Universities are important sources of many of the new ideas in science and technology that contribute to innovation in the United States. By producing new knowledge and exposing students to that knowledge, they not only generate new ideas but prepare knowledgeable, inventive, and motivated graduates who can carry those ideas into businesses, nongovernmental organizations, and governments. In addition, faculty members sometimes play a direct role by consulting with existing companies or even starting their own companies. A majority of the first workshop was devoted to considering the role of universities in the innovation ecosystem. Some of the common themes are cited in Box 4-1.
The Role of the University
• The culture of a university influences its success in producing innovation and can influence the culture of a region.
• A university does not merely prepare young people for the wide variety of roles in an innovative economy; its higher mission is to “discover and invent the future.”
• The technology transfer policies of universities may be in need of rigorous review and assessment.
• Universities might benefit from examining the rules governing the amount of time faculty can devote to outside activity.
• Faculty members and students can have conflicts of interest and conflicts of commitment that need to be understood and properly managed.
THE PREPARATION OF STUDENTS
The most important product of universities, said several workshop participants, is educated students. These students include not just the founders of new companies but the employees and customers of those companies. The first
few hires for a startup are critical, said Hennessy, but the next one hundred or one thousand people a company hires are also important for its success.
Universities should not be farm teams for industry, Hennessy said; they should not be engaged in designing the next product for a company’s line. Instead, they should seek the discontinuous innovations that create or transform an industry while teaching their students how to think critically and creatively. The job of a university is to “discover and invent the future,” Hennessy said—in part through research, in part through education, and in part through active efforts to move university-derived ideas into industry.
Investments by governments at all levels are essential if universities are to fulfill these missions. Eli Yablonovitch, director of the NSF Center for Energy Efficient Electronics Science at Berkeley, once calculated how much the government had invested in his education and concluded that the total was close to a million dollars. Furthermore, even after graduating, he had access to good jobs and good organizations where he was able to develop his expertise. “The human capital aspect is gigantic,” he said.
THE TRANSFER OF TECHNOLOGY TO INDUSTRY
Many universities have created offices to foster the transfer and licensing of technologies developed by faculty members and students to industry. These offices can establish beneficial relationships with individuals and organizations outside universities. As Borrus said, “If you build a relationship there, it’s like any other walk of life, you can get things done.”
Several speakers at the workshop also discussed problems with these offices and ways in which they could function more effectively. Mowery pointed out that technology transfer offices often have competing mandates. The university president sets one set of goals, licensing officers are evaluated on a different set of criteria, and state legislators mandate yet another set of objectives. Many participants pointed out that this reflects the fact that universities have multiple objectives that they want to achieve through technology licensing, including regional development, revenue generation, and recruiting and retaining faculty (for example, by supporting faculty spinoff companies), and these objectives are rarely prioritized. At the same time, technology licensing offices often are evaluated solely by the amount of revenue they generate, which creates the wrong incentives and can disrupt the academic environment and culture, thus suppressing interactions with industry rather than fostering them.
The revenue generated from technology licensing tends to be featured in publications from a university and talks by administrators, but almost all the data on licensing revenues consist of gross revenues. The net returns are much smaller after operating expenses and other liabilities are subtracted, said Mowery. With a few exceptions, such as Stanford, the net licensing revenues are much smaller than the amount of industry-sponsored research at a university.
This creates a tension between using licensing policy to maximize revenue and using licensing policy to develop broader relationships with industry.
Mowery pointed toward “an astonishing lack of experimentation” by universities of different ways of organizing their relationships with industry. As an example of such experimentation, he pointed to the University of California system, where faculty members have the option of not working with the technology licensing office. Instead of patenting an invention, they can put it in the public domain and seek benefits from exchanges with people who want to discuss an innovation and make a product based on that idea. This does not work in all fields of technology, and administrators may lament potential losses of revenue, but in some areas it is a valuable alternative to technology licensing.
Many universities have tried to emulate the licensing successes of Stanford when in fact they are very different kinds of institutions, said Mowery. The heterogeneity of U.S. higher education has always been one of its strengths and creates opportunities for experimentation. In addition, collaboration among different types of colleges and universities offers the potential to increase benefits and reduce costs.
