FRAND Commitments, Injunctions, and Rate Setting: How Standard-Essential Patents Get Priced

Document type: Article Practice area: Intellectual Property — Patents Jurisdiction: United States, with international context Last reviewed: 5 September 2026

Consider what a FRAND commitment actually is.

A company participating in a standards-setting organization promises that if any of its patents turn out to be essential to the standard, it will license them on fair, reasonable, and non-discriminatory terms. The promise is made before the standard is finished, about patents that may not have issued, to licensees who do not yet exist, at a price nobody has negotiated.

Then the standard is adopted, billions of devices implement it, and the parties discover they disagree about what "fair and reasonable" means by a factor of fifty.

That is the entire subject. Everything else — the injunction fights, the rate-setting trials, the antitrust claims, the jurisdictional races across three continents — is a consequence of pricing a promise that was deliberately left unpriced because pricing it in advance would have been a cartel.

The commitment is not a drafting failure. It is a bargain: implementers get certainty that the technology will be available; patent holders get their technology adopted into a standard that guarantees a market. What was traded away was the ordinary mechanism by which a price gets set, and the litigation is the replacement mechanism.

What the commitment is

A contract, in most analyses. A patent holder's declaration to a standards-setting organization creates a contractual obligation, governed by the SSO's intellectual property rights policy and by the law that policy selects — French law for ETSI, for example, which matters more than practitioners often expect.

Implementers are third-party beneficiaries. United States courts have generally held that an implementer may enforce the commitment as a third-party beneficiary of the contract between the patent holder and the SSO. That is the doctrinal foundation for most FRAND litigation in this country.

What the commitment does not do:

It does not set a rate.
It does not identify licensees.
It does not define essentiality.
It does not say whether the license must be portfolio-wide or patent-by-patent.
It does not say what royalty base applies.
It does not say whether an injunction is available.

Every one of those gaps has been litigated, in multiple jurisdictions, with different answers.

What the SSO policies do say, in broad terms: participants must disclose patents they believe may be essential, and must commit to license essential patents on FRAND terms or state that they will not. Policies differ on whether disclosure is mandatory or best-efforts, on the scope of the licensing commitment, and on whether the commitment binds transferees — which is why the identity of the SSO matters in every case.

Hold-up and hold-out

The framework exists to solve a problem, and the problem runs in both directions. A practitioner who describes only one half is describing a client's position rather than the law.

Hold-up

Once a standard is adopted, implementers are locked in. Redesigning around a patent essential to 5G is not a commercial option; the product would not work.

The consequence. A patent holder can demand a royalty reflecting not the technology's value but the cost of abandoning the standard. That is hold-up, and the FRAND commitment is the mechanism intended to prevent it.

Royalty stacking is the aggregate version. A standard may read on thousands of declared-essential patents held by dozens of owners. If each demands a royalty that looks individually reasonable, the total can exceed the device's price. The theoretical problem is real; whether it occurs in practice at the levels claimed is empirically contested.

Hold-out

The mirror image, and it took longer to be recognized.

The mechanism. An implementer can use the technology, decline to take a license, and litigate for years. Because injunctions are difficult to obtain against a FRAND-committed patent, and because the worst outcome of losing is paying the royalty it would have owed anyway, the implementer's incentive is to delay. Money later is worth less than money now, litigation is expensive for the patent holder too, and some implementers simply never pay.

Why it matters doctrinally. A rule that no injunction is ever available against an implementer of a FRAND-committed patent converts the patent into a compulsory license at a rate set by a court years later. Courts and agencies have recognized this, which is why the modern framework asks whether the implementer is willing rather than whether the patent is standard-essential.

The balanced view. Both problems are real. The doctrine has converged on a negotiation framework that assigns obligations to both sides and makes remedies turn on which side failed to meet them.

Injunctions

The most contested question, and the one where the answer differs most across jurisdictions.

The United States position. eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006) eliminated any presumption of injunctive relief in patent cases, requiring the traditional four-factor test: irreparable injury, inadequacy of legal remedies, the balance of hardships, and the public interest.

Applied to a FRAND-committed patent, the factors line up badly for the patent holder:

Irreparable harm is difficult to show when the patent holder has promised to license anyone on reasonable terms. A party that has committed to accept money has some difficulty arguing money is inadequate.
Adequacy of legal remedies points the same way, for the same reason.
The balance of hardships weighs an injunction that would exclude a product implementing an entire standard against a royalty dispute.
The public interest implicates the standards system itself and the availability of standardized products.

