Pull the phone out of your pocket and you are holding the most heavily litigated object in the history of commerce. Not because of what is printed on the box, but because of what hums invisibly inside it: a small forest of patented inventions that let your device whisper to a cell tower, hop onto Wi-Fi, pair with your earbuds, and stream a video that someone, somewhere, compressed using a patented codec. Every one of those handshakes follows a published technical standard—a shared rulebook agreed upon by hundreds of competing companies so that an iPhone made in California can connect to a network built by Ericsson in Sweden using a chip designed by Qualcomm in San Diego. Standards are why your phone "just works." They are also the reason that, on any given week in 2026, the same two companies might be suing each other in Munich, London, Shenzhen, Brasília, and the Eastern District of Texas simultaneously.

This article is about the patents that make those standards possible—standard-essential patents, or SEPs—and about the famous, slippery promise that is supposed to keep the whole system fair: the FRAND commitment to license those patents on terms that are fair, reasonable, and non-discriminatory. (Some standards bodies drop the "fair" and call the same idea RAND; the law treats the two as interchangeable, and you will see both, often abbreviated together as F/RAND.) We will define every term of art as we go, so that a judge, a licensing executive, and someone who has never read a patent claim in their life can all follow along. By the end you will understand why a single missing definition—what, exactly, counts as "reasonable"?—launched a global litigation arms race; how courts in five countries calculate royalties using three incompatible methods; why U.S. antitrust law once promised to be the field's great disciplinarian and then quietly retreated; why 2024-2026 may be remembered as the moment the United Kingdom, the World Trade Organization, and a Texas jury each rewrote the rules; and what a company should actually do when an SEP licensing letter lands on its desk.

If you want the surrounding context—how patents are enforced generally, how global IP wars are waged, and where standards came from in the first place—it helps to read this alongside our comprehensive guide to patent infringement litigation, our survey of global patent litigation strategies, the case-study lessons in global patent wars, and the surprisingly gripping story of POSIX and the standards that quietly hold computing together. This piece sits at the intersection of all four.

What Makes a Patent "Essential"—and Why That Changes Everything

Start with the ordinary patent. A patent is a time-limited legal right, granted by a government, that lets the inventor stop others from making, using, or selling the claimed invention without permission—generally for twenty years from filing, in the United States under 35 U.S.C. § 154. The classic justification is a bargain: society grants the inventor a temporary monopoly in exchange for public disclosure of how the invention works, which spurs the next round of innovation. If you do not like the patent holder's price, you usually have an exit. You can design around the patent, license a competing technology, or simply do without. Competition disciplines the price. (For the fundamentals, see our plain-English guide to patent basics and utility patent basics.)

A standard-essential patent breaks that escape hatch. A patent becomes "essential" when its claims cover technology that the published standard requires—when there is no way to build a compliant product without practicing the invention. If the 5G specification says a phone must encode a control signal in a particular way, and your patent covers that particular way, then every 5G phone on Earth infringes your patent the moment it powers on. There is no designing around a mandatory feature of the standard; to avoid the patent, a manufacturer would have to abandon the standard, which means abandoning the ability to connect to any 5G network. That is commercial suicide.

This is the heart of the matter. An essential patent confers something an ordinary patent does not: unavoidable leverage. Economists call the danger patent hold-up—the worry that once an industry has sunk billions of dollars into building products around a standard, an SEP holder can hold the whole industry hostage, demanding royalties far beyond the technology's real contribution simply because the alternative (an injunction shutting down the product line) is catastrophic. The mirror-image danger is hold-out (sometimes "reverse hold-up")—the worry that implementers will exploit the SEP holder's FRAND promise to stall negotiations for years, using the patented technology for free while litigation crawls along, secure in the knowledge that the worst case is paying the royalty they should have paid in the first place. Modern SEP law is, at bottom, a long argument about which of these two abuses is the greater threat. Where you stand often depends on whether you sell patents or sell phones—and, as we will see, U.S. enforcement policy has lurched between the two answers with every change of administration.

