Skip to main content
Helikon Labs
Sign inStart free

← Back to the Research Blog

partnershipsipcontracts

Who owns the IP when a company funds university research? Four ways a sponsor gets rights, and what they cost

Andrew T. Flowers ·

"We're paying for the research, so we'll own what comes out of it." It's the most common assumption a first-time sponsor brings to a university, and it's wrong at almost every US research university, for reasons that have little to do with the deal in front of you. This article explains the three reasons the university holds title, the four routes by which a sponsor gets rights anyway, what the universities that publish their prices charge, the one well-known US university that lets sponsors own the IP, and what the research says about whether the ownership model works at all. It's written for the sponsor. The university reader will recognize the arguments and may disagree with one of them.

We're not a law firm, and this isn't legal advice.

Why does the university usually own the IP, even when a company pays?

The law. Where U.S. federal money funds any part of a project, the Bayh-Dole Act lets the university elect to keep title to inventions, subject to disclosure, a license to the government, a preference for US manufacturing, and other conditions it can't contract away (35 U.S.C. 200-212 and 37 CFR 401.14). Most labs run on a mix of money, and the university's office will treat a project as federally funded unless it can prove otherwise. The Supreme Court added the detail that matters for sponsors in Stanford v. Roche (2011): Bayh-Dole doesn't vest title in the university automatically, because rights begin with the inventor, so the university's title depends on the assignment agreements it has with its researchers. A sponsor can only receive what the inventors assigned, and the first due-diligence question is whether those assignments exist.

The policy. Even with no federal money, nearly every US research university's IP policy says the institution owns inventions its employees make in the course of their work. That model predates Bayh-Dole. Mowery, Nelson, Sampat, and Ziedonis found that university patenting and licensing were already growing before 1980 and that the Act was an important but not a determinative factor in what followed (Research Policy, 2001). Thursby and Thursby found that the growth in licensing came mainly from universities and faculty becoming more willing to license and from business relying more on external research, not from a change in what faculty studied (Management Science, 2002). Institutional ownership is the operating system, not a negotiating stance.

The practical problem. The university can't price an invention that doesn't exist, and it can't assign away something a graduate student might invent next year under a different grant. What it can do is commit now to the terms on which you'll get rights later. That's what the pathways are.

What are the four ways a sponsor gets rights, and what do they cost?

1. University ownership with an option

The default. The university owns the inventions, and the sponsor gets a time-limited option, often six months from disclosure, to negotiate a license, exclusive or not, at terms to be agreed. The weakness is in the last four words: the price is set after the invention exists, which is when the two sides' views of its value diverge most. Hall, Link, and Scott found that of 25 research projects with no university partner, 12 reported IP as an insurmountable barrier to having one (Journal of Technology Transfer, 2001). The option-to-negotiate is usually where that barrier stands.

2. Field-of-use license

The sponsor takes rights in the field it cares about, and the university keeps the rest. Narrower, cheaper, and faster, because the university stays free to license the invention elsewhere and the sponsor pays only for the position it needs. For most industrial sponsors this is the right answer, and it's the one experienced sponsors ask for first.

3. Pre-priced menus

A growing number of universities publish the terms, so the IP negotiation becomes a selection. The University of Wisconsin-Madison's Badger IP Industry Advantage offers three. The Classic has no upfront fee and a license negotiated when an invention appears. The Varsity, for a fixed $10,000, lets the sponsor elect a non-exclusive license in a field of use at invention creation or the right to negotiate an exclusive one. The Bascom is a pre-paid exclusive license at 10% of the project budget (or $15,000, whichever is greater) for one field of use or 15% (or $22,500) for all fields, paid at the start, royalty-free unless annual sales pass $20 million, then 2%. Stony Brook's CLEAR-IP offers a pre-paid non-exclusive royalty-free license for the greater of $10,000 or 15% of the project budget (with a free six-month option to negotiate an exclusive one), an exclusive license for the greater of $15,000 or 20% with a 1% royalty once cumulative sales pass $20 million, and a traditional route. The University of Minnesota's MN-IP sets a one-time upfront fee, royalty-free non-exclusive licenses, and a 1% royalty on exclusive licenses only once product sales pass $20 million a year. When it announced the program's expansion in January 2014, the fee for its Create track was 10% of the agreement or $15,000, whichever was greater, and the track had produced 83 partnerships.

Read those prices against the option-to-negotiate and the logic is clear. For a $300,000 project, an all-fields exclusive at Wisconsin costs $45,000, known on day one, with no royalty until the product is selling $20 million a year. A sponsor who would otherwise spend that on counsel to argue about a hypothetical invention has every reason to pick from the menu.