The result in practice. Injunctions on FRAND-committed patents are rare in United States district courts, and agencies have expressed the view that they should be available only in limited circumstances — most importantly where the implementer is unwilling, refuses to take a license on any terms, or is beyond the court's jurisdiction.

The ITC is different, and it matters. Section 337 investigations produce exclusion orders rather than damages, and the Commission does not apply the eBay factors — it applies statutory public interest factors. A patent holder denied an injunction in district court may still obtain an exclusion order at the ITC, which is why SEP holders file there and why the public interest submissions in those investigations are heavily contested.

Foreign jurisdictions diverge sharply. Injunctions are more readily available in some European jurisdictions and in China, subject to frameworks that ask whether the implementer negotiated in good faith. The result is a global race: a patent holder seeks an injunction where injunctions issue, an implementer seeks a rate determination where rates are set, and both seek anti-suit and anti-anti-suit injunctions to stop the other. This is now a defining feature of major SEP disputes and it is genuinely disorderly.

The negotiation framework

Because remedies turn on whether the implementer is willing, courts and agencies have converged on a structured negotiation sequence. The European Court of Justice's framework in the Huawei v. ZTE line is the most influential formulation, and United States courts have applied comparable reasoning without adopting it wholesale.

The sequence:

1. The patent holder gives notice. It identifies the patents alleged to be essential and the standard, and explains how they are infringed. A demand letter that names a portfolio without claim charts is not notice in any meaningful sense.

2. The implementer expresses willingness to take a license on FRAND terms. This is the crucial step, and an implementer that does not take it — that ignores the notice, or responds with a general denial — risks being characterized as unwilling.

3. The patent holder makes a written offer specifying the royalty and the basis for calculating it, with an explanation.

4. The implementer responds promptly with a counter-offer if it disagrees, again in writing and with a rationale.

5. If the offer is rejected, the implementer provides security and renders an account of past use.

6. If the parties cannot agree, either may seek a determination of FRAND terms by a court or by arbitration.

What each side must actually do:

Patent holder: claim charts, an essentiality showing, a specific rate, the basis for the rate, comparable licenses if it relies on them, and a response to counter-offers within a reasonable time.

Implementer: an actual response, a counter-offer with a rationale, technical engagement with the charts, and security if it continues to practice the standard while disputing the rate.

What makes a party "unwilling." Not disagreeing about the rate — disagreement is the normal case. Unwillingness looks like: no substantive response over many months; a counter-offer of zero or a nominal amount without justification; refusal to enter a confidentiality agreement necessary to see comparable licenses; refusal to be bound by any court's determination; and continuing to sell while providing no security.

Why this framework matters more than the doctrine. In most disputes, the ultimate rate is negotiated. What the litigation determines is which side's conduct was unreasonable, and that determination shapes the settlement far more than any hypothetical injunction.

Valuation: two methodologies

When a court sets a rate, it uses one or both of two approaches. Understanding them is the substance of SEP practice.

Comparable licenses

The approach. Identify existing licenses to the same or similar portfolios and use them to derive a rate.

Why it dominates. Real-world licenses reflect what parties actually paid, which is the best available evidence of value. Courts prefer it when good comparables exist.

Why it is difficult:

Portfolio licenses cover more than the patents in suit — often thousands of patents, many not essential, across many standards and jurisdictions. Extracting a rate for the asserted patents requires apportionment.
Cross-licenses have no cash rate or a net balancing payment that reflects the relative strength of two portfolios rather than the value of either.
Lump-sum licenses must be converted to a running rate, requiring assumptions about volume.
Litigation settlements may reflect the cost of litigation rather than the value of the technology, and courts discount them accordingly.
Non-discrimination. Existing licenses at materially different rates to similarly situated licensees raise a discrimination question, which is the "ND" in FRAND and the least developed of the three letters.

What makes a comparable persuasive: a license to a similar implementer, for a similar scope, negotiated at arm's length without litigation pressure, close in time, with a cash rate.

Top-down apportionment

The approach. Determine the aggregate royalty that all essential patents on the standard should command, then allocate a share to the patents in suit based on their proportion of the total essential patents, adjusted for relative value.

The steps:

The aggregate royalty burden. What should all SEPs on this standard collectively earn? Evidence includes public statements by SSO participants about expected aggregate rates, industry studies, and the economics of the product.
The denominator. How many patents are actually essential? Declared-essential counts vastly overstate essentiality — studies of declared portfolios have found essentiality rates well below half, and sometimes far lower. Somebody must do the work of estimating.
The numerator. How many of the patent holder's patents are actually essential?
The share. The ratio, adjusted for relative technical importance rather than treated as a pure patent count.

