When a university patent becomes a company patent: what do the documents establish?

In 2003 the University of Nebraska filed a patent for a small robot that works inside the body during surgery; it was granted in 2006 and it names four inventors, Shane Farritor among them. In 2012 a second patent began, with eight inventors, Farritor again. A patent names two kinds of people: the inventors, who made it, and the owner, who holds the rights. It is tempting to read the second as the first — an inventor “leaving” a lab for a company. The register records no such thing. It records owners. So we follow one patent through its recorded owners, and nothing else.

The second patent was filed on 15 March and assigned to the university’s Board of Regents on 10 July: the inventors handed the rights to their employer, as university researchers usually must. On 9 August 2017 the university assigned it to Virtual Incision Corporation. It was granted on 10 March 2020, nearly seven years after filing. That is what “a university patent becomes a company patent” means in the record: one dated document, from one owner to another. What we can add from outside the register: the university’s own page for Farritor says he “is founder and has a financial interest in” Virtual Incision, “based on his research”. The company that received the patent was the inventor’s.

The nearest thing to a product in the public record is a regulator’s decision. On 23 February 2024 the US Food and Drug Administration granted Virtual Incision a De Novo classification for the MIRA Surgical System, which it describes as a table-mounted miniaturised surgical system. Our own record held none of this: the decision came from the regulator’s database, not from our sources, and a missing event in our record is not evidence that it never happened. So the chain we can defend runs: an invention at a university; the rights assigned to the inventor’s company; a device cleared for sale. Three documents, three dates, and no claim beyond them.

Of 10,144 inventor identifiers first named on a patent owned by a university or public research body, 665 are later named on a patent owned by a company — 6.6%. 570 of those later patents fall in one of the same broad robotics classes; 129 are co-owned with a research body, so the institution did not change so much as widen. These are identifiers, not people; the pattern selected them, so “the subject stays the same” is partly the selection talking; sharing one of five broad classes is not the same invention; and later cohorts have had less time to patent again. It is a description of the register, not a rate at which researchers enter industry.