R&D collaboration between academia and industry has benefits for both parties, yet there are some drawbacks.
As corporate spending on research and development dwindles, enterprises are reaching out to colleges and universities to supplement their R&D. And they often are finding eager partners in those endeavors, as professors and their graduate students look for help, financial and technical, in addressing long-term research projects.
“Pure research is just a luxury no one can afford anymore,” said Jim Hogan, managing partner of Vista Ventures and an investor in Angels by the Sea.
Many of the biggest corporate research organizations in the U.S., such as Bell Laboratories, IBM Research, and Xerox’s Palo Alto Research Center, are reduced in size and scope, Hogan noted. “The squeeze on expense was something that started happening in the 1990s. The reduced expense freed up cash, and if you look all the companies in the top 50 public companies, they rely almost totally on acquisition for innovation,” he said. “Not much real research is going on outside of Google. The public companies just can’t tell their investors that any spending is happening. Google is the exception, mostly because no one cares as long as 97% of their revenue remains advertising.”
Intel is one of the few companies that does more than pay lip service to the value of R&D. The chipmaker last year spent $12.13 billion on R&D, 21.9% of its revenue for 2015, and it has spent more than 20% of its annual revenue on R&D for the past three years, after spending about 15% to 16% in 2006 to 2012, and less than that in 2004 to 2005.
Most enterprises, however, can widen their R&D investments through collaborating with research universities. Hogan said many of the companies he works with as an investor are turning to the big universities for help, such as Carnegie Mellon University, the Massachusetts Institute of Technology, Stanford University, the University of California at Berkeley, and the University of Michigan.
“In the last 10 years, we have moved from government-funded to industry-funded,” said Venky Sundaram, principal research engineer at the Georgia Institute of Technology. “2006 was the last year of the government funding cycle. After that, we struggled to get a program up and running. We finally came up with a consortium model with a group of companies. We still do some work for the Department of Defense, but now we work on technologies that can be commercialized.”
Along with that comes a focus on technologies that are likely to be productized, rather than just research to prove something is possible. In fact, talk to any university research department today and you’re likely to hear that a technology has fallen out of favor because it’s considered too expensive or difficult to manufacture using existing technology. And this happens well before the research proceeds too far.
David Pan, a professor in the Department of Electrical and Computer Engineering at the University of Texas at Austin, said government funding “has not been picking up,” although there are occasional grants from the Defense Advanced Research Projects Agency (DARPA) and the National Science Foundation (NSF). Private industry has made up the difference. Joint R&D is “definitely a good thing,” he said. “I have been working closely with many companies, U.S. and international.” Pan and Nan Sun, an assistant professor at UT, last year received an NSF grant for a research project called “Automation for Synthesizable and Scaling Friendly Analog and Mixed-Signal Circuits.”
Through corporate funding, Pan has sent some of his UT-Austin students to Imec, the R&D organization in Belgium, to work during the current semester.
Greg Hinckley, the president of Mentor Graphics, said his company works with about 25 to 30 universities around the world, spending about $2 million to $2.5 million a year on such collaborations. The universities are in Armenia, Egypt, India, Poland, and Russia, in addition to the U.S.
“We get involved because it’s mutually advantageous for us, both for the university and Mentor,” Hinckley said. Joint R&D with universities has grown “more expansive, involving more universities, more programs, and more money,” he added.
“There is not a lot of centralized direction,” Hinckley said about the R&D programs. Mentor’s 15 business units are responsible for product roadmaps, engineering, product requirements, and outbound marketing, he noted, with each business unit involved in one or more universities. “We’ve gotten more dispersed as the company has increased its size and footprint through the years.”
More than funding is typically involved in joint R&D with universities, according to the Mentor president. “With the universities that we work with, what they’re interested in is getting access to current semiconductor problems. They want to know what are our big issues, in terms of technology, at the front of the progression of the semiconductor technology. They are interested in having access to collections or databases of advanced cases. So, if you’re working in place-and-route, for example, they would like to have access to designs that have been troublesome to have been completed, which are at advanced nodes like 10nm. These are big designs, which it isn’t what they would typically have in their own libraries.”
University researchers also get access to the internals of Mentor tools, Hinckley said.
Mentor isn’t interested in short-term projects, he noted. While the company’s preference is for long-term projects, they can’t be too long, and the problem at issue must be well-defined, he added. Such projects are looking two to five years out.
More universities are paying greater attention to intellectual property rights these days as they seek more funding, Hinckley said. “It gets more difficult. More and more universities have set up centralized IP groups for contracting. It takes longer and longer to reach agreement, and in some cases, you just don’t. We’ve had universities that we would like to work with, where the effort to reach mutual understanding over what the two sides bring and what they’re entitled to. Nothing ever happens. It breaks apart.”
Some IP issues between semiconductor companies and universities have escalated to litigation on patent claims. Marvell Technology Group and Carnegie Mellon University battled for years in the courts over patent-infringement allegations the university brought against Marvell in 2009. The parties reached a legal settlement in February, with the chip company paying a lump sum of $750 million to CMU.
