Munich's FRAND guidelines: a compass drawn too finely

On 13 August 2026 the 7th Civil Chamber of the Landgericht München I published guidelines on the FRAND defence, drawn from four judgments handed down between January and April 2026. The exercise in transparency deserves credit. The result raises a question of substance: at what point does the predictability of a forum stop serving the parties and start deciding for them?

What the chamber put in writing

The document sets out a complete method. The FRAND licence is assessed in two stages, with the licensee's financial posture examined before the content of its offer. Figures appear: the implementer must pay the undisputed royalties and provide security where its offer falls below 60% of the holder's demand, above ten million dollars. A hearing devoted to FRAND takes place early in the proceedings, with submissions there limited to twenty-five pages.

The substance is no less detailed. Comparable licences are confined to agreements less than five years old, with a range of plus or minus 50% around the median. Standardised product prices serve the top-down analysis. An aggregate royalty burden is put forward, in the order of 8% for 5G. A discount of 15% is applied to Chinese portfolios, on the ground that public filing incentives inflate their value.

The methodological inversion

The sequence adopted places financial conduct before substance. Yet Huawei v ZTE has the offer and the counter-offer come before the security: it is the exchange on terms that reveals what the parties intend, the security being only its consequence. By putting the payment threshold first, the chamber allows a case to be decided before the offers have been read.

The Bundesgerichtshof, in its judgment of 27 January 2026, itself refused a strictly sequential application of the European framework and held the FRAND defence ineffective for want of a serious willingness to take a licence. The lesson is one of method: willingness is inferred from a body of conduct, not from a percentage.

Parameters without an evidential basis

The figures are drawn neither from the four consolidated judgments, nor from a court-appointed expert, nor from any consultation of the industry. They express the chamber's own assessment, stated in advance. A median is worth no more than the sample from which it is taken, and the choice of comparable agreements belongs to the parties.

The procedural effect is considerable. Once those values are published, a party wishing to depart from them is no longer discussing an item of evidence, it is contradicting a position its own court has already taken. The debate on the FRAND rate and on portfolio licences shifts from the case file to the document.

Two blind spots: arbitration and non-discrimination

The guidelines rule out any safe harbour for an offer of bindingarbitration , which they treat as a delaying tactic. Yet proposing that a neutral third party determine what the undertaking requires is in itself strong evidence of good faith.

The 15% discount on Chinese portfolios raises a heavier difficulty. It sits awkwardly with the non-discrimination limb of the undertaking given to theETSI, which the chamber precisely intends to enforce. A presumption attached to the holder's nationality travels fast, and it will travel wherever a Chinese licensor litigates.

There remains the real reach of the document. The 21st Civil Chamber has not adopted it, nor have the appellate courts, and parties do not choose their chamber in Munich. The predictability announced is therefore partial.

The contractual alternative

The starting point should remain the undertaking itself: what the holder promised, in this relationship, at this time, in the light of what each side offered and of how each behaved. Comparables, the top-down analysis and the examination ofessentiality serve that demonstration, they do not precede it. That reading has the advantage of travelling: it works before the Munich chambers, before the divisions of the Unified Patent Court and before the other European courts, because it rests on a shared obligation and not on a local scale.

In practice, three consequences. The holder of standard essential patents litigating in Munich is well advised to document its licensing programme before the FRAND hearing, since the timetable will leave no time for it afterwards. The implementer must treat security as a matter of proving its willingness, and not as a threshold to be crossed at the last moment. And both must take in that the same defence, raised before the UPC or before the French court, is assessed without that scale, which gives the choice of forum back its strategic value.

The UPC's FRAND decisions are tracked one by one in our UPC Litigation Watch, for example the Panasonic v Oppo decision of 22 November 2024, and the outcomes by division in our UPC statistics. We act for German firms (Patentstreitigkeiten in Frankreich und vor dem EPG) and for Chinese holders (China patent litigation in France and before the UPC), in particular as correspondent in France for patents and the UPC. Contact us.

This article is an updated adaptation of an analysis by Matthieu Dhenne originally published on Kluwer Patent Blog on 1 October 2026: Munich's FRAND Guidelines: A Compass Drawn Too Finely. The guidelines of 13 August 2026 are not published online by the court. Official sources: CJEU, 16 July 2015, Huawei Technologies v ZTE, C-170/13, ECLI:EU:C:2015:477 ; Bundesgerichtshof, Cartel Senate, 27 January 2026, KZR 10/25.