2003, supported 60.5 percent of nationwide basic research and 63.5 percent of basic research done by universities and colleges. (The federal government also funds 30 percent of the country’s total research spending, much of it directed toward development of tactical nuclear weapons and space exploration.)
By comparison, industries devoted only 5 percent of their R&D budgets to basic research, accounting for 17 percent of total basic research spending nationwide. But in applied research development, industries accounted for 62.6 percent of the $67.8 billion spent in 2003, according to NSF.
“High Impact, High Risk” Research
For corporations, innovation results from many avenues€¦quot;acquisitions of start-up companies with revolutionary products, internal research and development, corporate venture funding of enterprises with related technologies and sometimes even by chance (the so-called Eureka moment). Labs within the Ivy walls are an important source of farther-reaching research, and indeed such key breakthroughs as the sequencing of the DNA gene originated in academic labs.
HP looks at university R&D as a way to do “high impact, high risk research,” says Scaglia. If a company works in a collaborative way with universities and other researchers, tremendous breakthroughs can result. “Every corporation has to have true invention and you want to be there when the next generation of technology comes, when the disruption comes,” Scaglia adds.
HP reserves about $200 million of its $4 billion annual research expenditure for what Scaglia calls “advanced research,” with a small percentage of that $200 million going toward university research. “What we are always striving to do is to get the right balance between developing the next generation of technology concurrently with adding value to the HP business,” says Scaglia, who points to the inkjet printer as one example of a successful product that originated in his advanced research group. “While we strive for impact, there is always the danger that we become too independent and too academic. The trick is balance.”
He cites the Itanium chip for PC servers, introduced in 2001 in a collaborative project with Intel, as an example of a successful technology that originated in the 1980s in academia and resulted in a new concept for PC architecture. HP today holds a 60 percent market share of Itanium-based servers.
Scaglia says the best university research is developed through open source or standards that can be used by anyone€¦quot;an approach advocated by leading science and engineering schools and a strong contrast to the patented and carefully protected research coming out of corporate internal labs. He argues that open source is the right approach because innovations do not happen in a vacuum but result from collaboration. He notes that the Internet, the Unix operating system and the first web browser all stemmed from open source research, mostly done at universities. “If you are working in the public domain, you can measure your technology with peer review,” he says.
Continuing to strengthen its ties to universities, HP funds research at “close to 100″ universities around the world, including some in Japan, China and South Korea.
Scaglia says he has “no idea” what they might develop. “We don’t think of it that way,” he says. Rather, the access gives HP a firsthand glimpse of, and a role in developing, tomorrow’s leading-edge technology. HP places many of its labs near such leading engineering and technology schools as Stanford University and Massachusetts Institute of Technology.
Intel, which invests $4.4 billion in R&D at corporate labs in the U.S., Russia, China, Korea and India, gets up close with academia by placing a network of four research labs at university campuses, where faculty, graduate students and Intel researchers work together on exploratory research. In addition to Berkeley, these lablets are at the University of Washington in Seattle, Cambridge University in Massachusetts and Carnegie Mellon University in Pittsburgh. Berkeley lab director Hellerstein says Intel has 200 people working on “off-the-road research.”
The universities and Intel have agreed to develop the bulk of the research coming from the labs without proprietary IP, says Hellerstein. “This gives our university Ph.D.s a way to collaborate with Intel labs without having to worry about not getting access to confidential research from other parties due to intellectual property protection,” he says, explaining that most software and hardware designs are developed today through open source.
The lab director’s position is rotated every two years among Berkeley engineering faculty, and the faculty member-director brings in four to eight graduate students to work with some 15 to 20 Intel researchers on projects. A large university project can take four to six years to develop, and then will be in testing and advanced research for another three to four years, Hellerstein says.
Among the research projects in development at the Berkeley lab are wireless sensors networks, which can be used to monitor the vibrations on manufacturing equipment to gauge their condition. British Petroleum is testing the sensors on a vessel in the North Sea, using it to measure the longevity of the super-tanker, a move that could help the company lower its insurance premiums.
