The global ramifications for the patent licensing industry are huge:
Patent holders who enter(ed) into agreements with chipset makers and other suppliers while hoping to reserve the enforcement of their rights against downstream customers--particularly for the purpose of collecting royalties from end-product makers--can no longer rely on a covenant to sue last (sometimes also called covenant to exhaust remedies) as a means of sidestepping patent exhaustion.
I'm grateful to German IP litigator Oliver Loeffel ("Löffel" in German) for flagging the following decision to me:
The Bundesgerichtshof (Federal Court of Justice of Germany) today published (PDF) a judgment (dated January 24, 2023) according to which
a covenant not to sue is, as a matter of law, not capable of circumventing patent exhaustion; and
a covenant to sue last (which the court refers to as a "covenant to be sued last", a term that I can't find elsewhere and which doesn't make sense to me linguistically) is not, as a matter of law, certain to avoid exhaustion. Instead of focusing on what might theoretically happen (i.e., that there are hypothetical scenarios in which the beneficiary of the covenant to sue last might find itself on the receiving end of an infringement action), it depends on the answer to the following question of fact: whether the beneficiary of the covenant must realistically expect--in the ordinary course of events--to be held liable for infringement.
This may also have implications (in the sense of potentially persuasive authority) for other jurisdictions. With a view to the United States, the first part is merely consistent with TransCore, Quanta, and the spirit of Impression Products, but the second part touches on what would raise a question of first impression in the Supreme Court. So far, patent holders such as Qualcomm were pretty confident that a covenant to sue last would not entail patent exhaustion under U.S. patent law. While persuasive authority from a foreign jurisdiction is normally given limited weight, Germany's Federal Court of Justice actually has a reputation for being patentee-friendly, and its patent-specialized judges regularly meet with the judges of the United States Court of Appeals for the Federal Circuit.
The relevant case numbers are X ZR 123 (Federal Court of Justice), 6 U 104/18 (Karlsruhe Higher Regional Court, appellate opinion of November 25, 2020), and 7 O 165/16 (Mannheim Regional Court, judgment of September 28, 2018). The plaintiff (and appellee, at least in the Federal Court of Justice) is Japan's IP Bridge, the defendant is HTC, and the patent-in-suit is EP2294737 on "control channel signalling for triggering the independent transmission of a channel quality indicator", a patent that has been asserted against various defendants and over which IP Bridge won a famous German patent injunction against Ford last year.
The law firm of Kather Augenstein (which by the way represented Ericsson against Apple last year) published an English translastion of key passages of the Mannheim court's decision (PDF). I have not been able to find the appellate decision, though it has been referenced by IAM (as a key post-Sisvel v. Haier FRAND ruling) and even by this blog (because the Munich I Regional Court applied the Karlsruhe Higher Regional Court's claim construction in the Ford case, though it is not even in the Karlsruhe circuit).
Previously, no one paid much attention to the question of patent exhaustion. But the Federal Court of Justice--after siding with IP Bridge on the question of infringement--has now remanded the case to the Karlsruhe court (Presiding Judge: Andreas Voss ("Voß" in German)) because the patent-in-suit may have been exhausted by a covenant to sue last. As a result, it is now known that HTC--on top of technical defenses and a FRAND defense to injunctive relief--argued that the patent had been exhausted (maybe by IP Bridge, or maybe by Panasonic, which originally obtained this patent) through contracts with two chipset makers whose products were incorporated into the accused products (HTC smartphones).
This is a very special subject, so let me start with why such a thing as a covenant to sue last exists in the first place, and then discuss the holdings of the Federal Court of Justice in this case.
The instrument that is called covenant to sue last (or covenant to exhaust remedies) was borne out of necessity: Qualcomm and other patent holders (primarily wireless standard-essential patent (SEP) holders) have always preferred to license their patents to smartphone makers (and, more recently, automakers) over licensing them to upstream suppliers such as baseband chipset makers. But they'd also like to be able to enter into contractual arrangements with chipmakers so long as those dealings don't trigger patent exhaustion, which would make it impossible to enforce patent rights against the downstream, particularly the end-product maker.
They knew early on that a contractual provision would not be capable of preventing exhaustion when a straightforward license is granted. The concept of patent exhaustion is that patent rights are exhausted with the first sale (if the patent holder makes a product) or the first authorized (licensed) sale. The German decision published today also recalls that fact.
So the first "workaround" that lawyers came up with was to enter into a covenant not to sue instead of granting a license. For some time, there was widespread belief that this would do the trick. But over the years, the U.S. Supreme Court's patent exhaustion caselaw became more expansive. The 2008 decision in Quanta Computer v. LG Electronics clarified that patent exhaustion applies to method claims (as opposed to only product claims). In Impression Products v. Lexmark International, the top U.S. court held that post-sale restrictions may be enforceable under contract law but do nothing to avoid patent exhaustion. It also took an expansive view on whether a foreign first authorized sale will exhaust U.S. patent rights.