A few numbers convey the scale. Over the past fifteen years the count of declared SEPs—patents whose owners have told a standards body they might be essential—exploded from a few thousand to roughly 95,000 patent families covering cellular standards alone by the mid-2020s, owned by several hundred companies. The leading holders of 5G SEPs are a familiar cast: Huawei, Qualcomm, Samsung, Ericsson, Nokia, ZTE, LG, and a tier of specialist licensing firms. Chinese entities now hold roughly a third of all declared 5G SEPs, a tectonic shift from the 3G era when the portfolio was overwhelmingly American, European, Korean, and Japanese. And crucially, the universe of licensees has burst past smartphones: automakers building connected cars, appliance and meter makers wiring up the Internet of Things (IoT), medical-device companies, drone manufacturers, and industrial-equipment firms now all implement cellular standards and now all receive licensing demands. The fight the phone industry has waged for fifteen years is now arriving, all at once, in a dozen new industries that have never fought it before.

The Bargain: Standard-Setting Organizations and the FRAND Promise

To see why FRAND exists, picture the alternative. Suppose there were no standards, and every networking company designed its own incompatible protocol. Your phone could call other phones of the same brand and nothing else. The value of a network grows with the number of devices on it (the famous "network effect"), so fragmentation would be a disaster for everyone. The entire point of a standard is to converge—to get the industry to agree on one shared method so that everything interoperates.

Standards are forged inside standard-setting organizations (SSOs), also called standard development organizations (SDOs). These are collaborative bodies—part engineering club, part diplomatic summit—where competitors hammer out technical specifications. The most important for our purposes is the European Telecommunications Standards Institute (ETSI), whose members, working through a partnership called 3GPP, develop the cellular standards: GSM (2G), UMTS (3G), LTE (4G), and 5G New Radio. Wi-Fi standards come from the Institute of Electrical and Electronics Engineers (IEEE); video codecs from groups like ITU-T and ISO/IEC's MPEG; and dozens of other SSOs handle everything from Bluetooth to automotive safety buses.

Here is the structural trap the SSOs had to solve. When a company contributes its patented technology to a standard and persuades the committee to adopt it, two things happen at once. The technology becomes vastly more valuable (the whole world will now use it), and the patent holder gains the hold-up leverage described above. Left unchecked, that leverage would make rational companies refuse to contribute their best technology to standards, or would let the winners extract monopoly rents afterward. Either way the standard suffers. Worse, because standard-setting necessarily involves competitors talking to one another about what technology the entire market will adopt, the process flirts constantly with the antitrust laws—an SSO is, structurally, a meeting of rivals coordinating on a common product input.

The SSOs' answer, refined over decades, is a quid pro quo written into their intellectual-property-rights (IPR) policies. A participating company must (1) disclose patents it believes may be essential to a standard under development, and (2) commit, in advance, to license any patents that turn out to be essential on fair, reasonable, and non-discriminatory terms. ETSI's policy, the one that anchors most cellular disputes, contains this undertaking in Clause 6.1 of its IPR Policy. The FRAND commitment is, in effect, the price of admission: you may steer the standard toward your technology, but in return you promise not to abuse the essentiality you gain.

Different SSOs draft these obligations with different teeth, and the drafting matters enormously. The IEEE rewrote its patent policy in 2015 to address hold-up directly—restricting when an SEP holder may seek injunctions against a willing implementer, instructing that a reasonable royalty be valued at the level of the smallest salable compliant unit rather than the end product, and providing that a FRAND commitment must be available regardless of the production stage at which the patented feature is used (a partial answer to the component-versus-device fight discussed below). The U.S. Department of Justice reviewed those changes and confirmed in a 2015 business review letter that it did not intend to challenge them, reasoning that clearer, more predictable licensing terms could speed standard adoption and reduce post-adoption litigation. SSOs routinely seek such DOJ business review letters to vet their policies in advance; the DOJ's letters concerning the VITA and IEEE disclosure rules, and its 2019 GSMA letter on the eSIM standard, sketch what the agency regards as a competitively healthy process—openness to all interested parties, a balance of interests, due process, an appeals mechanism, and consensus. (Congress added a separate inducement: the Standards Development Organization Advancement Act of 2004 lets an SSO cap its exposure to treble antitrust damages by filing a notification with the DOJ and FTC.)

