# How Is a FRAND Royalty Rate Actually Calculated in 2026?

iprs.cloud · September 27, 2026

> What a FRAND Rate Means A FRAND rate is the royalty charged for using one or more patents that the patent owner committed to license on fair...

## What a FRAND Rate Means

A FRAND rate is the royalty charged for using one or more patents that the patent owner committed to license on fair, reasonable, and non-discriminatory terms. Fair usually concerns the burden placed on the implementer, reasonable concerns whether the royalty is proportionate to the patented contribution, and non-discriminatory concerns equal treatment of similarly situated licensees. A FRAND commitment does not itself specify a price, require the lowest-possible rate, or guarantee that every portfolio qualifies. The commitment remains a voluntary contractual assurance under policies such as the IEEE patent policy, but its enforcement can become a complex dispute when essential patents, technical contributions, and comparable licenses are disputed. After the UK Supreme Court’s 2020 Unwired Planet decision, UK rates may also address global FRAND terms, subject to the territorial scope and governing law of the relevant license. In practice, the number is therefore not a tariff printed on the standard.

**Also worth reading:** [How should discount rate adjustments be calculated and applied in intellectual property valuation?](https://iprs.cloud/knowledge/how_should_discount_rate_adjustments_be_calculated_and_applied_in_intellectual_property_valuation.php) · [How Do FRAND Royalty Evidence and SEP Valuation Methods Affect Royalty Rates in 2026?](https://iprs.cloud/knowledge/how_do_frand_royalty_evidence_and_sep_valuation_methods_affect_royalty_rates_in_2026.php) · [How does top-down FRAND royalty calculation work in standard essential patent licensing?](https://iprs.cloud/knowledge/how_does_top-down_frand_royalty_calculation_work_in_standard_essential_patent_licensing.php)

A defensible calculation must separate several questions: which patent rights are asserted, what value the patented technology contributes to a particular product, which royalty base best represents that value, and what rate produces a fair return if applied to that base. Courts and arbitrators may also examine the parties’ bargaining positions, existing licenses, and the commercial relationship surrounding the FRAND undertaking. No single factor is universally decisive. The central discipline is to avoid using a convenient revenue figure merely because it produces a large royalty while failing to reflect the patent’s technical contribution. For counsel and product teams, the practical issue is usually not abstract valuation; it is building a record that can withstand opposing experts, judges, auditors, and procurement reviewers.

## How the Calculation Usually Works

Most FRAND analyses begin with the relevant patented feature and trace how it operates in the accused product. The examiner identifies what the claim covers, what the product would do without that feature, and whether the feature improves performance, interoperability, reliability, storage, power consumption, or another measurable attribute. A multi-element portfolio may require separate allocation before any portfolio-wide rate is considered. The analysis then tests the royalty denominator: unit sales, product revenue, component revenue, net sales, or a narrower base tied to the value-bearing feature. Net sales may itself require deductions such as discounts, returns, taxes, and independently sold components, but courts do not apply one accounting convention across every jurisdiction without regard to the governing law. The same product can therefore support different legal rates in different proceedings because the scope of the portfolio, territorial rules, and evidentiary record differ.

After selecting an appropriate base, the analyst considers evidence about acceptable royalty rates. Comparable licenses can be informative when they involve similar patent scope, technology, stage of commercialization, and business model, but a license is not automatically a market price merely because it exists. Some comparables are heavily confidential, involve cross-licenses, bundle paid and unpaid rights, or cover litigation risk rather than only the relevant patent family. Cost-based approaches can examine the patentee’s R&D and Patent Office expense, then estimate what portion of that cost is recovered by the royalty revenue at a stated margin. These methods are aids rather than arithmetic answers: sunk development costs do not directly establish the value of an incremental improvement, and percentage-of-cost calculations can reward expense without asking what the implementer gained. The most credible approach usually triangulates technical value, comparable licensing, commercial context, and the parties’ actual FRAND conduct.

| Calculation input | Component-centered approach | Portfolio revenue approach | Cost-based cross-check |
| --- | --- | --- | --- |
| Royalty base | Revenue for the feature or component | Net revenue for products using the portfolio | Royalty income divided by qualifying R&D and prosecution cost |
| Strength | Closely links payment to technical value | Faster for a large portfolio | Provides an independent reasonableness check |
| Main weakness | Can be technically complex and contestable | May include value unrelated to the patents | Cost is not the same as benefit or commercial value |
| Common result | Lower base, potentially higher percentage | Higher base, potentially lower percentage | Used to test whether proposed returns are plausible |

## The Main Valuation Methods Compared
A top-down calculation applies one rate, such as 3%, 5%, or 8%, to a defined revenue base. It is easy to audit and useful for negotiation, but the number carries little meaning until the base, patent coverage, and non-discriminatory basis are established. A bottom-up calculation starts with the smallest technically relevant value-bearing unit and asks what share of its commercial value comes from the patented improvement. The share attributable to each patent is then aggregated where the portfolio is genuinely connected. This approach can better reflect contribution, though it creates disputes over whether the technical improvement, its implementation, the standards requirement, and the resulting product value can be separated.

