Patent Licensing and Royalties Explained

A patent is fundamentally a right to exclude — but that right is worth little if you never turn it into money. Licensing is how most patents actually pay, letting you earn from an invention you may lack the capital, factory, or sales force to commercialize yourself.

How does patent licensing actually work?

A patent grants its owner the right to stop others from making, using, selling, offering to sell, or importing the claimed invention for its term (generally 20 years from the earliest non-provisional filing date, under 35 U.S.C. § 154). A license is simply the owner agreeing not to enforce that right against a particular party, on agreed terms.

Nothing about ownership changes. Unlike an assignment — an outright sale that transfers title to the patent — a license leaves you as the owner and creates a contractual permission you can shape, limit, and renew. That distinction drives everything downstream, from taxes to who can sue an infringer. If you’re weighing licensing against selling the patent outright, our companion guide on how to license or sell a patent walks through the decision in depth.

Licensing appeals to inventors and companies who want revenue without the cost and risk of building a product, and to operating companies who want to add a proven technology without inventing it from scratch. It also spreads risk: instead of betting on a single product line, you can put the same invention to work across many markets at once. Done well, it converts a legal right into a recurring income stream — and a well-drafted license can outlive the founder, survive an acquisition, and be assigned or securitized like any other asset.

Exclusive vs. non-exclusive vs. sole license: what’s the difference?

The scope of exclusivity is the single biggest lever in any patent deal. There are three basic structures:

• Exclusive license. Only one licensee may practice the patent within the defined scope — and, depending on the drafting, even the patent owner may be excluded. Because the licensee gets a monopoly, exclusive licenses command the highest royalties and the largest up-front payments. A licensee holding “all substantial rights” may even sue infringers in its own name.
• Non-exclusive license. The owner can license the same patent to as many parties as it wants. Each licensee pays less, but the owner can build volume across many deals. This is the norm for standardized or widely used technologies.
• Sole license. A middle ground: the owner grants rights to one licensee but also keeps the right to practice the patent itself. No third parties, but the licensee isn’t the only user.

Exclusivity can — and usually should — be narrowed by field of use, territory, and time, so you can grant an exclusive license in, say, medical devices in North America while licensing the same patent non-exclusively elsewhere. One practical warning: the word “exclusive” on its own is ambiguous. Spell out precisely whether the patent owner retains the right to practice the invention, whether the exclusivity is capped to a field or region, and whether the licensee can enforce the patent — courts have repeatedly split hairs over exactly these questions when a licensee tries to sue an infringer.

What royalty structures are available?

“Royalty” is shorthand for several very different payment mechanics, and most sophisticated deals combine them:

• Running royalty (percentage). The licensee pays a percentage of net sales of licensed products. This is the most common structure and aligns both sides with commercial success. Nail down exactly what “net sales” means — deductions for returns, shipping, and taxes materially change the math.
• Per-unit royalty. A fixed dollar amount for each unit sold (e.g., $2 per device). Simple to audit and predictable, useful when unit prices swing.
• Lump-sum / paid-up license. A one-time payment for the life of the license. It removes ongoing accounting but shifts the risk of over- or under-estimating volume.
• Up-front fee plus running royalty. A signing payment (to compensate for access and reduce the licensor’s risk) layered on top of ongoing royalties.
• Milestone payments. Payments triggered by events — regulatory approval, first commercial sale, hitting a revenue threshold — very common in pharma and biotech deals.
• Minimum annual royalties. A floor that guarantees income and pressures an exclusive licensee to actually work the patent rather than sit on it.

There’s no universally “right” structure. The best deals match the payment model to how the technology makes money and to each side’s tolerance for risk.

What is a typical patent royalty rate?

Everyone wants the magic number, and there isn’t one — but there are useful ranges. Running royalties most often fall between 2% and 10% of net sales, with a large share of technology deals clustering around 3% to 6%. Industry matters enormously:

• Pharmaceuticals and biotech: frequently 5% to 15%+, reflecting huge margins and high R&D risk.
• Medical devices: often 5% to 10%.
• Software: commonly 5% to 15% depending on how central the patent is.
• Consumer electronics and low-margin hardware: often 1% to 5%.

Rate drivers include the patent’s strength and breadth — a broad, clearly patentable claim that is hard to design around is worth more — whether it’s exclusive, how much profit the invention actually enables, the existence of non-infringing alternatives, and raw bargaining power. To ground these ranges in your specific asset, see our valuation guide on what is my patent worth, which covers the income, market, and cost approaches to putting a number on a patent.

Do the 25% rule and Georgia-Pacific factors still set royalties?

Two frameworks dominate royalty discussions, and you should understand both — including the limits of one.

