What would you do if you were the university official in charge of company research partnerships and a Fortune 100 tech company offered to fund a research center at your university to the tune of $2.5 million dollars a year? If it were me, in these tough financial times, I’d say “great, tell me more,” visions of a nice press release dancing in my head.
But wait, there’s A Catch: the company has made it a condition that in order to receive the millions, your university must open source any resulting software and inventions that come out of this research funding. Yes, open source. Your university cannot stake claim to any patents. There will be no intellectual property clauses, no negotiations, no… nonsense.
This situation is not hypothetical. Since January, four U.S. universities have agreed to host Intel Science and Technology Centers (ISTCs) that will be funded at the rate of $2.5 million a year for five years. ISTCs will conduct research in various areas related to cloud and embedded computing. Each ISTC is led by two principal investigators, one from Intel and one from the faculty of the host university. Each Center will host three Intel researchers who will act as technology scouts.
In Intel’s own words:
“The IP policies and practices within the ISTCs will typically be designed to level the playing field for all of the participants, thereby enhancing cooperation and open collaboration. The preferred IP policy is to conduct open research wherein ISTC researchers, whether from academia or Intel, agree to not file patents and to publish all patentable inventions. All significant software developed in the course of conducting research will be released under an open source license.”
Intel may be the latest, but it’s not the first big tech company to prefer to sidestep university intellectual property (IP) clauses involving the negotiation of patent rights. Intel follows in the footsteps of HP and IBM, who also fund open source software collaborations with university researchers. Applications for additional ISTCs were submitted last week and more will be announced before year’s end. Here are current recipient universities of Intel funding:
- Stanford, January, 2011, Visual computing,
- UC Berkeley, June 2011, Secure computing,
- Carnegie Mellon, August 2011, Cloud computing,
- Carnegie Mellon, August 2011, Embedded computing,
To learn more, I spoke with Intel’s Matt Hancock, Director of Intel’s Science and Technology Centers program office. Hancock explained that “Whatever technologies are produced in the ISTCs will be made public, meaning anybody else can continue to build off of it, including Intel.” Hancock said that Intel had three major goals in funding university-based Science and Technology Centers.
- To increase the flow of ideas between the academic community and Intel
- To build a robust hiring pipeline at top-notch universities
- To change the way that the global research community thinks about cloud computing
Of course Intel isn’t purely altruistic. After all, the company has a bottom line to maintain and shareholders to answer to. Intel’s investments in software development via ISTCs signal a new direction in Intel’s product strategies. In search of better margins, Intel hopes to beef up its software practice to combat increasingly thin profit margins from chip sales.
According to Hancock, Intel is funding research in areas aligned with Intel’s future. “We hope that the research will indirectly benefit Intel’s product development efforts and perhaps Intel Capital, but that’s not the short term goal. We’re aiming at a long pipeline, 7 – 10 years out.”
Why not open source everything invented in university/industry partnerships?
Open sourcing research results sounds so reasonable. In fact, what’s not to like about this arrangement? After all, in many sponsored research negotiations that take place between university administrations and company legal teams, intellectual property (IP) clauses remain a sticky and sometimes time-consuming negotiating point.
What many people don’t realize is that few new patents arise from company sponsored university research projects. Although university patents get a lot of press as vehicles of innovative technology, according to the professional organization of university research adminstrators, NCURA, only 3% of industry sponsored research projects generate patents. Those are pretty slim odds, aren’t they?