Why it appeals. It addresses royalty stacking directly and it produces a number that can be sanity-checked against the product's economics.

Why it is difficult. Every step requires an estimate that is itself contested, and the aggregate royalty figure is frequently drawn from statements made for other purposes.

In practice, courts use both — comparables as the primary evidence and a top-down analysis as a cross-check, or vice versa. A party presenting only one and dismissing the other is usually less persuasive than one that reconciles them.

The royalty base

The most technically contested question in SEP damages, and the one with the largest dollar consequence.

The competing positions:

Patent holders generally argue for the end-user device — the handset, the vehicle, the connected appliance — on the ground that the standardized functionality contributes to the value of the whole product and that industry licensing has historically been at the device level.

Implementers generally argue for the smallest salable patent-practicing unit — the baseband chip, the modem module — on the ground that Federal Circuit apportionment law requires damages to reflect the value of the patented feature rather than the unpatented components with which it is combined.

The doctrinal anchor. The entire market value rule permits use of the whole product as the base only where the patented feature drives demand for the product. Where it does not, apportionment is required — either by using a smaller base or by applying a rate that accounts for the patented feature's contribution.

The practical resolution. The base and the rate are two variables producing one number, and courts have accepted that a device-level base with a very small rate can be equivalent to a component-level base with a larger rate. What courts have not accepted is a device-level base with a rate derived from component-level value, or evidence that presents a large base to a jury without apportionment — the risk that apportionment doctrine exists to prevent.

Where the dispute is live. The automotive supply chain is the current flashpoint: whether SEP holders may license the vehicle manufacturer or must license the component supplier that actually implements the standard. This is a question about licensing level, not just about the base, and it has been litigated on multiple continents with different answers.

Essentiality

An assumption underlies every SEP dispute, and it is usually wrong.

Declared-essential is not essential. SSO policies require participants to declare patents they believe may be essential, and the incentives favor over-declaration: declaring costs little, failing to declare risks losing the ability to enforce.

The consequence. Declared-essential portfolios contain a large proportion of patents that are not essential — because the claims do not read on the standard, because the relevant standard section is optional, because the patent covers an implementation choice rather than a requirement, or because the patent is invalid.

Why this matters:

Top-down denominators built on declared counts overstate the number of essential patents and understate each patent's share.
Portfolio valuations premised on declared counts are unreliable.
An implementer's best technical defense is frequently essentiality, not invalidity. If the claim does not read on the standard, practicing the standard does not infringe — and the patent holder must then prove infringement the ordinary way, product by product.

What essentiality analysis requires. A claim-by-claim mapping to specific sections of the standard, with attention to whether the section is mandatory or optional and whether the claim requires more than the standard does. It is expensive and it is the work that distinguishes a serious SEP position from a declared-portfolio assertion.

Antitrust and the FRAND commitment

Whether breaching a FRAND commitment is an antitrust violation has been a moving target, and the answer matters because the remedies differ enormously.

The theories that have been advanced:

Deception of the standards body. A participant that conceals essential patents during standard development, allowing the standard to be adopted, and then asserts them — this is the strongest theory and it has succeeded. The conduct is the deception, not the later assertion.
Reneging on the commitment. A patent holder that made a FRAND commitment to induce adoption and then demands non-FRAND terms. Courts have divided on whether this states a monopolization claim under 15 U.S.C. § 2 or is simply a breach of contract.
Refusal to license. A refusal to license a competitor at any price, where the commitment required licensing to all comers.
Discriminatory licensing. Charging different rates to similarly situated licensees — the "ND" theory, under § 2 or under 15 U.S.C. § 1.

The countervailing view, which agencies have at times embraced: a FRAND commitment is a contract, breach of contract is not an antitrust violation, unilateral refusal to license a patent is generally lawful, and treating rate disputes as antitrust matters over-deters the standards participation that makes standards possible.

Agency positions have shifted, more than once, on whether SEP-related conduct warrants antitrust enforcement and on whether injunctions on FRAND-committed patents raise competition concerns. Statements and policy documents in this area should be checked for current status before being relied on.

The practical guidance. Plead breach of contract first; it is the securest theory. Add antitrust claims where the facts support deception of the SSO or a genuine refusal to deal, and recognize that a rate disagreement dressed as an antitrust claim is usually dismissed.

