When Uber announced plans last fall to invest $200 million on its Advanced Technology Group (ATG) in Toronto, the company made a point of showcasing a key hire—Raquel Urtasun, an associate professor of computer science at the University of Toronto. The MIT and UC Berkeley scholar is a leading light in the world of machine learning-based vision systems, which figure prominently in the future of automated vehicle navigation. Under the agreement, Urtasun became Uber ATG’s chief scientist, while retaining her academic positions at both the university and the Vector Institute, which she co-founded.

The move gives Uber unrivaled access to Urtasun’s research and to her graduate students. It reflects the university’s emergent philosophy of tech transfer—that is, to forge close partnerships with private firms with the market experience and capital to commercialize academic research. The university’s approach, says Derek Newton—U of T’s assistant vice-president of partnerships, innovation and entrepreneurship—“has worked very well.”

The Uber investment fits neatly into a broader narrative about how international tech firms have brought jobs and capital to AI research hubs in the Toronto-Waterloo corridor, Montreal and Edmonton.

Yet such deals also raise questions about how Canadian universities are managing tech transfer and the difficult business of licensing inventions to startups, investors or large firms with an eye to commercializing university-generated intellectual property.

Many leading Canadian research universities have made a point of touting the ever-growing network of incubators, accelerators, corporate research deals and entrepreneurship programs that have cropped up on campuses in recent years. As Newton says, “We’ve seen an explosion of this kind of activity.”

But the picture gets fuzzier when it comes to licensing IP generated by university researchers. As a recent McGill University board of governors report noted, licensing revenue by Canadian universities is “at least an order of magnitude” lower than at U.S. institutions—and, what’s more, there’s no predictability to this form of income. As the report points out, the University of Saskatchewan led all Canadian schools from 2014 to 2016, with almost $50 million in revenues from licensing a single patent.

U of T, meanwhile, saw its royalty stream dwindle sharply from more than $34 million in 2014 to less than $2 million in 2015 while the University of Waterloo, which famously allows its scientists to own their inventions, generated just half a million from 2014 to 2016. (In a high-profile exception to the rule, UW recently sued a startup that emerged from one of its incubators, citing patent infringement. The two parties have since settled.) A 2008 Statistics Canada survey calculated that all Canadian universities and research hospitals generated $53.2 million in royalty income. Less clear is what the national picture is 11 years since then: StatsCan discontinued the survey in 2012.

In general, Canada doesn’t have a great record when it comes to patenting its intellectual property. According to the Conference Board of Canada, we rank 14th out of 16 OECD nations in so-called “triadic” patents—those filed to offices in the U.S., U.K. and Japan—and lag behind not just tech powerhouses like Japan and Germany, but also smaller nations like Switzerland and Sweden. The Netherlands, which has half of Canada’s population and half its gross domestic product, generated more than twice as many triadic patents in 2015, according to data from the OECD.

Many patents, of course, are registered by private firms. Furthermore, as Charles Plant, senior Impact Centre fellow at U of T, points out, the mere fact that an invention is patented, either by a corporation or a university, doesn’t mean it will be successfully commercialized.

Plant and other critics say universities aren’t good at commercializing research by licensing and marketing IP, and probably shouldn’t be in this business, anyway. “Universities are not engines of innovation,” says Waterloo IP lawyer Jim Hinton, who has conducted a study of Canadian university patents for the Centre for International Governance Innovation (CIGI). He and others point out that academics focus on fundamental research, and tend not to worry about the elements that drive innovation-driven companies, like marketing and product development. “We’re putting universities in a position where they’re not going to be successful.”

Yet this dynamic is hardly universal. Those in the field can recite the royalty windfalls generated south of the border by landmark infringement lawsuits or homerun inventions. Among them is Lyrica, an anti-epileptic drug developed at Northwestern University that has netted the Chicago institution US$1.3 billion in revenues. Others include Remicade (New York University; US$1 billion), Gatorade (University of Florida, US$228 million), according to a ranking compiled by Bloomberg. Stanford University, in fiscal 2017, earned over US$45 million in royalty revenue, and has seen its portfolio of income-generating patents grow from 371 in 2001 to almost 800 in 2016.

The list also features some high-profile legal actions, including a US$750-million settlement to Carnegie Mellon University against Marvell Technology, a software firm that the Pittsburgh research powerhouse sued for patent infringement in 2009. There are no such blockbuster results in Canada.