The 25% rule of thumb held that a licensee should pay roughly 25% of the expected profit from the licensed product as a royalty, with the licensor keeping a quarter and the licensee three-quarters. It was popular for decades because it was simple. But in Uniloc USA v. Microsoft (Fed. Cir. 2011), the Federal Circuit rejected the 25% rule as a legal basis for calculating patent-infringement damages, calling it a “fundamentally flawed” tool untethered to the facts of a case. Negotiators still use it privately as a rough sanity check, but you cannot lean on it in court.

The Georgia-Pacific factors are the enduring standard. From Georgia-Pacific Corp. v. United States Plywood Corp. (S.D.N.Y. 1970), these 15 factors guide how U.S. courts construct a “reasonable royalty” — the statutory floor for infringement damages under 35 U.S.C. § 284. They include established royalties the patentee has charged, rates for comparable patents, the exclusivity and scope of the license, the commercial relationship of the parties, the profitability and advantages of the invention, and the hypothetical arm’s-length negotiation the parties would have had just before infringement began. Because litigation royalties are built on these factors, they heavily shape voluntary deals too — smart negotiators bargain in the shadow of what a court would award. These same principles underpin the damages analysis in our overview of patent enforcement and monetization.

Which license terms matter most beyond the money?

The royalty gets the attention, but these clauses often decide whether a deal is good or a trap:

• Field of use. Limits the license to a specific application or market (e.g., “automotive sensors only”). Lets you slice one patent into multiple non-competing deals.
• Territory. Geographic scope. You can license the U.S. exclusively and Europe to someone else.
• Sublicensing. Whether the licensee can grant rights to third parties — and how any sublicense revenue is shared with you. Absent express permission, licensees generally cannot sublicense.
• Grant-backs. Whether improvements the licensee develops flow back to you (as a license or assignment). Overbroad grant-backs can raise antitrust concerns, so they need careful drafting.
• Audit rights. The right to inspect the licensee’s books to verify royalty reports. Without this, running royalties are effectively unenforceable — build in audit access, reporting cadence, and interest on underpayments.
• Improvements and new patents. Does the license cover future patents in the family, or only the listed ones?
• Term, termination, and defaults. When it ends, what triggers termination, and what happens to inventory and sublicenses on exit.
• Representations, indemnities, and validity. Note that under Lear v. Adkins (1969) and MedImmune v. Genentech (2007), a licensee can often challenge a patent’s validity while continuing to pay — so “no-challenge” clauses are limited.

What are standard-essential patents and FRAND?

Some patents are unavoidable if you want to build a compliant product. A standard-essential patent (SEP) is one that must be practiced to comply with an industry technical standard — think 5G, Wi-Fi, LTE, USB, or the H.264/H.265 video codecs. You literally cannot make a standards-compliant phone without touching thousands of SEPs.

To keep standards from becoming hostage situations, standards-setting organizations (like ETSI, IEEE, and ITU) require members to commit their SEPs to licensing on FRAND terms — fair, reasonable, and non-discriminatory. That commitment is a contractual promise, generally enforceable by any implementer as a third-party beneficiary, and it means the patent owner:

• Must license to any willing licensee, rather than refusing or picking favorites;
• Cannot charge excessive or discriminatory royalties; and
• Faces real limits on getting injunctions against a willing licensee.

Courts and regulators worldwide police the balance. Landmark decisions like Microsoft v. Motorola (9th Cir. 2015) in the U.S. and Unwired Planet v. Huawei (UK Supreme Court 2020) have set global FRAND royalty rates, and antitrust authorities scrutinize both “hold-up” (owners over-charging) and “hold-out” (implementers refusing to pay). If your technology reads on a standard, FRAND is not optional background — it reshapes what you can charge and how you enforce.

The bottom line

Licensing is how patents earn their keep. Decide first how much exclusivity to grant — exclusive, sole, or non-exclusive — because that choice drives price and enforcement rights. Then match the payment structure (running royalty, per-unit, lump sum, milestones, or a blend) to how the invention actually makes money, and benchmark your rate against real industry ranges rather than the discredited 25% rule, keeping the Georgia-Pacific factors in view because a court would. Finally, treat field of use, territory, sublicensing, grant-backs, and audit rights as first-class terms, and know whether FRAND applies before you name a number. For the strategic view of monetization and to study how these deals fare in litigation, see our patent enforcement and monetization hub and the patent case analysis archive.

This guide is general education, not legal advice, and does not create an attorney-client relationship. Royalty rates, FRAND obligations, and license enforceability turn on your specific technology, industry, and jurisdiction — consult an attorney licensed in your jurisdiction before signing or enforcing a patent license.