MFN & Anti-Stacking Clauses in Patent Licenses

A university or research institution negotiating an exclusive patent license rarely faces just one hard-fought term. Two provisions come up disproportionately often once a licensee has real leverage — a most-favored-nation (MFN) clause, which ties the licensee’s terms to whatever better deal the licensor might give someone else later, and an anti-stacking clause, which protects the licensee from paying an unsustainable combined royalty burden when a product requires more than one licensed technology. Both are licensee-protective provisions that a technology licensing office (TLO/TTO) has to evaluate carefully, because both can quietly erode the institution’s return on a licensed invention years after signature.

CASRAI’s Biotech/Pharma Licensing vs. Standard Tech Licensing Deal Structures comparison introduces royalty stacking and anti-stacking mechanisms briefly, as one of three dimensions distinguishing biotech/pharma deal shapes from standard technology licenses. This guide goes deeper on both anti-stacking clauses and the related, less commonly discussed MFN clause specifically — what each provision actually does, why licensees ask for them, why university licensors often resist or limit them, and the negotiation dynamics that typically shape how they end up drafted.

What a Most-Favored-Nation (MFN) Clause Does

A most-favored-nation clause (sometimes called a “most-favored-licensee” clause) is a contractual promise that if the licensor later grants a different licensee more favorable terms for comparable rights, the original licensee becomes entitled to those better terms too — either automatically or by election, depending on how the clause is drafted. The concept is borrowed from international trade law, where an MFN commitment between countries guarantees that neither party will be treated worse than the “most favored” trading partner; the licensing version applies the same logic to a private contract between a licensor and a licensee.

In university patent licensing, MFN provisions appear most often in non-exclusive licenses, where the licensor is expected to grant the same or similar rights to other parties over time and an early licensee wants assurance it isn’t disadvantaged relative to a competitor who licenses the same technology later on better terms. They can also appear in exclusive licenses as a narrower assurance tied to specific deal terms (for example, a right of first negotiation or first refusal on improvements), though a true MFN-on-royalty-rate provision is less common once exclusivity is already granted, since exclusivity itself is the licensee’s primary protection against a competing licensee.

Why Licensees Want an MFN Clause

The commercial logic is straightforward: a licensee that commits significant capital to commercializing a licensed technology wants confidence it isn’t paying more, or accepting worse terms, than a competitor who licenses the same underlying IP later — whether because the licensor grew more experienced at pricing the technology, needed the deal to close for programmatic reasons, or simply negotiated with a less sophisticated counterparty. An MFN clause converts that concern into a contractual guarantee rather than leaving it to trust in the licensor’s future consistency.

Why University Licensors Resist or Narrow MFN Clauses

Technology licensing offices generally do not refuse MFN language outright, but they routinely push to narrow its scope, because an unqualified MFN commitment creates real administrative and strategic problems for the institution:

Defining “more favorable” is genuinely hard. A later license might have a lower royalty rate but a larger upfront fee, a narrower field of use, different milestone triggers, or different diligence obligations. Comparing two license agreements holistically, rather than term-by-term, is a real drafting and interpretive problem — licensing-practice commentary on preferential-rights clauses (see the Licensing Executives Society International’s coverage of preferential rights in IP agreements) treats this ambiguity as one of the clause’s chief practical risks, and TTOs typically respond by defining comparison narrowly (e.g., “royalty rate only,” or “terms for substantially the same field of use and exclusivity scope”) rather than agreeing to compare agreements as a whole.
It constrains future deal-making. A university licensing the same platform technology into multiple, genuinely different fields of use or markets may have good reasons to offer different economic terms to different licensees — an MFN clause that isn’t carefully scoped to “the same field of use” can force the institution to either extend better terms it didn’t intend to extend, or avoid deals it would otherwise want to make.
It creates an ongoing monitoring and disclosure burden. Honoring an MFN clause requires the licensing office to track every subsequent license for the same technology and assess, deal by deal, whether it triggers the clause — without breaching the confidentiality obligations owed to those other licensees. This is nontrivial administrative overhead that a TTO with a small staff has to weigh against the value of the deal.
It can suppress the value of later deals. If a licensor knows granting slightly better terms to a strategically important second licensee will retroactively hand the same terms to the first, it has a real incentive to avoid improving terms even when a later deal justifies it — flattening pricing across the technology’s licensing history in a way that isn’t necessarily in the institution’s interest.