Cost-based methods provide a third route. Suppose qualifying R&D and prosecution costs are $10 million, estimated annual license revenue is $20 million, and the selected profit margin is 20%. A simple cost-recovery calculation would show $2 million of return, or 20% of cost, before taxes and operating burdens; translated against license revenue, that is a 10% royalty. These are illustrative figures, not legal thresholds, and a court may reject the cost estimate, margin, portfolio scope, or revenue attribution. Other methods may discount future cash flows, divide benefit among several technical contributors, or use the incremental value that a particular feature adds to the final product. The selection must fit the technology. For a semiconductor feature that reduces power use, evidence might connect performance gains to demand or product value; for a connectivity standard, the analysis may focus on interoperability and access to an ecosystem.

## Practical Steps for a Defensible Calculation

First, define the exact licensing proposition before gathering valuation evidence. Record the jurisdiction, product generation, patent families, relevant standard version, asserted claims, and whether the proposed license covers portfolio access or only specified rights. A rate negotiated for one product family cannot safely be transferred to a different product without considering volumes, value, patent coverage, and other legal factors. Next, prepare a claim chart that identifies where each asserted feature appears and separates elements supplied by the patentee from elements supplied by the implementer. This step is essential because a standards-essential patent is only essential to the extent that every conforming implementation must include the claimed feature, as the relevant court or tribunal construes it.

The team should then select and reconcile royalty bases rather than mixing them silently. Prepare a product-revenue case, a component-revenue case, and a feature-centered case if the evidence permits, showing assumptions and reconciling each result. Comparable licenses should be normalized for portfolio size, payment structure, caps, minimums, exclusivity, cross-licenses, timing, and confidential terms. After calculating the returns, stress-test the result against at least one plausible alternative method and exclude figures that lack a logical connection to the patented contribution. Courts may consider whether the implementer would have licensed another comparable portfolio and whether the patentee’s offer was within the range supported by that evidence. The output should be a reasoned range, often supplemented by a central negotiation point, rather than a scientifically exact rate.

## Numbers That Matter but Have No Universal Threshold

There is no universal FRAND royalty of 3%, 5%, 8%, or 10%. Those figures recur in negotiations and cases because the denominator, technology, patent scope, and evidence differ. A royalty rate stated against total handset revenue can produce a much larger payment than the same percentage applied to the value of a small component. For example, a 5% royalty on $10 million in qualifying net sales produces $500,000; a 3% royalty on $2 million produces $60,000. Both are arithmetically valid, but only one may reflect the evidence. A rate should therefore always be reported together with its monetary result and royalty base. Per-unit prices, annual volumes, minimum annual payments, caps, and escalation clauses can materially change the actual burden and may be commercially relevant even when the headline percentage remains unchanged.

Dates also matter because a 2022 license, a 2024 portfolio, and a 2026 product are not automatically comparable. The evidence should account for the number of active families, standard changes, litigation history, remaining patent term, and the commercial maturity of the technology. A rate that represented a reasonable compromise in 2019 may become less persuasive where the product now includes several additional patented features, or more difficult to justify where the accused technology contributes little. Inflation and changes in market value do not mechanically revise a FRAND rate. Courts often ask what a reasonable royalty would have been at the relevant royalty-base date, while the law applicable to damages and limitation periods may point to a different time. The report must identify which date answers which question.

## Common Mistakes and Weak Assumptions

A frequent mistake is to treat the implementer’s entire profit as value attributable to the patent. That approach confuses the product’s value with the patented contribution and can produce a rate that fails technical and legal scrutiny. Another error is comparing headline percentages without normalizing the bases. A 4% royalty on total revenue and a 10% royalty on a narrowly defined component may produce the same current payment while embodying very different theories. Using an expired license, a small research collaboration, or a transaction covering thousands of unrelated rights as a direct comparable also weakens the analysis. Public awards can inform the commercial background, but an award is not binding on a later court and may reflect a settlement designed to end expensive litigation.

