[I’ve added some additional comments, and some light editing for clarity (I hope)]
I expanded a discussion of the odd wording of Bayh-Dole’s 35 USC 204, which restricts the US manufacturing requirement to exclusive licenses “to use or to sell” in the United States. If the holder of patent rights on a subject invention does not license at all–a small company makes and sells product that practices the subject invention, say–then there is no US manufacturing requirement. Similarly, if the holder of patent rights on a subject invention grants non-exclusive licenses, still no US manufacturing requirement. If the holder of rights exclusively grants all substantial rights (make, use, sell), then the transaction is in effect an assignment and a different part of Bayh-Dole controls–35 USC 201(b), having to do with what happens when a contractor makes an assignment, and 35 USC 202(c)(7)(A) when that contractor is a nonprofit. This stuff may sound hairy, but really it isn’t, except that a whole lot of people who should know better (and worse, some do) have mucked everything up by chronically misrepresenting Bayh-Dole so that in addition to following things clearly, anyone who has happened to listen to these folks has to also unlearn what they thought they knew. Like finding out your elementary teachers really didn’t know squat about science or literature and you had the misfortune to be their really good student.
Anyway, here’s the expansion, given its very own post.
As a point of practice, there are multiple ways for a university to set a company up with exclusive commercial rights to practice an invention. One can expressly assign the invention and patent rights (and patents) and be rid of it all. The company runs with it, and provides whatever compensation has been agreed to. Another term for this sort of transaction is “sale.” Or, one can grant an exclusive license to all substantial rights–to make, to use, to sell–and in effect transfer ownership of the invention to the company, even if title to any patent does not formally change. Such a deal still assigns ownership of the invention.
A third method is to grant an exclusive license but reserve rights for university “non-commercial” or “educational” or “research” uses. While this might look like a transaction for less than all substantial rights, courts have not been persuaded and view such deals as an assignment and a grant back of a limited non-exclusive license. The “commercial” rights to make, to use, and to sell are the substantial rights.
If one expands the rights not licensed–so that for instance, a license grants exclusively the right to sell, but grants only non-exclusively the rights to make and to use, then the license does not act as an assignment. Similarly, if the term of an exclusive license is substantially less than the full term of the patent, such as, say, eight years, then again the license is exclusive but may not be regarded by a court as an assignment.
Or one may grant a sole license–the company is the only entity other than the patent holder that may practice the invention. In this variation, the patent holder and the licensee both may practice the invention–the sole license does not exclude the patent holder from also practicing the licensed invention, and so long as the patent holder does not sell, the sole (non-exclusive) licensee has, in effect, exclusive commercial rights. At this point, the sole license does not appear to act as an assignment, even if the patent holder (a university, say) agrees conditionally not to grant additional non-exclusive licenses. It may be worth noting (you decide) that a sole license may also come at the end of a string of non-exclusive licenses. In effect, the sole licensor buys out all future non-exclusive licenses.
[There’s an interesting way to value such a sole license, too, taking into account the potential value of those future non-exclusive licenses, less the cost to get them, and reduced by the likelihood one can get them. Consider that–even thinking in terms of non-exclusive licensing ended by a sole license creates a pretty easy value pattern for any potential sole licensee to figure out on their own. At least, then, whatever consideration is asked, it appears reasonable–“buy me out of what I reasonably expect to make without you.”]
Yet another approach is to grant a non-exclusive license while agreeing for a time to “stand still” and not grant any other non-exclusive licenses while the licensee meets whatever criteria. This approach is also what is used when a company takes an “option” to an exclusive license, or has a non-exclusive license with a first right to negotiate an exclusive license, or even has a limited “evaluation” license prior to negotiating for the right to sell (or to make or to use other than for the purpose of evaluating the invention to determine whether to take a license). The value of the deal then is divided between the value of standing still (loss of other licensing (and research) relationships plus a share of the value realized by the licensee (making, using, selling, excluding all others for a time).
In the long of it, there are a number of ways by which a company can have an exclusive market position. Some of those ways involve assignment (even if by an instrument labeled “exclusive license”), others involve only an exclusive license (short of assignment), and still others involve some variation on granting a single non-exclusive license on whatever conditions, such as standstill or an option to upgrade the license later to exclusive).
The effect of the 35 USC 204 wording is that if an exclusive license does not rise to an assignment, then 204 controls. If a nonprofit’s exclusive license does assign (make, use, sell; right to enforce or settle; right to sublicense, full term of the patent, in substantially all fields and territories), then 204 does not control–instead, 202(c)(7)(A) controls–the licensee is an assignee of the invention (though maybe not title to the patent), and consistent with the mechanism in 201(b) and (c) by which a contractor may add parties to a funding agreement by “any assignment,” the assignee becomes a party to the funding agreement (must accept the nonprofit’s patent rights clause) and therefore also a federal contractor, with the rights (such as 202(a) right to retain title) and obligations (disclose any subject invention made in the federally funded project which it acquires, even if the company-assignee has not received any federal funding itself–one of the implications of “in part” in 35 USC 201(b)). Again, 204 controls only certain kinds of exclusive licenses; 202(c)(7)(A) and 201(b) control assignments and exclusive licenses that assign; nothing much controls non-exclusive licenses, especially if not bounded by a standstill agreement.
These differences among the varieties of exclusive license make AUTM’s licensing survey that merely counts “licenses” all that more useless. There’s a huge difference even among exclusive licenses–exclusive licenses that create multiple opportunities for investment vs exclusive licenses that assign vs licenses that create for a time an exclusive commercial position but aren’t necessarily exclusive or are, well, non-exclusive. If a university granted non-exclusive public licenses to every company that wanted to practice, then the number of licenses would be a coarse proxy for the extent of a university’s company relationships related to technology transfer. But even that count would not distinguish between one company taking multiple licenses and a number of companies taking one license each. And we still would not know whether the university licensed just one invention a hundred times (with 99 unlicensed) or licensed a hundred inventions one time each. And of course, if universities and their inventors didn’t go off and establish ownership positions in research results, and relied on possession of information, on tacit knowledge, on the excitement of getting into something new, then there would be no “licensing” as a barrier to access. There would be other barriers–withholding information, waiting to publish first, inherent complexity, instability of results, need for specialized skills or access to custom instrumentation–but licensing would not be one of them.