Consider the potential upsides if open source licenses became a standard requirement of university/industry research collaboration:
- open sourcing fits nicely with the university’s non-profit, tax-exempt status; not patenting and licensing enures that sponosored research results will be made freely available to everyone at a fair market price
- open sourcing supports a core university mission, to disseminate innovative technologies quickly, cheaply and broadly
- open sourcing software lays the foundation for a sustained innovation ecosystem by creating a vibrant global community of researchers and tinkerers
- publicly sharing research results avoids publication delays induced by patent applications and removes restrictions on disclosing research results
- mandated open sourcing means no haggling over IP terms between university and company researchers, nor concern about patent rights
- sponsoring companies could more easily bring in additional companies to jointly sponsor open source consortiums since IP terms would not have to be negotiated for every member company
- a heartily endorsed, decentralized, open-sourced approach would open up multiple paths to commercialization, increasing the odds that university inventions would eventually find a commercial application
- open sourcing inventions would ease industry and public concern about the university’s ability to commercially broker technologies that originate in sponsored on-campus research
- open sourcing inventions would save the host university money on patent fees
- open sourcing inventions would minimize administrative costs to the host university since there’s little reporting, less oversight and no need for additional staff to manage the resulting research output
Universities are not required to file patents; they choose to
Universities are not required to patent what comes out of their research labs. Contrary to what many believe, there’s no law that obligates university researchers to feed potentially patentable knowledge into the university’s patent pipeline. However, you may hear people incorrectly claim that the Bayh Dole Act of 1980 made it mandatory that universities file for patents on on-campus research. True, the Act requires universities to take certain steps as beneficiaries of publicly funded research (see more info). However, rather than mandating that universities file for patents on university research, instead, the Bayh Dole Act gives universities the *option* to elect to take title to patents and the opportunity to commercially license them. The decision whether to patent or publish remains ultimately an inventor’s choice; many university inventors elect to do both.
Mandates to file university patents are actually a product of dictates at the local, not federal level. First, many universies require that their employees sign an employment contract in which they agree to give the university title to anything they invent during the course of their employment. Second, a contractual arrangement laid down by a specific IP clause in a single industry sponsored research agreement may require that sponsored research results be patented, if possible. Finally, many universities have campus IP policies that, at least in writing (typically not in practice), require researchers to get the permission of a central technology transfer office before they open source a technology.
Why patents have their place
But wait! Not all companies want their sponsored research to be shared freely. Although the private sector is increasingly comfortable working with open source technologies, many companies continue to prefer to fund university research that will be patented. Fair enough. A well-known example of a patent-reliant sector is Big Pharma. Given the long time-frame and high costs to get a drug through the FDA approval process, many pharmaceutical companies prefer to exclusively license patents in order to maintain a temporary monopoly position.
Open source works best in situations in which the company sponsor is not seeking crisply defined deliverables, but is seeking to broaden the existing knowledge base across the entire industry. Open sourcing results will not be appealing in cases in which a company wants a university research lab to do straightforward measurement or product testing or data crunching. Nor will open sourcing research appeal to companies looking to university research labs for proprietary improvements to existing commercial products, or to help develop an entirely new product.
As indicated by the approach used by Intel, HP and IBM, the software industry currently best lends itself to open source licensing. First, software is complex and a large part of its value lies not in viewing the source code, but in applying it productively. Second, software companies live and die by their operational prowess and executional abilities, not solely on the strength of their code. Third, software patents are notoriously difficult to enforce, therefore they are of relatively little value in providing a short-term, temporary monopoly, not to mention the fact that the software industry moves at the speed of light.
Ironically, although universities are the leading hotbed of open source licensing, formal university technology commercialization operations continue to focus on patents. (It’s important to point out that at the individual level, there are a number of forward-thinking technology transfer practitioners who are advocates of the commercial value of both patents and open source software licenses.) Nor do all faculty and students automatically embrace open source. In fields where a patent holds greater commercial value and does not disrupt the research process, a university researcher may elect to follow the formal technology transfer process and get a patent for her invention or technology.
The lesson from the Intel Centers is not that patents are bad, but that open source is a increasingly viable way to spark industry innovation and economic growth. A second lesson is that there’s no single correct way to handle the intellectual property issues when companies sponsor university research. Open source hardware licenses loom on the horizon. Let’s be open to them all.
Melba Kurman writes and speaks about innovative tech transfer from university research labs to the commercial marketplace. Melba is the president of Triple Helix Innovation, a consulting firm dedicated to improving innovation partnerships between companies and universities.