Cost-based reasoning is similarly misused. Patent costs are sunk, and prosecution cost depends on jurisdiction, claim count, and agency fees. Dividing total legal expense by revenue does not show what the patented feature is worth to users. Conversely, a low incremental-value estimate can also be extreme if the patented feature enables an entire compliant product, is indispensable to market adoption, or is embedded across millions of units. The calculation should identify what evidence would change the answer. A defensible model distinguishes facts known with confidence from negotiated assumptions, provides sensitivity cases, and explains why selected comparables are more probative than rejected ones. The aim is not to maximize the probability of winning through numerical complexity; excessive assumptions can make a model less transparent.

## Negotiation, Costs, and Timing

Early engagement is usually preferable because unresolved identification, product-version, and royalty-base issues become expensive later. Counsel should establish an evidence calendar, preserve product and licensing records, and align technical experts with the commercial team before positions harden. A FRAND negotiation may be faster than litigation, but it still requires claim construction, technical mapping, comparable-license analysis, and clear authority to settle. Outside technical analysis, legal drafting, translation, and case management can range from tens of thousands to hundreds of thousands of dollars for a substantial multi-jurisdiction dispute. A narrowly scoped calculation may cost less, while a portfolio spanning several standards, thousands of products, and multiple jurisdictions can require a seven-figure budget once discovery, experts, and parallel proceedings are included.

Timing should reflect both commercial urgency and evidence readiness. Start before a new product launch if the product introduces new claims, standards versions, or supply-chain entities, because launch decisions can be constrained by interim licensing arrangements. For an active dispute, however, negotiating before the patents, claims, and technical evidence are sufficiently defined can produce a weak anchor. A staged process often works better: document the implementation, narrow the disputed portfolio, exchange comparable licenses, calculate several bases, and then negotiate within an evidence-supported range. If a deadline is imminent, use an interim agreement only after confirming exclusivity, audit rights, payment mechanics, patent coverage, and the consequences of later rate-setting. The act of paying under a reservation-of-rights clause does not automatically concede the final legal rate.

## How iprs.cloud Fits the Working Method

For B2B intellectual-property teams, FRAND calculation is best treated as a repeatable registry and review workflow rather than a one-off spreadsheet. Useful records include patent-family status, product and standard mapping, claim evidence, asserted jurisdictions, licensing events, comparable-license terms, royalty bases, and the date of every calculation. Version control matters because a change in claim construction, product volume, or denominator can alter the result even when the percentage stays constant. A structured record also helps counsel distinguish a negotiated proposal from an adjudicated rate and helps product teams understand the cost impact without exposing privileged valuation analysis to every stakeholder.

Registry software should not be represented as calculating a legally binding rate without a case. It can organize evidence, flag missing comparables, preserve assumptions, and make range calculations reproducible, but the methodology and legal judgment remain the work of qualified professionals. Confidentiality controls are also important because license economics, unresolved claims, and settlement positions may be competitively sensitive. The strongest implementation joins technical mapping with financial modeling and legal metadata while maintaining separate access levels. Teams should validate totals independently and record why a figure was selected. Used that way, the system reduces spreadsheet duplication and audit friction; it does not replace the court-specific analysis needed to determine whether the offered terms are FRAND.

## Quick answers

### Is there a standard percentage for a FRAND royalty?

No. Rates such as 3%, 5%, or 8% are fact-specific, not legally prescribed benchmarks. The result depends on the patent contribution, royalty base, portfolio scope, comparables, jurisdiction, and evidence presented.

### Should a FRAND royalty be based on total product revenue?

Not automatically. A product-centered base may be used in some cases, while a component or feature-centered base may better reflect technical value. The base must be legally and economically supportable rather than chosen only to raise or lower the payment.

### Can R&D costs determine a FRAND rate?

Costs can provide a reasonableness cross-check, but they do not directly measure the benefit created by the patent. The analysis must justify which costs qualify, the applicable profit margin, and any assumptions about portfolio size and revenue recovery.

### Does paying a license under reservation of rights accept the offered FRAND rate?

Usually not by itself, but the agreement language matters. Courts generally distinguish interim payment or conduct from a final determination, so counsel should confirm the payment, reporting, and reservation-of-rights provisions before invoicing.

### Can a negotiated FRAND rate be reused for another product?

Only with careful analysis. Differences in patent coverage, technical contribution, product value, volumes, jurisdictions, and license scope may justify a different result. A prior agreement can be evidence, but it is not automatically a fixed tariff.

Canonical: https://iprs.cloud/knowledge/how_is_a_frand_royalty_rate_actually_calculated_in_2026.php
Markdown: https://iprs.cloud/knowledge/how_is_a_frand_royalty_rate_actually_calculated_in_2026.php/index.md