Two features of the FRAND promise matter enormously and recur throughout the cases. First, it is generally enforceable not only by the SSO but by implementers as third-party beneficiaries—the manufacturers who build standardized products can sue to enforce the SEP holder's promise even though they were never parties to the original ETSI undertaking. U.S. courts treat this as a matter of contract law, often applying the law of the SSO's home jurisdiction (French law for ETSI). Second, and fatally, the words "fair," "reasonable," and "non-discriminatory" are never defined. The committees that drafted these policies could agree royalties should be FRAND; they could not agree on a number, a formula, or even a methodology. That deliberate vagueness was the only way to get the bargain signed—and it has kept patent litigators employed ever since. FRAND is, in essence, an agreement to agree to a number nobody can define.

The Royalty Problem, Illustrated

Let us make the abstraction concrete with a deliberately simple hypothetical. (This scenario and the parties named in it are invented to illustrate the math; any resemblance to real companies is coincidental.) Suppose Acme Mobility makes a 5G phone that sells for $400 and yields $40 of profit. The 5G standard incorporates contributions from, say, forty different SEP holders. Beacon Wireless, one of those forty, owns a portfolio of genuinely useful 5G patents and sends Acme a licensing demand: it wants 2% of the phone's selling price, or $8 per unit.

Is $8 fair? On its own, it does not sound outrageous—2% of revenue. But now do the arithmetic that keeps implementers awake at night. If each of the forty SEP holders demanded a "reasonable-sounding" 2%, the aggregate royalty would be 80% of the phone's selling price—$320 on a $400 phone that earns only $40 of profit. The standard would be unbuildable. This is royalty stacking: individually plausible demands become collectively ruinous because no single negotiator accounts for everyone else's slice. It is the SEP equivalent of the tragedy of the commons, and it is why "is this rate reasonable?" cannot be answered patent-by-patent in isolation. Reasonableness for one holder depends on what is left over for the other thirty-nine.

Three more wrinkles complete the picture, each generating its own line of cases. First, the royalty base: should the 2% be measured against the $400 phone or against the $20 baseband chip that actually performs the 5G function? U.S. courts, drawing on the "smallest salable patent-practicing unit" idea from cases like LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51 (Fed. Cir. 2012), have often insisted that royalties be tied to the value of the patented feature, not the price of the whole gadget—lest a patent on a minor component capture the value of the camera, screen, and brand it has nothing to do with. SEP holders counter that portfolio rates are conventionally quoted per-device and that the per-handset convention is itself a market norm. Second, apportionment: of the value Beacon's patents add, how much is intrinsic to the invention versus how much is merely the windfall of having been blessed by the standard? The foundational U.S. principle, articulated when Judge James Robart set rates in Microsoft Corp. v. Motorola, Inc., is that a FRAND royalty must capture the ex ante value of the technology itself—what it was worth before the standard locked it in—stripped of the hold-up premium that essentiality created. Third, essentiality and validity: many declared SEPs are not actually essential (they were declared defensively, "just in case"), and many are invalid over prior art. Technical studies of declared SEPs routinely find that only a minority truly read on the standard. A licensee who pays for a thousand "essential" patents may be paying for two hundred real ones—which is why a hard-nosed essentiality and validity review (the SEP cousin of a freedom-to-operate analysis) is the implementer's single most valuable defensive move.

Hold these four problems—stacking, base, apportionment, essentiality—in mind. Every methodology and every case below is an attempt to wrestle one or more of them to the ground.

Three Ways to Set a FRAND Rate

No statute prescribes how to compute a FRAND royalty. Courts have improvised, and three rival families of method have emerged. Sophisticated tribunals now use several at once, treating each as a cross-check on the others.

The Comparable-Licenses Approach

The first instinct of any valuation professional is to look at the market: what have actual, willing parties paid for this same portfolio in arm's-length deals? If Beacon has signed twenty existing 5G licenses averaging $0.50 per phone, that is powerful evidence that $0.50—not $8—is the FRAND rate. Comparable licenses are the gold standard for many courts precisely because they reflect real-world bargains rather than litigation-driven theory.

The difficulty is that SEP licenses are messy. They are almost always confidential. They bundle hundreds of patents across many countries; mix lump-sum payments with running royalties; throw in cross-licenses (where each party licenses the other, so no cash changes hands at all); and reflect idiosyncratic leverage—a settlement signed on the courthouse steps is not the same as a freely negotiated deal. To use a comparable, a court must "unpack" it: strip out the cross-license value, normalize for sales volume and geography, and adjust for portfolio differences. This is as much art as science.