More recently, Harvard University filed suit in June against GlobalFoundries and Micron Technology, claiming the chipmakers infringed upon patented technology developed by Professor Roy G. Gordon and his students for depositing thin films of metals and other materials, specifically covering alkylamide percursors and using those precursors in thin films of high-κ dielectric materials.
“Partnerships between academia and industry have brought tremendous benefit to the world by advancing science and facilitating the development of countless innovations that took shape in university labs,” Harvard said in a statement. “In this situation, Harvard has reached out to each of the named companies outside of the context of litigation and invited them to engage in good-faith licensing discussions. The companies have refused to engage and have, so far, continued their infringement activities.”
GlobalFoundries and Micron declined to comment on the pending litigation, which is being heard in the U.S. District Court for the District of Massachusetts.
In mid-August, the Electronic Frontier Foundation launched the Reclaim Invention initiative, calling on universities to not license or sell their patents to patent assertion entities or patent holding companies. Those firms are commonly known as “patent trolls,” which have no business other than litigation over patent infringement claims. The initiative seeks to have universities take the Public Interest Patent Pledge, part of which reads, “We will strive to ensure that any company we sell or license patents to does not have a history of litigation that resembles patent trolling….Instead, we will partner with those who are actively working to bring new technologies and ideas to market, particularly in the areas of technology that those patents inhabit.”
Universities realized patent-licensing revenue of $2.2 billion in 2014, compared with $130 million in 1991. The number of U.S. patents awarded to universities skyrocketed from 1,307 in 1991 to 5,898 in 2014.
The University Industry Demonstration Partnership (UIDP), based in Washington, D.C., is a project-oriented organization, established in 2003. Among its aims are “The development of a deep understanding and respect of the diverse goals, missions, and cultures between our universities and companies, and appreciation of the synergy that they can afford,” and “Innovation for the public good, maximizing to the greatest extent possible, the information and products that will ultimately be available to the public,” according to its website. Its board voting members include representatives of BASF, Boeing, DuPont Pioneer, MIT, Iowa State University, the University of Southern California, and the University of Virginia.
The Semiconductor Research Corporation has served as a conduit of joint R&D between academia and industry for more than three decades. Mentor Graphics is the only electronic design automation company that is a member of SRC. “We’re very active with the SRC,” Hinckley said, noting that Chairman and CEO Wally Rhines serves on the board of directors. “The principal purpose of SRC is to raise money through dues and other things from the semiconductor industry, match that with government funding, and do directed research,” he said.
The University of Massachusetts at Lowell hosts the Printed Electronics Research Collaborative (PERC), which endeavors to bring together academic, government, and industry funding in the areas of additive manufacturing and printed electronics. The field is similar to semiconductor manufacturing, while also being distinctively different. BAE Systems, Creative Materials, FLEXcon, MicroChem, Raytheon, Rogers Corp., SI2 Technologies, and Triton Systems are the corporate members of the R&D program.
Among the benefits in working with universities, through the SRC and other channels, are the longer-term perspective provided by university researchers, according to Hinckley. While the industry is entrenched in problems two to three years out, universities take the longer view, peering ahead five years into the future.
“We’re very involved with foundries,” Hinckley added, and the process issues that have come up in recent years. “Our customers see it as important that we’re working with universities,” he said. “We get credibility.”
At Mentor, “we don’t do fundamental research,” Hinckley observed. “We’re in the tools business.” One benefit of working with universities is gaining entrée to talented students, who may be hired by Mentor in the future, he said.
Drawbacks in joint R&D with universities
Still, good relationships with universities aren’t always easy to maintain. “It requires great patience,” Hinckley said. “Universities move slowly. They’re run by the academics, not the administration.” The academic decision process is consensus-driven, he noted, which doesn’t lend itself to quick deals.
Mentor Graphics hasn’t had any patent disputes with universities or any other parties. “We haven’t had the trolls chasing us,” Hinckley said. EDA, he noted, is a much smaller business than the semiconductor industry, and less likely to attract litigation by outside parties.
The Mentor president serves on the advisory board of Portland State University’s engineering school, and personally funds some scholarships to PSU. He also sponsors a prize in emulation.
Of collaborative R&D, Hogan said, “University research is never ready for prime time. You have to get the graduate students as well as the IP and surround that around a productization team to have any hope of success. I’m generally a person that has spent a lot of time and treasure on IP and patents. It is, in my opinion, the one key advantage that a startup has. This protects startups. My answer and preference is to turn to startups. Since the startup has limited money and time to death, their focus is generally very clear. In big companies, projects have a hard time dying. “
Universities, meanwhile, have a different perspective and usually far less legal clout. “Universities do have patents but they generally are insufficient in the market to be either defensive or offensive,” Hogan said. “With the universities I have listed they have reasonable expectations and a long history of listening to their technology. With that said, universities that haven’t done much are totally out of line with reality.”