Even at universities that have had great successes with technology licensing, the returns can be deceptive. The current chairman of COSEPUP, Richard Zare, the Marguerite Blake Wilbur Professor in Natural Science at Stanford, said that he had looked at the greatest returns to Stanford from licensing. The greatest single return came from patents that Stanley Cohen and Herbert Boyer were granted on the recombinant DNA techniques they pioneered. The second biggest return came from the School of Music for the invention of technologies used in the Yamaha synthesizer. The third largest return was for licensing of the Stanford logo.
Zare insisted that Stanford graduates return much more to the university than does its licensing. “It’s not Hewlett-Packard but Hewlett and Packard individually and their families who have given money back to Stanford.” As Hennessy said, the philanthropy from Hewlett and Packard to Stanford dwarfed any licensing fee that could have been charged for the discoveries they used to start their company.
Some workshop participants made the case that technology licensing offices can be counterproductive if they cause negotiations to be so complex that agreements are scuttled. They also pointed to the negative consequences of a belief widespread among state legislators that licensing offices could produce abundant resources for universities, which then could displace government funding. Even senior university officials and technology licensing officers sometimes have a “somewhat naïve view” that all areas of technology have the potential to create successes that in the past have occurred in just a few particular circumstances. In fact, the objectives of technology licensing offices are much broader. These offices can forge connections with industry, help faculty members move their ideas beyond the walls of the university, and support regional and national economies.
Participants also mentioned other ways some universities accelerate the transfer of technology to industry where commercially promising ideas or technologies can be developed, either within academic or commercial settings, such as incubators, accelerators, or proof-of-concept centers. These institutions can have different characteristics depending on the technology they are fostering. For example, the biotechnology sector requires expensive access to laboratories, equipment, and staff, while the information technology sector may require much smaller investments. Hennessy described one approach Stanford has pursued to support student-driven innovations for three to six months to see if they can be developed to the point where they would attract commercial interest or the interest of a venture capital firm or angel investor. It remains an interesting educational experience for the students with a very minimal investment, Hennessy observed. For example, he mentioned a group of students who were working on a way for students to pay each other money using their phones. The students got money to spend three months working on the project, and they lived in a house together and spent 12 hours a day working to see what the technology could do. It cost $5,000 to support the students, and they had an educational experience that resulted in a potential business opportunity.
The biggest problem with incubators, according to Hennessy, is that they need to be shut down if they are not producing results, and shutting them down can be difficult. “You give them a hard deadline and say, go out and get funding by this deadline or you are out.” The history of incubators at universities is not promising, because they often require that university staff members shut down projects led by faculty members. “You could do it and make it work, but it would take tough love,” said Hennessy. One possibility suggested by Zare would be to have the industrial part of an incubator report directly to the administration of a university and be judged over a five-year time frame rather than on a year-to-year basis. Another possibility, suggested at the second workshop on research parks, would be for an industrial park to be part of a university-government entity that could take equity in early start-up companies.
THE ROLES OF FACULTY MEMBERS
Faculty members transfer technology directly to industry when they help start new companies or consult with existing companies.5 For example, Yablonovitch described the four companies he helped start beginning in 2000. Yablonovitch remained a professor at Berkeley even as he was forming his companies. He said that his job was to conceive of a valid business idea and then hire the technical team. The team ran the company, while he served on the
5 It should be noted that, although the participants did not discuss it explicitly, throughout their discussions they implicitly acknowledged that the freedom given to individual faculty members to explore, coupled with a culture in which failure is at least occasionally acceptable, is a striking component of the U.S. research and innovation system.
boards of the company. This approach has worked well in the past, he said, though it may not be the right model for all sectors.
Yablonovitch said that he has always favored business ideas that have a strong scientific component. This can be problematic, he admitted, in that society wants its needs to be fulfilled. Researchers, in contrast, are often interested in ideas that are different, elegant, or clever, but “society didn’t ask for that. There’s a bit of conflict there, and sometimes you end up being too far ahead or not focusing on basics, where there is money to be made just doing the basics.”
His first company, Ethertronics, was established to pursue new designs for cell phone antennas. Initially the company secured a patent on a promising design, but the patent turned out to be less useful than anticipated. For four years the company worked on new ideas with few returns until a “deeper understanding of Maxwell’s equations” led to a radically new antenna concept that increased efficiencies from approximately 33 percent to 50 percent. At the time of the workshop, the company had shipped 700 million antennas, most of which are still in operating phones, meaning that about one-tenth of humanity is using the technology. But the company has not been as profitable as might be expected, because it has relatively few customers, and the customers are able to dictate the price they will pay for the antennas. “It’s certainly a technical success and has impacted society, but it hasn’t made that much money.”