And on the other side: a patent holder should assume that aggressive conduct — refusing to negotiate, demanding a rate wildly out of line with comparables, seeking injunctions against a demonstrably willing licensee — creates antitrust exposure it can avoid by following the negotiation framework.

The global dimension

Major SEP disputes are now multi-jurisdictional by default, and the procedural race is a substantive part of the strategy.

What each side wants, and where:

Patent holders want an injunction, in a jurisdiction where injunctions issue and cases move quickly.

Implementers want a global rate determination, in a jurisdiction willing to set worldwide terms — which converts the dispute into a licensing question and removes the injunction threat.

The escalation that follows. A party facing an injunction abroad seeks an anti-suit injunction at home, ordering its counterparty not to pursue the foreign action. The counterparty seeks an anti-anti-suit injunction. Courts have issued both, and the resulting conflicts have drawn diplomatic attention and complaints to international trade bodies.

Whether a court can set a global rate is itself disputed. Some courts have held that where a party has committed to FRAND terms, a court determining those terms may set them for a worldwide portfolio, and that a party refusing to accept a global determination may be an unwilling licensee. Others have declined to reach beyond their own territory.

Practical consequences for counsel:

Forum selection is the first strategic decision, and it is made by whoever files first.
Filing sequence matters enormously. An action filed a week earlier can determine which court sets the terms.
Coordinate across jurisdictions from the outset, because positions taken in one forum are exhibits in another.
Assess anti-suit exposure before filing anywhere.
Expect the case to settle on a global portfolio license, because that is what both sides ultimately need, and the litigation is about the number.

A worked negotiation

Thorncroft Wireless holds a portfolio it declares essential to a cellular standard — roughly 1,200 declared-essential families. Marchbank Devices manufactures connected industrial sensors, about 4 million units a year at an average selling price of $85, each containing a $9 cellular module purchased from a third-party supplier.

Month 0. Thorncroft sends a letter: it holds essential patents, Marchbank's products implement the standard, and it seeks 1.2 percent of the device price — about $0.90 per unit, or $3.6 million a year.

What Thorncroft's letter did not contain: claim charts, an essentiality analysis, identification of specific patents, or an explanation of the rate.

Marchbank's counsel, Ndidi Vasquez-Thornbury, responds within three weeks. The response:

States willingness to take a license on FRAND terms. This sentence is the most important one in the letter, and it is free.
Requests claim charts for representative patents and an essentiality showing.
Requests the basis for the 1.2 percent rate and any comparable licenses, offering to enter a confidentiality agreement.
Notes that Marchbank's supplier may already be licensed, and asks whether Thorncroft licenses at the component level.
Does not deny infringement generally or dispute essentiality without analysis.

Month 3. Thorncroft provides charts for eight patents and a list of licensees. Two of the eight charts map to optional sections of the standard.

Months 3–8 — the technical work. Marchbank's analysis: of the eight charted patents, three are arguably essential, two map to optional features Marchbank does not implement, two do not read on the standard as written, and one is likely invalid over prior art. Extrapolating, Marchbank estimates Thorncroft's genuinely essential portfolio at 20 to 30 percent of the declared count.

Month 8 — the counter-offer. Marchbank offers $0.11 per unit — about $440,000 a year — supported by:

A top-down analysis: an aggregate royalty burden for the standard drawn from public statements by major participants; an essentiality-adjusted denominator; Thorncroft's adjusted share.
A comparables analysis from the two publicly reported licenses in the licensee list.
An apportionment argument that the standardized functionality resides in the $9 module, not the $85 device.

Months 8–14. Thorncroft rejects and threatens an ITC filing. Marchbank places $1.1 million in escrow representing its offer for the disputed period — which materially strengthens its willingness position — and proposes binding arbitration of the rate.

Month 16. Thorncroft files at the ITC and in district court. Marchbank counterclaims for breach of the FRAND commitment and seeks a rate determination.

Month 22 — settlement. A portfolio license at $0.19 per unit, plus a release for past sales, with a five-year term and a most-favored-licensee provision.

Why the outcome landed where it did:

Marchbank's early willingness statement and escrow made an unwillingness finding unavailable, which removed the injunction threat that was Thorncroft's leverage.

The essentiality work was the substance. Extrapolating from eight charts to a portfolio estimate is imprecise, and it was the only quantitative analysis in the room.

Thorncroft's opening letter — no charts, no rate explanation — cost it. A patent holder that begins with a supportable position negotiates from a better place than one that begins with a demand.

