Most guides on how to commercialize university research describe a process. This one describes a set of decisions, because the process belongs to your institution and the decisions belong to you.
There are three real routes out of the lab: license the technology to a company that already exists, spin out a company built around it, or assign the rights outright. Which of those stays available to you is decided by one deadline, and most of the guides on this topic bury it under a definition. If you want the definition first, we cover what research commercialisation means separately. This page is about choosing.
The one deadline that decides how many routes you have
Disclose to your technology transfer office before any public discussion of the invention. Before you submit the paper, before the preprint goes up, before the conference poster.
This is not administrative caution. Late disclosure destroys patent eligibility in most countries and starts a 12-month clock in the United States that often expires before the technology is licensable. The disclosure form commits you to nothing except notifying your institution. It preserves your options, which is the entire point.
The institutional view is the same one. Case Western's technology transfer office advises early contact precisely because public disclosure risks precluding patenting, including through journals and presentations, and it is during the assessment window that follows disclosure that the licence-or-startup strategy is actually chosen. You want to be in the room for that conversation, not receiving its conclusion.
If you have already published, the option is narrowed, not closed. Talk to the office this week.
Does this research belong in a company at all
Three signals separate research that should be commercialized from research that should simply be published well.
A concrete application. Not a new approach to studying something, but a product a named buyer would pay for.
A finite path to it. You can describe what has to happen for this to become a product, and that list ends. "Three years of development and clinical validation" is a path. "Solve an open problem in the field first" is not.
A reason exclusivity matters. If a competitor can replicate the work within a year of reading your paper, an exclusive licence buys nobody anything. If reproducing it takes non-trivial expertise, exclusivity protects the investment a licensee has to make.
Now the denominator. Across participating US institutions in FY2025, more than USD 112 billion in research expenditure produced just over 28,000 invention disclosures, with total licences up 7 percent and new startups up 15 percent. Of the licences that do get signed, roughly 5 to 10 percent generate more than USD 50,000 a year in royalties. The rest produce less, or nothing. Trent University's commercialization handbook puts the startup side just as plainly: a very large proportion of start-ups fail, and the route that carries the highest risk is also the one that carries the highest potential reward.
None of that is a reason not to proceed. It is a reason to choose the route deliberately rather than defaulting into one.
Route one: license to an existing company
You keep ownership and grant a defined set of rights. The licensee carries development cost and market risk. In exchange, the returns to you and your institution are substantially lower than if you commercialized it yourselves.
That is the trade, stated honestly, and the handbook above is right to lead with it. Licensing is the correct answer when an established company in the space has the resources and the motivation to develop the technology, when the invention is incremental rather than category-creating, and when you want to keep your academic career rather than acquire an operating job.
A licence is not a handover. NC State's guide is explicit that license agreements stipulate the licensee must diligently bring the technology into commercial use and provide a reasonable return, with income flowing back to inventors, departments and central administration. Performance obligations are the part inventors under-negotiate. A licensee who shelves your technology has still technically licensed it.
The economics come down to a number nobody publishes as a standard. Inventor share of net royalty income at US universities is typically 30 to 50 percent, but that percentage applies to whatever rate your institution negotiates, and the rate does far more work than your share of it. Before you accept a term sheet, understand the average royalty rate for licensing intellectual property in your sector, because a generous split of a weak rate is still a weak outcome.
Route two: spin out a company
Spin out when no existing company is the right vehicle: the technology is too early, too uncertain, or too category-creating for an established player to absorb, and it needs dedicated focus that would not survive inside a multi-product firm.
There is one criterion that outranks the others, and it is personal. Half-committed spin-outs are the most common failure mode in university technology transfer. If you would rather stay in your lab and the only available chief executive is someone who will not commit fully, the venture will struggle regardless of how good the science is.
The route is well capitalised where it works. UK university spin-out companies attracted GBP 3.4 billion in 2024, a 44 percent year-on-year rise, against USD 39.3 billion invested into US university start-ups in 2021. The AUTM FY2025 data shows institutions leaning further in: licences and options including equity rose 16 percent, and the number of startups reported as non-operational fell almost 10 percent, which suggests the ones being formed are surviving longer.
Before the company exists, do the four-step work the handbook lays out: map the product or service options the technology enables, develop the two or three strongest into concepts, run a feasibility analysis, then pick one focused market entry. Early-stage companies fail by trying to serve every application at once. What you give up in exchange for the upside is equity and control, and the terms are institution-specific, so read how university spinout equity is split before you sit down to negotiate.
Route three: assign, or partner without either
Selling or assigning ownership outright is the third channel, and it is the right one when neither a licence nor a company fits: you want a clean exit from the technology, or the buyer needs full ownership to justify the investment.
There is also a hybrid the US-facing guides tend to skip. The European R2B commercialisation guide names a third route combining licensing with equity, alongside industry partnership as a destination in its own right rather than a step toward one. A co-development agreement with a company that already owns the channel can put your technology in the market faster than either classical route, and it is frequently the only route open to a technology whose market is small.
The funding stack from disclosure to first revenue
Take non-dilutive money first, in this order.