Between Quanta and Impression, the Federal Circuit equated, in its TransCore decision, a covenant not to sue to a license for exhaustion purposes, The instrument of a covenant to sue last became popular. This is how it works:
The beneficiary neither gets a license nor a 100% reliable commitment not to be sued.
But recovery will be sought from the beneficiary only if it cannot be obtained from the downstream.
Take automotive patent licensing, for example. If Nokia had (which I believe is not and was not the case) entered into a covenant to sue last vis-à-vis the chipmaker, it would only have been able to sue that company after seeking and failing to obtain recovery from Daimler (end-product maker), Continental (tier 1 supplier that makes telematics control units), and whatever tier 2 supplier supplied the Network Access Device (NAD) to Conti.
In the dispute between IP Bridge and HTC, the former succeeded in persuading the appeals court (Karlsruhe Higher Regional Court) that a covenant not to sue would already be sufficient to avoid patent exhaustion and, therefore, a covenant to sue last was even more certain to do the job.
HTC appealed, and the Federal Court of Justice (which is based in Karlsruhe like the court below) reversed and remanded with the following holdings and instructions:
The Federal Court of Justice said it was not merely an obiter dictum that the Karlsruhe Higher Regional Court said a covenant not to sue would not trigger exhaustion. Instead, it was the bedrock of the lower court's conclusion that a covenant to sue last could not result in patent exhaustion either.
The Federal Court of Justice says the focus must be on whether the patent holder has authorized the "release into commerce" ("Inverkehrbringen") of products that implement the patented invention. Typically, that will be the case if the patent holder entered into a covenant not to sue, and contractual provisions according to which the patentee reserves its rights against third parties are invalid.
As a result, the Karlsruhe Higher Regional Court's conclusion that a covenant to sue last does not exhaust patent rights because even a covenant not to sue would not trigger exhaustion has been reversed.
The Federal Court of Justice then turns to questions that the court below didn't have to reach:
On remand, the Karlsruhe Higher Regional Court will have to determine whether the patent-in-suit falls under the capture clause of the patentee's relevant contracts with the two chipmakers.
Patent exhaustion is not ruled out only because the chipmakers don't make phones and the patent claims-in-suit involve technical components that are not found in chips, but only in phones. The Federal Court of Justice acknowledges that patent exhaustion generally applies only to the particular product that was sold on an authorized basis. Exhaustion does not necessarily apply to downstream products that contain that product as one of multiple components. But that may be the net effect if the only meaningful way in which the chipsets in question can be used is to incorporate them into mobile end-user devices. In that case, the patentee's authorization of the sale of such chipsets may amount to tacit consent to the distribution of mobile devices that incorporate those chipsets. What could weigh against that conclusion, however, would be a disclaimer of liability in the agreements between the chipmakers and HTC.
Even if the relevant agreements were to be interpreted to the effect that the patentee did not consent to the incorporation of those chipsets into mobile phones, the patent could still be exhausted if the technical effects of the patent-in-suit are materially achieved by those chipsets, with all other components of those mobile phones not playing a determinative role.
Now comes the most important part:
The Karlsruhe Higher Regional Court will have to determine whether the patentee gave its contractual partners (the chipmakers) an assurance that it would not assert the patent-in-suit against them--and such determination will have to focus on practical as opposed to exclusively theoretical considerations. The first and foremost question will be whether in the ordinarily expected course of events the chipmakers must fear to be sued over an infringement of the patent-in-suit.
In other words, the merely hypothetical possibility of the immediate beneficiary of the covenant to sue last being sued is not enough to avoid exhaustion. If it's rather unlikely that the immediate beneficiary of the covenant ever gets sued, then the covenant to sue last will be treated like a covenant not to sue, which in turn is treated like a license.
How can the chipmaker have a reasonable apprehension of being sued in the ordinary course of events? The normal course of events is that if you hold a patent and sue a device maker (and your infringement allegations have merit), you get paid--unless that device maker goes bankrupt along the way.
The short version is: the Federal Court of Justice does not want to let patent holders circumvent exhaustion through a covenant to sue last, but unless there are special circumstances that suggest otherwise, will treat it like a covenant not to sue, which is treated like a license.
Is this a desirable outcome? This post is already long enough without discussing the policy implications, but I do want to mention one: there may be cases in which chipmakers want peace of mind and where it would be in the public interest to allow patentees to give it to them, but in which patent holders may now tell them that after IP Bridge v. HTC it's too risky for them to enter into a covenant to sue last because they may exhaust their patent rights.