In practice, MFN language that does survive university review is usually narrowed in one or more ways: limited to a defined time window (e.g., MFN rights that sunset after a fixed number of years or after the licensee hits a diligence milestone), limited to a specific comparison term (royalty rate rather than the whole agreement), limited to licenses for substantially the same field of use and scope of exclusivity, or structured as a right to be notified and negotiate rather than an automatic entitlement to match.

What an Anti-Stacking Clause Does

Royalty stacking is what happens when a single commercial product requires rights under more than one licensed patent or technology, each owed to a different licensor — a combination therapeutic that embeds a licensed delivery platform, a licensed cell line, and a licensed formulation patent is a common example, but the same dynamic shows up in software (multiple licensed components in one product) and hardware (multiple standard-essential or component patents in one device). Each individual royalty obligation may be reasonable in isolation; combined, they can erode the product’s margin to the point of making it commercially unviable — a risk that falls on the licensee, not any single licensor, since no individual licensor bears the cumulative effect of the others’ royalty rates.

An anti-stacking clause is the licensee’s contractual protection against that cumulative burden. Licensing practice generally uses two related mechanisms, both already introduced briefly in CASRAI’s biotech/pharma vs. standard tech licensing comparison:

Royalty offset (anti-stacking) provisions — allow the licensee to reduce the royalty owed to one licensor by some percentage of the royalties it separately owes third-party licensors for other IP necessary to make, use, or sell the product. Secondary licensing-practice sources describe a common structure where the licensee can offset roughly 50% of third-party royalty payments against the amount owed to the academic licensor, subject to a floor below which the primary royalty cannot be reduced further — the exact percentage and floor are heavily negotiated and vary deal to deal, so treat any specific number as illustrative of typical practice, not a fixed convention.
Stacking caps (global royalty ceilings) — set a maximum combined royalty burden across all licensed inputs in the product. If the sum of all royalty obligations would exceed the ceiling, a pro-rata reduction mechanism is triggered so that the licensors collectively share a smaller pie rather than the product becoming structurally unprofitable to sell.

What Gets Negotiated Within an Anti-Stacking Clause

The headline concept — “the licensee gets some protection against cumulative royalties” — is the easy part to agree on. What actually gets fought over in drafting is narrower and more consequential:

What counts as a “necessary” third-party license. Licensors generally want the offset limited to patents the licensee is legally required to license to avoid infringement (freedom-to-operate necessity). Licensees often push to include any IP that is “reasonably useful” to commercializing the product, even where a non-infringing workaround theoretically exists — a materially broader category that increases how often the offset applies.
The offset percentage and the floor. A licensor wants both the percentage offset and the floor (the minimum royalty that survives regardless of stacking) set high enough that the deal still returns meaningful value even in a worst-case stacking scenario; a licensee wants both set to genuinely absorb the risk of an unpredictable future stacking situation it can’t fully model at signature.
Whether the mechanism is bilateral or one-sided. Some anti-stacking language is symmetric across all of a product’s licensors (each accepts a pro-rata reduction under the global-cap approach); other language protects the licensee against just one specific licensor’s royalty, leaving the licensee to negotiate separate protection with every other licensor individually — a materially different risk allocation.
Disclosure obligations. To invoke an offset, a licensee typically has to disclose the existence and terms of the third-party license(s) triggering it, which raises the same confidentiality tension that MFN monitoring does, just from the licensee’s side of the relationship this time.