Another project in the works is the use of scientific monitoring devices called “macroscopes” which, among other things, measure tree respiration. “We have hundreds of these sensors in Redwood trees and they give us an ecosystem reading of the trees,” he says, adding that this basic research could lead to a breakthrough in “precision agriculture.”
Hellerstein says that the number of products that have been commercialized would “fit into your hand.” Some of the more farsighted research and technology, such as the sensors, is “just popping out now” from the lablets, he says, which were set up in 2001.
Other research from the lab may not materialize for at least another decade. One other-worldly project is Planet Lab, which has the support of Berkeley, HP, Intel, Princeton University, the University of Washington and more than 60 universities. Designed to be a global test bed for inventing and testing prototype Internet applications and services, the Planet Lab researchers aim to spark a new era of innovation by using “overlay” networks to upgrade and expand the Internet’s features and capabilities.
At IBM, which spends some $5 billion annually on R&D, university research is supported in numerous ways€¦quot;through student fellowships, grants to students and faculty, and donations of equipment for faculty-led research groups doing work that is strategic to the corporation. IBM also has faculty working part time at its Almaden Research Center.
Some IBMers also teach part-time at campuses. Such close links give IBM an edge in recruiting the brightest Ph.D.s from the top engineering and computer science departments, says Williams at the IBM Almaden research labs.
With a research division of nearly 3,000 people in eight labs around the world and a record for the most patents in the U.S.€¦quot;3,277 last year, according to IFI Claims Patent Services€¦quot;IBM doesn’t need to depend upon universities for research, he says. But the links to academia help with “general progress in research,” which keeps IBM linked in to basic research developments that originate in its own walls.
Two research programs IBM sponsors are CITRIS with Berkeley (see sidebar, page 43) and a nanotechnology project with Stanford University, called Spintronics, where about 30 professors, graduate students and IBM scientists are working on ways to magnetize electrons to improve the performance of computers. That could be as important as the transistor 50 years ago. “Invention is no longer enough,” says Williams. “One has to add business insight and make it useful and bring it to market. This is innovation.” And the U.S. needs more of that kind of innovation if it is to remain ahead of ambitious challengers around the world.
West Coast innovators bet on a university collective
It’s Cal Day at the University of California Berkeley, and A. Richard Newton, dean of the College of Engineering, is standing in front of a welcome desk, greeting prospective students at the open house. The mid-April weather is glorious, as is the view overlooking the aquamarine San Francisco Bay. Newton, sporting a bright smile and a Hawaiian-style shirt plastered with palm trees and school logos, shouldn’t face too many obstacles in recruiting the best talent to the campus on a day such as this.
Under Newton’s leadership, Berkeley, one of the world’s top engineering schools, has recently broken ground on new headquarters for a six-year-old multi-disciplinarian research institute called the Center for Information Technology Research in the Interest of Society, or CITRIS. The mission of the center is “to tackle six grand challenges through technology,” says Newton, a cofounder of two major semiconductor companies, Cadence Design Systems and Synopsys. The six areas are energy, the environment, transportation, natural disasters, education and health care.
Among the corporate sponsors that have lined up to support CITRIS are Intel, Microsoft, Sun Microsystems, IBM, Ericsson, Hewlett-Packard, Nortel Networks and Infineon Technologies, plus 11 associate corporate sponsors such as Cisco, Ford, BT and Siemens. The founding corporate partners have each pledged $1.5 million over four years and get to direct where their dollars go, says Newton. Patrick Scaglia of HP Labs calls CITRIS “the most expansive and ambitious multi-disciplinarian approach of the many universities” working on information technology breakthroughs.
Newton says CITRIS does not typically grant exclusive licenses, preferring to operate in an open source environment to encourage quicker adoption of new developments. “We are not looking to make money on technology licensing€¦quot;what we call kilogram dollars€¦quot;but what we want to do is to create a billion-dollar industry and have impact in society,” he says. On the rare occasion when Berkeley has granted an exclusive license, it’s because the technology had the most impact with that corporate owner and because the company had lined up state and federal funding to maximize R&D from the university.