And the number: $0.19 against a $0.90 demand and a $0.11 offer. That is the ordinary shape of these resolutions, and it is why both sides invest in the analysis rather than in the rhetoric.

What each side should do, and when

The implementer

Before any letter arrives:

Know your standards exposure. Which standards do your products implement, at what level, and through which components? Most companies cannot answer this.
Know your supply chain's licensing status. If your module supplier is licensed, that may substantially reduce or eliminate your exposure — and the answer is in the supply agreement.
Negotiate IP indemnities that reach SEPs. Many standard indemnities carve out patents essential to standards, which is precisely the exposure you need covered. Read the carve-out.
Budget for it. SEP royalties are a cost of implementing a standard, and a company that has never modeled them will be surprised.

On receiving a letter:

Respond within weeks, and state willingness. This single sentence is the difference between a negotiation and an unwillingness finding, and it commits you to nothing about the rate.
Ask for charts, essentiality, the rate basis, and comparables. Offer a confidentiality agreement immediately; refusing one is evidence of unwillingness.
Do not deny generally. A blanket denial without analysis is the response that makes an implementer look unwilling.
Do the essentiality work. It is expensive and it is where the value is.
Make a supported counter-offer, with a methodology.
Consider escrow or security if you continue selling while disputing. It is the strongest available evidence of good faith.
Assess the forum risk early, including anti-suit exposure, before anyone files anywhere.

The patent holder

Before asserting:

Do your own essentiality analysis. Asserting a declared portfolio without knowing what is actually essential produces a weak position that the implementer will expose.
Prepare charts for representative patents and be able to explain the selection.
Develop the rate basis. Comparable licenses, a top-down analysis, or both — with the apportionment reasoning stated.
Confirm your licensing level position. Device or component, and be consistent with your existing licenses.
Check your own license history for discrimination exposure. Different rates to similarly situated licensees is the "ND" problem and it is discoverable.

In the negotiation:

Give real notice. Charts, patents, standard sections, and a rate with an explanation. A demand letter without them starts you behind.
Respond to counter-offers substantively and promptly. Silence is evidence.
Enter the confidentiality agreement so the implementer can see the comparables.
Do not seek an injunction against a demonstrably willing licensee. It rarely works in the United States, it creates antitrust exposure, and it damages the negotiation.
Document every step. The record of who did what and when is what a court assessing willingness will read.

Where the law is unsettled

Whether a court may set a global rate, and whether refusing a global determination makes a licensee unwilling.

Anti-suit and anti-anti-suit injunctions, and whether any principle constrains the escalation.

The licensing level — device versus component — particularly in the automotive supply chain, where different jurisdictions have reached different results.

The scope of the non-discrimination obligation, which is the least developed element of FRAND and the one most likely to produce new doctrine.

Whether and when a FRAND breach is an antitrust violation, on which agency positions have shifted more than once.

Whether the FRAND commitment binds transferees, and what happens when a portfolio is sold to an entity that did not participate in the standard.

The practical response to all of it. The doctrine is unsettled, but the negotiation framework is not. A party that gives real notice, responds substantively, offers a supported rate, and documents the process will be treated as reasonable in any jurisdiction. That is a durable strategy in an unstable area, and it is available to both sides.

Patent pools

A significant fraction of standardized technology is licensed through pools rather than bilaterally, and the mechanism deserves separate treatment because it changes the analysis.

What a pool is. An administrator aggregates patents from multiple holders essential to a standard and offers a single license at a published rate, distributing royalties among contributors according to an agreed formula.

What it solves:

Transaction costs. One negotiation instead of forty.
Royalty stacking. A published aggregate rate is visible in advance and can be budgeted.
Non-discrimination. Published terms available to all comers largely resolve the "ND" problem for pool patents.
Certainty. An implementer taking a pool license knows what it has and what it costs.

Why pools are lawful. Aggregating complementary patents — those that must all be practiced to implement the standard — is procompetitive: it reduces transaction costs and eliminates the double-marginalization problem. Agency guidance has long recognized this, subject to conditions.

The conditions that keep a pool lawful:

Only essential patents. Including non-essential patents converts the pool into a tying arrangement and aggregates substitutes rather than complements. Pools use independent essentiality evaluators for exactly this reason.
Non-exclusive licensing. Contributors remain free to license bilaterally, so the pool does not become the only channel.
No downstream coordination. The pool must not become a mechanism for exchanging competitively sensitive information or coordinating on price in product markets.
Reasonable and non-discriminatory terms, published.
Grantbacks limited to essential patents and non-exclusive.