The seminal comparables decision is the United Kingdom's Unwired Planet International Ltd. v. Huawei Technologies Co. Ltd. Justice Birss (as he then was) conducted a marathon trial, examined a stack of Ericsson and other licenses, derived benchmark rates, and then held something startling: there is only one truly FRAND rate for a given portfolio and licensee, and a UK court can set a global portfolio rate and enjoin a defendant who refuses to take a worldwide license on those terms. The UK Supreme Court affirmed in [2020] UKSC 37, cementing England as a jurisdiction willing to adjudicate the world's royalty, not merely the UK slice. That single holding is the reason so much of the action narrated below now plays out in London.

The Top-Down Approach

The comparables method answers "what have others paid?" The top-down method asks a different question: "what should the whole industry pay, and what is this holder's fair share of it?" It begins with an estimate of the aggregate royalty that all SEPs for the standard, combined, ought to command—say, an industry consensus that total 5G royalties should not exceed roughly 6–10% of handset revenue. It then allocates that pie among holders according to each holder's share of the truly essential patents. If Beacon owns 5% of the essential 5G patents and the total acceptable royalty is 6%, Beacon's slice is 0.3%.

The appeal is that top-down attacks royalty stacking head-on: by fixing the size of the pie first, it guarantees the slices cannot exceed the whole. Judge James Selna used a top-down analysis (cross-checked against comparables) in TCL Communication Technology Holdings Ltd. v. Telefonaktiebolaget LM Ericsson and arrived at rates dramatically lower than the UK's Unwired Planet rates for an overlapping Ericsson portfolio—a vivid demonstration that methodology, not just evidence, drives the number. (The Federal Circuit later vacated TCL's rate determination on Seventh Amendment grounds, holding that Ericsson was entitled to a jury trial on the past-damages component—itself a reminder of how unsettled U.S. procedure remains.)

Top-down has a soft underbelly. Counting patents treats a pioneering invention and a trivial tweak as equal "votes," when their value is wildly different. The aggregate-royalty number is itself a guess, often pieced together from companies' own promotional statements about what they intend to charge. And patent counting cannot address the "non-discriminatory" prong at all—it tells you Beacon's share of the standard, not whether Beacon charged Acme more than it charged Acme's competitor.

The Modified Georgia-Pacific Framework

The third approach imports the workhorse of ordinary U.S. patent damages: the fifteen factors of Georgia-Pacific Corp. v. United States Plywood Corp., 318 F. Supp. 1116 (S.D.N.Y. 1970), which model a hypothetical negotiation between a willing licensor and willing licensee. Courts apply a modified version, because several Georgia-Pacific factors must be neutralized in the FRAND setting—the hypothetical negotiation must assume the patent is FRAND-encumbered, must exclude the hold-up value of the standard, and must account for all the other SEPs competing for the same royalty headroom. The Federal Circuit blessed this adaptation in Ericsson, Inc. v. D-Link Systems, Inc., 773 F.3d 1201 (Fed. Cir. 2014) and Commonwealth Scientific & Industrial Research Organisation v. Cisco Systems, Inc. (CSIRO), 809 F.3d 1295 (Fed. Cir. 2015), which together stand for the proposition that FRAND damages must apportion to the patented feature and strip out the standardization premium.

In practice these three approaches are not rivals so much as instruments in one orchestra. A careful court will set a rate using comparables, sanity-check it with a top-down calculation, and structure the reasoning through modified Georgia-Pacific factors. When the methods agree, confidence is high; when they diverge by an order of magnitude—as they sometimes do—the divergence itself becomes the litigation. Readers who want the mechanics of patent damages outside the FRAND context will find them in our discussion of intellectual property damage statistics, which surveys reasonable-royalty and lost-profits theory across the IP landscape.

Two Foundational Decisions: Robart and Huawei v. ZTE

Before tracing the recent wars, two decisions deserve their own headstones because everything builds on them.

In the United States, Microsoft Corp. v. Motorola, Inc. proved a court could actually do this. Motorola had demanded roughly 2.25% of the end-product price for its Wi-Fi (802.11) and H.264 video SEPs—which, applied to every Xbox and Windows PC, would have come to billions. Microsoft sued for breach of the FRAND commitment. Judge Robart of the Western District of Washington held a bench trial and, in a 2013 opinion, did the previously unthinkable: he calculated FRAND rates himself, deriving figures orders of magnitude below Motorola's ask (fractions of a cent for the Wi-Fi patents). The Ninth Circuit affirmed the framework in Microsoft Corp. v. Motorola, Inc., 795 F.3d 1024 (9th Cir. 2015), establishing that FRAND is a justiciable contract obligation, that a court can set the rate, and—critically—that an opening demand grotesquely above FRAND can itself breach the commitment.