Luxtera, which is a pioneering company in silicon photonics, originated when a venture capitalist came to Yablonovitch and expressed interest in starting a photonics company. Though Yablonovitch was worried that the technology was not yet developed enough for commercialization, he started the company and began developing a silicon chip that included optical components as well as semiconductor components. The initial product was a cable that converts electronic signals to photonic signals, which travel along a fiber optic strand until they are converted back into electronic signals on the other end of the cable. The cable is much faster than a USB cable and has attracted the attention of people who run supercomputers and large data centers.
A major issue in this case, said Yablonovitch, is that many companies have become involved in this area, and each has taken a slightly different approach to the technology. At this point it is difficult to say which approach is best. Luxtera therefore could become known as the company that pioneered the technology, but another company might end up making more money from the technology.
A third company, developed with a mathematician at the University of California, Los Angeles, creates the patterns for photolithography to manufacture sub-wavelength features on silicon wafers. These patterns are not at all intuitive, because when light shines through them it creates a quite different pattern on the substrate. The business model was therefore to provide the patterns to semiconductor manufacturers that wanted to build semiconductor chips. Furthermore, because the patterns are not unique, as typically happens
with mathematical problems involving inverse transformations, the patterns can be engineered to be insensitive to errors in depth of field, easier to manufacture, and so on.
The technology was successful, and it ended up being particularly useful for chips that contain repetitive elements, such as memory chips. However, this company also had relatively few customers, which meant that it had relatively little power to set prices. Eventually, the part of the company making the inverse patterns was sold to another company.
Finally, Yablonovitch’s fourth company, AltaDevices, has broken the world record for solar-cell efficiency, taking it from 25.1 percent to 28.8 percent efficiency. The technology is based on the idea that a solar cell achieves its highest efficiency when it emits a small amount of light, which was “very counterintuitive,” said Yablonovitch. Today the cells are being used in applications like space satellites, and an inexpensive way of producing the cells could lead to a much broader range of applications.
However, solar cell technologies also have run into problems because of the large government subsidies that other countries, and especially China, have devoted to this area. Interest-free loans, production subsidies, and other governmental investments have driven the price of conventional solar cells down to a level where they cannot be profitably manufactured without subsidies. As a result, said Yablonovitch, “all the competing technologies are losing their shirts. . . . No matter how good your technology is, you can’t deal with a heavily subsidized competitor.” In addition, ill-advised overinvestment by U.S. venture capitalists in the first decade of the century have scared current investors away from the field.
Today, China is overproducing the world’s needs for solar cells by a factor of two, said Yablonovitch, and the subsidies appear likely to continue for years. “Even if you have good technology and you are successful technically, the market conditions have changed.” Other countries seem to be willing to make investments in the field even though they know they will lose money. But they hope eventually to gain a powerful market position. As a result, in the future innovators may need to go to Asia to secure investment funds, just as they once came to the venture capitalists clustered around Stanford to seek funding. And will the federal government continue to invest money in research if the companies based on that research end up going offshore, Yablonovitch asked. “These are very vexing issues for which I don’t have answers.”
Traditionally, faculty members have been allowed one day a week for work on outside projects. Workshop participants discussed whether these guidelines should be modified to give university innovators more time to work on commercializing an idea. However, Hennessy warned against 50-50 splits. It would be better for a professor to take a leave from the university to work on an outside project, possibly while spending a limited amount of time maintaining ties with a university. However, leaves typically have to be accompanied by a time limit, since otherwise they can drag on for technologies that are proving difficult or time consuming to develop. If people decide that they need to stay
with a company once a time limit is reached, they can break their ties with the university, and if they are successful, they may be able to rejoin the faculty later. “I’m fairly flexible with respect to that,” said Hennessy.
Mowery said that he is skeptical about efforts to use institutional resources to turn faculty into entrepreneurs. Faculty should certainty not be discouraged from engaging in innovation activities and should be supported when they do, but evaluating faculty on the basis of these activities is not a good idea, he said. An evaluation based on patents obtained, for example, would certainty generate more patenting, but “patenting is not necessarily a way of either supporting technology transfer or of reducing the operating expenses of a technology transfer office.” Many faculty members would not be good industrial managers. In fact, access to outside managerial talent is an important contribution that venture capitalists bring to the interface. “Trying to fit faculty who may or may not be square into square boxes is very much to be avoided.”