What a pool license does not do. It covers only the contributors' patents. Major holders frequently stay out of pools and license bilaterally, so an implementer with a pool license may still face several bilateral negotiations. Check the contributor list before assuming coverage — this is the most common misunderstanding about pools.

The practical guidance for implementers: take the pool license where one exists and the rate is reasonable. It is cheaper than bilateral negotiation, it establishes a comparable for the bilateral discussions that follow, and it is strong evidence of willingness.

And for patent holders: joining a pool trades a lower per-patent rate for lower transaction costs and higher collection rates. Whether that trade is worthwhile depends on portfolio strength and on enforcement capacity — a holder with a strong portfolio and the resources to litigate generally does better bilaterally, which is why the strongest portfolios are frequently outside the pools.

Frequently asked questions

Does a FRAND commitment mean we can use the technology without paying? No. It means the patent holder must offer a license on fair, reasonable, and non-discriminatory terms. You still owe a royalty; the dispute is about the number.

Can we be enjoined? In United States district courts, rarely — the eBay factors weigh against injunctive relief for a patent whose owner has promised to license anyone. At the ITC, an exclusion order is a real possibility because the Commission applies statutory public interest factors rather than eBay. Abroad, injunctions are more readily available. The variable that matters most everywhere is whether you are a willing licensee.

What makes us a willing licensee? Responding promptly, stating willingness to take a FRAND license, engaging technically with the charts, making a supported counter-offer, entering confidentiality agreements, and providing security if you continue selling while disputing. Disagreeing about the rate does not make you unwilling; ignoring the patent holder does.

Our component supplier says it is licensed. Are we covered? Possibly, through exhaustion — but only if the license covers the products, the field, and the downstream use. Read the license and the supply agreement. This is the first question to ask and it is frequently dispositive.

Do we have to pay on the price of the whole product? The base and the rate are two variables producing one number. Apportionment doctrine requires that damages reflect the value of the patented feature, not the unpatented components with which it is combined. A device-level base with a very small rate can be equivalent to a component-level base with a larger one; what is not permitted is a large base with a rate derived from component-level value.

Are all declared-essential patents actually essential? No, and not close. Over-declaration is rational under SSO policies, and essentiality studies consistently find that a large share of declared patents are not essential. Essentiality analysis is frequently the implementer's best technical defense.

Should we file first? Forum selection is decided by whoever files, and it is one of the most consequential decisions in a major SEP dispute. Assess anti-suit exposure before filing anywhere, and understand what relief each candidate forum can grant.

Is a FRAND breach an antitrust violation? Sometimes. Deception of the standards body during standard development is the strongest theory. A disagreement about the rate is generally a contract dispute, and courts have divided on the middle ground.

How do these cases end? With a portfolio license on a negotiated rate, in the overwhelming majority of cases. The litigation determines the leverage, and the leverage determines the number.

What is the single highest-value thing an implementer can do? Answer the first letter within three weeks, state willingness, and ask for charts. It costs almost nothing and it removes the patent holder's strongest leverage.

The short version

A FRAND commitment is a contract made to a standards body, enforceable by implementers as third-party beneficiaries, that promises a license on fair, reasonable, and non-discriminatory terms without saying what those terms are.

It exists to prevent hold-up — the leverage a patent holder gains once implementers are locked into a standard — and the framework has since had to account for hold-out, the mirror problem of implementers who use the technology and never pay.

Injunctions are difficult in United States district courts after eBay, available at the ITC on different factors, and more readily granted abroad — which is why major disputes are multi-jurisdictional races with anti-suit injunctions in both directions.

The variable that decides remedies is willingness, assessed through a structured negotiation: real notice with charts, a prompt statement of willingness, a supported offer, a supported counter-offer, and security if the implementer keeps selling.

Rates are set by comparable licenses, top-down apportionment, or both, and every input in each method is contested.

The royalty base and the licensing level — device or component — carry the largest dollar consequence and are the least settled questions in the field.

Declared-essential is not essential, and essentiality analysis is frequently an implementer's best defense and a patent holder's most neglected preparation.

Pools resolve much of this for the patents in them, and cover only the contributors.

Antitrust theories exist and are strongest where a participant deceived the standards body; a rate disagreement is usually a contract case.

And the durable advice in an unsettled area: follow the negotiation framework. A party that gives real notice, engages substantively, offers a supported number, and documents every step will be treated as reasonable in any forum — and reasonableness is what the leverage, and therefore the price, actually turns on.