Counsel for the
frontier of computation
Research

University and startup: the license terms that matter in quantum

Many quantum computing companies begin with a university license. The terms that most often need working out between the two, and why several of them are harder in quantum than in other fields.

Many quantum computing companies begin in a university lab. IonQ’s core trapped-ion patents are licensed from the University of Maryland and Duke University. Oxford Ionics was founded by two Oxford physicists in 2019 and sold to IonQ in 2025 for $1.075 billion. The license between the university and the startup sets the terms of that relationship for years.

Both sides want the company to succeed, but they come to the table with different responsibilities. The university has to protect its researchers’ freedom to publish and teach, meet the conditions of its research funding, and see the technology put to use. The startup has to protect what it builds, raise money and ship a product. This post covers the terms that most need working out, and why several of them are harder in quantum than in other fields.

Who owns the next invention

The license covers the patents that exist when it is signed. The lab keeps working, and founders often keep their faculty posts. One recurring question is who owns what the lab invents next.

The startup’s concern is that the next invention may be the product. Quantum hardware changes quickly, and a control scheme developed two years after the license may matter more than the original patent. The university’s concern is that it cannot commit the future work of every student and researcher in a lab, some of it funded by sponsors with their own rights. Common middle grounds are an option or a right of first negotiation on improvements, limited to named people or a defined project and to a fixed period. IonQ’s filings describe one version of this. Alongside its main license, IonQ had an option agreement with Duke, entered in 2016 and running to July 2026, that let it add Duke’s interests in later patents relating to quantum information processing devices to the license, including work by three named researchers and the people they supervised.

Publishing and filing

The university’s mission is to publish. The startup needs to file before anything is published. Most agreements give the company a review period before a paper or thesis is released, so that an application can be filed first.

In quantum the length of that window matters, because the papers are often the prior art an application has to get past. They also define the field’s terms. Rigetti Computing, a US company that builds superconducting quantum computers, holds European patents that third parties challenged after grant through the European Patent Office’s (EPO) opposition procedure. In one of those cases the EPO’s opposition division read a key term in Rigetti’s patent by reference to the field’s published papers, including Rigetti’s own 2012 publication, and the reading went against Rigetti. A workable agreement keeps the review period short, so publication is not held up, and asks both sides to use consistent terminology in papers and patent applications.

Diligence milestones

Universities want licensed technology developed, and federal funding rules expect it. Licenses usually set milestones, and a missed milestone can let the university narrow the license or end it.

Quantum development timelines are long, and hardware roadmaps change. A milestone written around a 2020 roadmap may not fit the machine a company is building in 2026. Milestones tied to development steps the company controls are easier for both sides to measure than dates for a particular qubit count. An agreed process for extensions, and a remedy that narrows the field rather than ending the whole license, protect both sides.

Who runs and pays for the patents

Universities often control prosecution and pass the costs to the licensee. Universities file broadly and cannot fund every case everywhere, and the public record shows the consequences. At the EPO, universities lost 99 of their 246 concluded quantum applications, about 40%. 71 of the 95 withdrawals happened before any examination report, which usually means a fee was not paid or the application was not taken forward after the international phase. At the United States Patent and Trademark Office (USPTO), the same universities abandoned 11% of concluded applications.

A startup that licenses a patent family may find that the European half lapsed before the company existed. Agreements can address this by giving the licensee a say in where to file, a right to take over any case the university plans to drop, and a clear split of costs.

What the university keeps

Universities almost always keep the right to use the technology for research and teaching, and often the right to license it in other fields. The field of use therefore matters to both sides. IonQ’s license covers “the application of the licensed intellectual property in ion trap quantum computing.” A field written around one qubit modality leaves sensing, networking or another hardware platform available for the university to license elsewhere. That may suit the university’s mission and the startup’s current plans, or it may not suit the company’s later ones.

The agreement should also say who has the first right to sue an infringer, who pays for it, and how any recovery is shared.

Money and federal funding

Payment usually combines equity, royalties and a share of sublicense income. UK practice offers a reference point: a 2023 government review recommended that universities take 10% to 25% of a spinout, the average UK university stake fell to 16% in 2024, and a 2024 guide for software spinouts recommends 5% to 10%. Structures vary. IonQ’s license is royalty-free, and the universities received IonQ stock.

Federal funding adds obligations that bind both parties. Under the Bayh-Dole Act the government keeps a royalty-free license to federally funded inventions, and agencies hold march-in rights, although no agency has used them since the Act was passed in 1980. The more practical rule is in 35 U.S.C. 204: an exclusive US licensee of a federally funded invention must agree that products embodying it will be “manufactured substantially in the United States,” unless the funding agency grants a waiver. The university has to report and comply; the company has to manufacture accordingly. For quantum hardware built with cryogenics, lasers or fabrication from abroad, that clause deserves early attention. IonQ’s annual report says these rules “may limit our exclusive rights and our ability to contract with non-U.S. manufacturers.”

When the company is sold

Many licenses cannot be assigned without the university’s consent, and some treat a change of control as an assignment. The university wants assurance that the technology goes to a company that will develop it. The startup needs to remain acquirable. Agreeing in advance on when consent is required, and on what terms, avoids negotiating it in the middle of a sale.

Revisiting the agreement

A license signed at founding reflects what both sides knew then. Either side can propose revisiting it as the company grows, and doing so before a financing or a sale avoids negotiating against a deadline.

Source: CipherLaw; based on public filings and the USPTO’s and EPO’s records for quantum applications filed by universities.