Jacob Rooksby, Gonzaga University’s law school dean and author of The Branding of the American Mind: How Universities Capture, Manage, and Monetize Intellectual Property and Why It Matters, says university tech-transfer offices (TTO) in the U.S. have become far savvier and better resourced over the past 20 years. While many universities still lock down patents on inventions but then can’t find licensees, he says the paradigm has largely shifted since 1980, when the U.S. Congress passed the Bayh-Dole Act. That law was intended to let universities and their researchers commercialize federally-funded science. Until then, discoveries effectively belonged to the U.S. government, which was unwilling to license inventions to private firms. No comparable restriction ever existed in Canada.

Much has changed since the law came into effect. Rooksby says TTOs today have rebranded themselves, and are more likely to be staffed with PhDs instead of people with general business backgrounds, meaning they have more in-house capacity to identify discoveries with commercial potential. Corporate-funded research partnerships, moreover, have become commonplace. He also points out that these departments now rely more on highly-specialized external patent lawyers, rather than in-house generalists, and they’ve become far more assertive about using the courts to defend patent infringements.

“There’s no getting around the fact that in U.S. higher education, the entrepreneurial spirit has gone from being seen as a bad thing to a positive,” Rooksby observes. “We know about the things that correlate with success.”

The moral of this story is complicated. While Rooksby and others stress that no university should ever rely on royalty income and blockbuster discoveries as a predictable source of revenue, the overall economic impact of such activity is substantial. According to a 2017 study for the Biotechnology Innovation Organization and the Association of University Technology Managers, the total contribution of academic licensors to gross industry output from 1996 to 2015 ranged from US$320 billion to US$1.33 trillion (2009 dollars), including millions of person years of employment created. In the last six decades in fact, the share of total U.S. R&D activity performed by universities grew from 5.3 per cent to 14.2 per cent.

What about the wider impact of Canada’s TTOs? In a recent paper, CIGI senior fellow Karima Bawa noted that the $12 billion spent on university and hospital research in Canada each year is well-regarded internationally. Universities now have “significant patent portfolios,” the study notes. But, she says, Canadian university TTOs tend to be less well-resourced and less litigious than their U.S. counterparts, partly out of a fear of alienating potential corporate donors and research sponsors.  

In his forthcoming CIGI study, Hinton looked at patents registered by the leading 15 Canadian research universities. He found that about half didn’t find an assignee. For the other half, about two-thirds were scooped up by foreign firms. Hinton says the main corporate consumers of Canadian patents are BlackBerry, IBM, Xerox, Ericsson and Huawei, which has attracted intense scrutiny for the extent of its research partnerships with Canadian universities.

Canadian university-originated IP that gets locked down by large multinationals like Google effectively hinders Canadian startups, which may find themselves paying royalties to much larger competitors to use patented algorithms or other building blocks. “That has a real impact on Canadian companies,” Hinton says. “The universities are complicit in this whole thing.”   

Hinton argues that Canadian public policy should be more proactive about directing Canadian IP to domestic firms, for example by requiring university TTOs to add riders on licences that the technology be exploited here by Canadian firms. But others are skeptical about whether the exodus of patents is a solvable problem. “There’s no closing the loop,” says Plant, whose research shows that Canada ranks third in the world when it comes to assigning patents abroad. As he notes, our failing isn’t transforming inventions into innovations; rather, the problem is the country’s long habit of allowing international firms to snap up Canadian IP. “The system wasn’t designed to be successful.”

He argues that the marketing of university and hospital research should be carried out not by individual institutions, but by specialized regional agencies with the resources, budget and expertise to connect university and hospital-generated IP to potential end users.

The private sector is generating its own solutions. In the past five years, TandemLaunch, a Montreal venture capital firm, has developed a model of launching startups that begins with the company searching for promising university patents, which it then licenses as part of an extended process that involves funding proof-of-concept research, recruiting startup executives and providing early-stage capital.

According to technology director Omar Zahr, TandemLaunch has graduated 14 companies, has six more in its pipeline and has raised two funds with a total of $25 million in available venture financing. While Zahr says Tandem also looks internationally for IP, it’s had a positive relationship with Canadian institutions. “They get our model quickly and understand our pain points,” he says. But, he adds, the American university ecosystem is “more developed,” with TTOs and researchers who are better at marketing their IP to prospective end users.

Ultimately, Plant points out, the weakness in Canada’s approach reflects a long-standing and well-recognized dilemma, which is that innovation-driven firms founded here grow slower, and then tend to get acquired by larger international companies before they’re large enough to become significant acquirers of domestic discovery.

“We have lots of small companies,” he says. “We don’t have [as many] larger companies.” That situation, he adds, isn’t getting much better over time. “Our approach to solving this problem is not evidence-based,” Plant adds. “It’s opinion-based. We end up identifying the symptoms, but not the root causes.”