Academic economic literature on the topic (Economics Letters published a 2022 analysis, “On the inefficiencies of anti-stacking royalty clauses,” examining how these provisions affect licensing outcomes) treats anti-stacking clauses as a real, studied mechanism with genuine efficiency trade-offs rather than a purely licensee-favorable device — worth knowing if a TTO negotiator is asked to justify a position on one internally, since “this only helps the licensee” is not the settled economic consensus.

Why These Two Provisions Cluster in the Same Deals

MFN and anti-stacking clauses aren’t the same mechanism, but they tend to surface together in the same category of license: platform technologies and enabling technologies licensed to more than one party, and complex products (combination therapeutics, multi-component devices, layered software stacks) that realistically require several independently owned licenses to bring to market. A licensee facing genuine uncertainty about both “will I get the best deal anyone gets” (MFN) and “will my total royalty burden across every input stay commercially sane” (anti-stacking) is often the same sophisticated, well-advised counterparty — typically an industry partner or a well-funded startup working with outside licensing counsel, rather than a first-time academic spinout negotiating its own founding license.

Practical Considerations for a University Technology Licensing Office

General principles that recur across licensing-practice literature and TTO guidance, offered here as considerations rather than a fixed playbook — every institution’s actual position depends on its own policies, the specific technology, and the counterparty:

Scope MFN language narrowly and specifically (defined comparison term, defined field of use, defined time window) rather than agreeing to an open-ended “most favorable terms overall” commitment.
Define “necessary” IP for anti-stacking purposes as tightly as the licensee will accept — freedom-to-operate necessity is a materially different (and lower) exposure than “reasonably useful” IP.
Set an offset floor that protects a minimum royalty return regardless of how much stacking ultimately occurs, so the institution isn’t exposed to a scenario where its royalty is effectively negotiated away by licenses it isn’t even a party to.
Coordinate with the university’s other outstanding licenses on the same technology before agreeing to MFN language — the licensing office needs to know what it’s actually promising to compare against.
Treat both provisions as connected to the broader deal structure, not isolated boilerplate — how a technology is priced and staged (see CASRAI’s guide on how TTOs evaluate and price a license and on royalty rate–setting methodology) directly affects how much exposure an MFN or anti-stacking clause actually creates.

As with any specific license term, institutional policy and legal counsel — not general guidance like this page — should govern the actual language used in a given agreement.

Frequently Asked Questions

Is an MFN clause the same thing as an anti-stacking clause?

No. An MFN clause compares one licensee’s terms against terms the same licensor grants to a different licensee for the same or similar technology. An anti-stacking clause addresses a completely different problem: the cumulative royalty burden one licensee faces from multiple, unrelated licensors whose IP is all needed for the same product. They’re both licensee-protective, but they solve different risks and are negotiated separately, even though they often appear in the same category of deal.

Do most university licenses include an MFN clause?

No — MFN clauses are not a default or universal term in university patent licensing. They appear more often in non-exclusive licenses, where multiple licensees for the same technology are expected, and in deals with sophisticated, well-resourced licensees who specifically negotiate for one. Many university licenses, especially exclusive licenses to a single field of use, do not include MFN language at all.

Why would a university ever agree to an anti-stacking clause?

Because refusing one doesn’t eliminate the underlying risk — it just shifts where the pressure lands. If a licensee’s product genuinely can’t absorb the combined royalty burden of every license it needs, an unyielding royalty term doesn’t produce a healthier deal; it can produce a product that never launches, generating no royalty income for anyone. A negotiated, capped anti-stacking mechanism is often a more realistic way to preserve some return than holding a royalty rate that assumes the licensee is the only IP owner in the product.

What’s the difference between an anti-stacking offset and a stacking cap?

An offset reduces what’s owed to one specific licensor based on what the licensee separately pays other licensors, usually with a floor limiting how far it can be reduced. A stacking cap instead sets a ceiling on the combined royalty burden across every licensed input in the product, triggering a pro-rata reduction across all licensors if the ceiling would otherwise be exceeded. Some agreements use one mechanism, some use both together.

Most-Favored-Nation and Anti-Stacking Clauses in Patent Licenses