Though CITRIS research typically has its eye on future decades, the center already has commercialized some products. One is called “smart dust,” which Crossbow Technology licensed to create a tiny wireless sensoring device that can determine a building’s energy efficiency, detect the earthquake response of a retrofitted building and monitor wild fires.
Another campus R&D project has sparked the interest of UTStarcom, a provider of cell phones to third world countries. Berkeley students, through a project called the Technology Peace Corps, have found a cost-effective way of delivering phone service to remote rural areas: by using a balloon instead of a village base station to transmit signals.
UTStarcom division president Jack Mar says the company has its own solution for transmitting signals inexpensively but even so, CEO Hong Liu, an alumnus of Berkeley, is studying whether the company might deploy the balloon after it conducts trials of the technology this summer. “If it works out,” says Mar, “we can apply what they come up with and integrate it into our system.” €¦quot;R.F.
|Tapping the R&D Gold Mine|
While other options exist, the most direct way to gain access to technologies developed on university campuses is to license them.
Some 3,000 universities, including leading business and engineering schools such as MIT, Harvard and Stanford, have an office dedicated to technology licensing. Their job is to facilitate movement of lab inventions to the marketplace. The programs vary, but there are some basic guiding principles.
First and foremost, the university owns the intellectual property to the product or technology developed on campus. The licensing and technology office typically files for a patent after combing through dozens of inventions. At Harvard, the licensing office “receives reports on about 150 inventions each year, but we only file patents on about 40 percent of them,” says Erik Halvorsen, director of business development at Harvard’s Office for Technology and Trademark Licensing. “We license about 65 percent of those, and about one in 10 of those is successful.” Those successes include two large biotech companies in Cambridge, Mass., Millennium Pharmaceuticals and a spin-out group from Genzyme, says Halvorsen. Novartis is also making use of licensing at Harvard, most recently with a treatment for multiple sclerosis.
Most leading research universities also maintain the right to publish news of any technological or scientific advance, typically in an academic or scientific journal. “Corporate managers need to be aware that we are an academic institution and not a business,” says Lita Nelsen, director of the MIT Technology Licensing Office. In other words, universities will not delay publishing breakthrough research so that a company can make a splash with a commercial launch. Nelson has had to explain the competing priorities to irritated corporate representatives. “It takes a while for them to get over it the first time regarding the university policies, what they can do and what they can’t do,” says Nelson, who manages some 170 patents at the 30-employee office. Among her successes: a bladder cancer detection product for Matritech, a hip replacement compound for Zimmer and a Bristol-Myers Squibb medical imaging product called Cardiolite, which has generated more than $20 million in royalties for joint sponsors MIT and Harvard.
When a licensing office shops a new technology, it’s typically offered first to any corporate sponsors involved. If there are no takers, the office looks for other corporate researchers who may want to bring the technology in-house.
The office will then grant exclusive or nonexclusive rights to the technology, with negotiable terms. “We will give an exclusive license if a company commits to developing the technology and has the resources to do it,” says Nelsen.
Harvard’s Halverson says there have been “heated arguments about exclusive and nonexclusive licenses” and about fees. He is currently negotiating a license with a major pharmaceutical company. “We may have to walk away because they want everything,” says Halvorsen.
A company with a nonexclusive license receives access to the technology without having to pay royalties, which can otherwise mount quickly. At Stanford, the licensing office takes 15 percent of royalties to cover overhead. The remaining royalties are divided up evenly between the inventors, the university department where the discovery was made and the School of Engineering, notes Luis Mejia, associate director at Stanford’s office of technology licensing.
Mejia says he’s currently managing some 120 inventions in his portfolio, but he doesn’t expect anything to surpass Google in terms of royalties generated for another 20 or 30 years. The last such major breakthrough was in the 1980s, when two professors came up with recombinant DNA cloning discovery, and that led to the birth of the biotech industry and products for Genentech, Amgen, Eli Lilly, Johnson & Johnson and Schering-Plough. Not so insignificantly, it also generated $255 million in royalties for Stanford€¦quot;money that went back into fueling the next generation of possibly commercial technology.