Internal proof-of-concept funds come first because they are the least competitive. Then the national translation programmes: NSF I-Corps in the US, Innovate UK grants, and in Europe the ERC Proof of Concept scheme and the EIC Pathfinder and Transition instruments covering roughly TRL 1 to 5. The R2B guide records over 1,350 innovations supported through EIC programmes with more than 100 start-ups created to commercialise the results, which is a real conversion rate, not a brochure number.
The amounts are smaller than founders expect and larger than researchers expect. Arizona State publishes concrete figures: NSF I-Corps Teams grants of USD 50,000, and sustainability-focused non-dilutive grants of USD 100,000 to 250,000 for hitting technical and commercial milestones. SBIR and STTR sit behind those at the federal level.
Budget for the fact that patent prosecution will consume a serious share of this before any of it reaches the bench. Once an application is filed, an issued US patent typically takes several years and tens of thousands of dollars, and foreign patents can run to hundreds of thousands. If the funding stack is where your route lives or dies, the European side of it is worth its own read: see EU funding routes for deep tech startups.
What has to be true after the licence is signed
Here is the part every ranking guide on this keyword stops short of, and it is where outcomes are actually decided.
Patent geography is now a commercial decision
Read the FY2025 numbers again, this time against each other. Licences rose 7 percent and disclosures rose nearly 8 percent, but new patent applications fell nearly 5 percent and foreign patent filings fell 40 percent. Institutions are not commercializing less. They are protecting the same inventions in far fewer countries, because prosecution costs force them to concentrate resources where commercial potential is strongest.
For an inventor, that is a live risk rather than a statistic. If your technology's real market is Germany, Singapore or Nigeria and your institution files only in the United States, you have a licence to an asset that is unprotected everywhere it matters. Raise the question of jurisdictions during the assessment window, while the filing strategy is still open. It is the same discipline that governs patent strategy for deep tech startups: file where you will sell, not where filing is cheapest.
Capital and operators, not enthusiasm
A signed licence is a starting gun. Expect years 0 to 1 for filing and marketing, 1 to 2 for term sheet and execution, 2 to 5 for the licensee's development, validation or scale-up, and 5 to 10 before first commercial sales and any royalty flow. The AUTM survey counted nearly 2,000 licence amendments in a single year, which is what it looks like when deals have to be renegotiated to survive contact with reality.
Two things determine whether anyone crosses that gap. The first is deployment capital, which is a different instrument from research funding: it buys units, sites and certifications rather than experiments. The second is an operating team that has built the thing before.
This is the half of the process we run. EX EPIC finances, patents and deploys breakthrough science across four continents on a EUR 160M+ capital track record, with 200+ waste-to-energy units deployed across 11 countries, 24 patent families filed with around 100 more validated in the pipeline, and 250+ trained operators placed into portfolio companies. The pattern that repeats is not that good science fails to find a licensee. It is that the licence gets signed and then nobody funds the two years between a working prototype and a certified, installable unit. That is the specific job of a deep tech venture builder, and the reason to understand it now is that the version of your route that includes a builder is negotiated at the term sheet, not afterwards.
A short decision checklist
- Disclose to the technology transfer office before anything becomes public. Everything else is reversible; this is not.
- Test the three signals honestly. Concrete application, finite path, a reason exclusivity is worth paying for.
- Choose the route on exclusivity and on your own appetite, not on which sounds more ambitious.
- Stack non-dilutive funding before you take equity money, and stack it in order of least competitive first.
- Decide patent geography against the market you actually intend to serve, and say so while the filing strategy is open.
- Negotiate performance obligations into any licence. A shelved technology is the quiet failure mode.
- Plan the post-signature years. Name who funds them and who builds, before you sign.
FAQ
Do I own the patent on research I did at a university? Generally not, if university resources or federal funding were involved. Under standard employment agreements and the Bayh-Dole framework, the institution owns inventions arising from work in the lab. You receive a share of licensing income, commonly 30 to 50 percent of net, but you do not have unilateral authority over the invention.
I already published. Is it too late to patent? In the United States you have 12 months from public disclosure to file. In most other countries international patent rights ended the moment you published. Speed matters, so contact the office immediately rather than assuming the option is gone. The standard preventative measure is a provisional application filed before publication, which establishes a priority date and buys 12 months.
How long does it take to reach royalty income? Roughly years 0 to 1 for patent filing and marketing to potential licensees, 1 to 2 for term sheet negotiation and licence execution, 2 to 5 for the licensee's development work, and 5 to 10 before first commercial sales. Most major institutions commit to an initial assessment within 60 to 90 days of your disclosure.
Can I commercialize my invention without going through the technology transfer office? Generally no, and attempting it is usually a contractual breach with serious consequences. Even where the office is slow, the legal structure makes it the institution's invention to license. Engage rather than circumvent.
Which route pays the inventor most? Spinning out, in the rare cases it works, because you hold equity rather than a royalty share. Licensing pays less but far more reliably. Assignment pays once. Rank them by risk tolerance and by whether you want an operating job, not by headline outcome.