Every week, Prof. Amnon Shashua convenes his class on the Givat Ram campus of the Hebrew University in Jerusalem. A computer science professor and newly minted billionaire, Shashua teaches an advanced seminar on deep learning. The class is theoretical, but he could tell them a thing or two about turning theory into fact, and how much their research could change the world one day.
That is what Shashua aims to do with Mobileye, which he founded over 17 years ago with Ziv Aviram. Mobileye developed driving-aid systems and today its technology is being used to develop autonomous cars. That’s why Intel just offered to buy it for $15.3 billion.
Shashua came from MIT, where he did his PhD in brain and cognitive sciences, and post-doc in biological and computational learning, then joined the Hebrew University in 1996 – three years before establishing Mobileye.
Mobileye’s technology is based on his research, which raised questions about intellectual property at the company’s beginning. The norm at universities is that IP developed during research doesn’t belong only to the researchers but to the institution employing them.
But though some of the development of Mobileye’s technology began at Hebrew University, it’s not getting much from the company’s meteoric growth and imminent sale in the biggest takeover in Israeli history.
Compare that with the Weizmann Institute of Science, which over the years enjoyed large royalties from Copaxone, a proprietary drug made by Teva for multiple sclerosis.
What happened? “Mobileye wasn’t exactly a Hebrew University venture. It started as a private venture by Shashua and Aviram, and there was an argument to what degree the university had a part in it,” says Prof. Menachem Magidor, the university’s president at the time.
“That’s because much of the venture was done outside the university. Shashua claimed some of the research had been done while he was at the Technion [in Haifa], some on sabbatical and some at Mobileye itself. It was complicated.”
Universities protect their intellectual property through in-house technology transfer companies. Tel Aviv University has Ramot; Weizmann has Yeda; and Hebrew University has Yissum to commercialize research.
Transfer companies may operate by royalties, which is often the model used with drug development – the university keeps the patent and sells usage rights, in exchange for say 30% or 40% of sales. Or the university may hand over its technological rights in exchange for shares.
In Mobileye’s case, the Hebrew University elected for shares – though not before a dispute over its stake, which was settled by arbitration in the early 2000s. The compromise left Hebrew University with only a few percent.
But the university didn’t want to battle on and sue its own staffer; they held Shashua in the highest regard and wanted him to stay on, Magidor says. And the shares they got were diluted over time because the university didn’t invest more in the company.
He adds that the university isn’t in the business of making money, but of conducting research and teaching, and that the academic consideration must prevail.
Hebrew University commented that the controversy is long over. Mobileye declined to comment.
How many percent the university actually got isn’t known; it also may have sold shares some before Mobileye floated in New York. Anyhow, it arrived at the IPO owning 0.6% of the company. From the flotation to date, Hebrew University has made $40 million from selling shares.
Prof. Shmuel Peleg, Shashua’s colleague in the computer science and engineering department, says Hebrew University is very supportive of its researchers founding companies, certainly compared with other universities.
When a researcher founds a company – as he himself has done twice – he negotiate with the university over its share in it. When selling a drug patent to a company, the scientist simply shares the income with the university. Technology is different.
“In such cases, the researcher invents a mathematical formula, which is useful mainly for publishing a paper,” says Peleg. “Somebody else comes along and suggests using the formula in automated driving, for example – a new technology of unproven demand; even the formula’s ability to solve a problem hasn’t been proved. A company is set up, it applies the formula to the problem and persuades the world to move to automated driving. The argument will be over the university’s part in its success, what is its part in the idea, the implementation and the marketing.”
Some investment formulae tie stakes to future investment in the company’s research: if the future investment is high, then the technology is based less on know-how gained at the university and its share will be smaller.
“Some universities demand 40% of any company and no companies get set up there,” says Peleg, who advocates negotiation. Granted, arguments ensue, but the fact is that Hebrew University people produce a lot of companies, showing it’s on the right track, he says.
Furthermore, income from intellectual property rights is a tiny fraction of any university’s total budget, adds Magidor. Most universities that made anything from IP did it through two or three blockbuster successes, and Hebrew University is in a very distinguished position from that perspective – though Magidor admits that Weizmann beats it hand over fist.
“In principle, as a proportion of the budget, the return on the investment in research is usually far from covering it,” says Magidor. “It doesn’t pay, economically speaking. Hebrew University owns thousands of patents that it doesn’t try to commercialize. It isn’t good at it and that isn’t its relative edge. We try to commercialize a few dozen things that we believe may proffer commercial potential. Sometimes you miss. It’s like a venture capital fund. In Mobileye’s case, it’s hard to say we were confident.”
The university uses such income to improve labs, etc. Some years ago, a dispute arose between the computer science school and the university administrtration. The school claimed the administration owed it tens of millions of shekels and some senior teachers at the school took an unusual step: They sued. The case was settled out of court in secret.
It’s well documented that technology giants also fund the school. “We’re very strong in computer vision. The company that’s invested the most in us is Intel, but Google also gives research grants,” says Peleg.
And what do they get in return?
“From the perspective of IP, they don’t get anything,” responds Peleg. “Intel’s condition is that everything developed will be open to the public and everyone can use it – meaning developments can’t be patented. The companies do get access to the researchers who give lectures, and to the students, who join then afterward.”
They’re not contractors, he notes, but researchers. The companies can’t just ask them to develop projects, says Peleg, “because we’d refuse to do it for them.”
Want to enjoy 'Zen' reading - with no ads and just the article? Subscribe todaySubscribe now