4. Sponsor assignment, with conditions

One well-known US research university went further. In December 2011 Penn State removed the requirement that the university own IP from industry-sponsored research, so that, in the words of its then vice president for research, "the ownership of IP developed with industrial funds" could "flow back to the sponsor" (Foley, 2012). The policy still stands, as an option rather than a default: for projects funded with the sponsor's own money, Penn State may assign the IP or grant an option to an exclusive license on pre-negotiated terms, provided all project personnel agree, no one on the project has a significant financial interest in the sponsor, and the work isn't human-health IP. Where the industry money is a pass-through or match for federal funds, Bayh-Dole applies and the university keeps title. The rationale Foley gave is worth quoting to any contracts office that says it can't be done: the university decided to value "the net present value of the interactions and relationships" over the future value of IP proceeds.

Outside the US, the University of Glasgow went the other way in November 2010 with Easy Access IP, offering a large share of its existing IP to companies free of charge, royalty-free, on a one-page agreement, with the University of Bristol and King's College London joining in 2011 and the UK's Defence Science and Technology Laboratory in 2015. It covered existing inventions rather than sponsored projects, and it's the clearest demonstration that a university can decide its IP is worth more deployed than defended.

Translation Box: "ownership"

On the academic side: the university owns it, the inventors share the royalties, the technology transfer office decides whether to patent, and the researchers keep the right to use it in further research and to publish. "Ownership" is a stewardship role with legal obligations attached.

On the industry side: ownership means freedom to operate, the right to exclude competitors, and a line on the balance sheet. Most sponsors who say "we need to own it" need the first two, and a field-of-use exclusive license delivers both.

Does university ownership of IP work? What the evidence says

A sponsor who thinks the university's position is arbitrary should know that scholars have argued the same, and that the evidence from countries that tried the alternative is mixed.

Kenney and Patton argue that the current model, "in which universities maintain de jure ownership of inventions, is not optimal either in terms of economic efficiency or for advancing the social interest of rapidly commercializing technology," and that it's "plagued by ineffective incentives, information asymmetries, and contradictory motivations" among universities, inventors, licensees, and licensing offices (Research Policy, 2009). Their alternative is inventor ownership, which is what the Swedish "professor's privilege" provides and what Germany abolished in 2002.

The comparisons don't settle it. Åstebro, Braguinsky, Braunerhjelm, and Broström compared US and Swedish STEM PhDs and found the rate of entry into entrepreneurship from academia "slightly lower in the U.S. than in Sweden," with negative mean economic gains for academic entrepreneurs in both countries (ILR Review, 2019). Czarnitzki and colleagues studied Germany's shift of patent rights from professors to universities and found "no evidence that the new infrastructure resulted in an increase in start-up companies by university researchers," while the shift "substantially decreased the volume of patents," with the largest decrease in faculty-firm patenting (European Economic Review, 2016). Universities can point to their licensing growth. Critics can point to what the German experiment cost. Neither side has shown that the other model clearly produces more innovation.

For a sponsor, the practical conclusion is that the ownership question is a policy argument you won't win in a negotiation, and that the pathways exist because the universities know it. Pick the route that gives you what you need, in the field you need it, at a price you know on day one.

What should a sponsor settle before drafting the IP clause?

What you actually need. Freedom to operate in your field, an exclusive position, or just the results. Sponsors who ask for everything get the slowest negotiation and often the least.

Which pathway the university publishes. If it has a menu, pick from it. If it doesn't, ask whether it will accept a field-of-use exclusive at a fixed fee set at signing, which is what the menus amount to. And for AI-assisted work, where the USPTO's November 2025 guidance restates that only natural persons can be inventors, say in the agreement who the human inventors are expected to be and how the tools will be documented, or ownership of the foreground will be contested before it's assigned.

The agreement's terms, once signed, live with the project. A project on Helikon records the agreement type and the signed contract with its dates, so the team running the project can see what the negotiators agreed. You can set one up on a free account. For guidance on the whole agreement, start with the sponsor's guide.

Sources

Start free on Helikon Labs

Book Office Hours, the first 30 minutes are free

ShareXLinkedInEmail

Related posts

Can a sponsor block publication of university research? What universities accept, and what the model clauses say

Background IP vs foreground IP: what each side brings, keeps, and shares

The STTR allocation of rights agreement, explained for founders

Who owns the IP when a company funds university research? Four ways a sponsor gets rights, and what they cost · Helikon Labs