In Europe, the keystone is Huawei Technologies Co. v. ZTE Corp., Case C-170/13, decided by the Court of Justice of the European Union (CJEU) in July 2015. The question there was a competition-law one: when does an SEP holder's suit for an injunction abuse a dominant position under Article 102 of the Treaty on the Functioning of the European Union? The CJEU answered with a choreographed dance—the Huawei framework—that the parties must perform before an injunction is permissible. The SEP holder must (1) notify the implementer of the specific infringement; if the implementer signals willingness to license on FRAND terms, the holder must (2) make a specific written FRAND offer with rates and a royalty basis; the implementer must (3) respond diligently and in good faith, and if it disagrees, make a FRAND counter-offer; and if that counter is rejected, the implementer must (4) provide security (a deposit or bank guarantee) for its ongoing use. An implementer who skips its steps is an "unwilling licensee" exposed to an injunction; an SEP holder who skips its steps abuses its dominance and is denied one. The genius of Huawei v. ZTE is that it converts the hold-up/hold-out standoff into a procedure: behave well, and you are protected; stall or overreach, and you lose. Virtually every European SEP case since is litigated as a referendum on which side broke step.

Notice the deep structural difference these two cases embody. In the United States, FRAND is policed mainly as a matter of contract (and, as we are about to see, was once policed as a matter of antitrust). In Europe, FRAND is policed as a matter of competition law, woven directly into the test for abuse of dominance. Same promise, two legal homes—and the choice of home shapes everything from who bears the burden of proof to what remedy is on the table.

The Road Not Taken: U.S. Antitrust and the SEP

For a heady decade, many believed American antitrust law would become the chief disciplinarian of SEP abuse. The theory was elegant. An SEP holder who deceives an SSO into adopting its technology—concealing patents during the standard-setting process, or making a FRAND promise it never intends to keep—has used deception to acquire monopoly power over the standard, the classic stuff of Section 2 of the Sherman Act. Standard-setting itself, being a meeting of competitors, can also raise Section 1 conspiracy concerns when the process is rigged. The Supreme Court had already shown, in American Society of Mechanical Engineers v. Hydrolevel Corp., 456 U.S. 556 (1982), that an SSO and a member could be held liable for manipulating a standard to crush a rival. The pieces seemed to be in place. (For the doctrinal background, our library practice notes on antitrust risks in standard-setting and on IP licensing under the antitrust laws map this terrain in detail.)

Three lines of authority developed, and each turned out to be more complicated than it first appeared.

The "patent ambush" cases. The marquee example is Rambus. The Federal Trade Commission found that Rambus had sat through years of an SSO's work on memory standards while quietly steering its own patent applications to read on the technology under discussion, then sprung its patents on the industry once the standard was locked in. The FTC condemned this as unlawful monopolization and ordered Rambus to license at capped rates—only to be reversed by the D.C. Circuit in Rambus Inc. v. FTC, 522 F.3d 456 (D.C. Cir. 2008), which held the agency had not shown that Rambus's deception actually caused its monopoly, as opposed to merely letting it charge more. The lesson, sobering for antitrust enthusiasts, is that proving causation in a standard-setting deception case is genuinely hard. By contrast, in the private-litigation track, the Third Circuit in Broadcom Corp. v. Qualcomm Inc., 501 F.3d 297 (3d Cir. 2007), held that an intentionally false FRAND commitment, relied upon by the SSO when it incorporated the patentee's technology, can state a monopolization claim—keeping the deception theory alive even after Rambus narrowed it.

The "renege on FRAND" cases. A separate theory targets not deception at adoption but bad faith afterward—seeking injunctions against willing licensees, or repudiating an inherited FRAND promise. The FTC pursued both through its consumer-protection authority under Section 5 of the FTC Act. In In re Negotiated Data Solutions (N-Data) (2008), it secured a consent order against a company that had bought patents subject to a predecessor's RAND commitment and then tried to charge far more, holding that walking away from an assumed FRAND obligation could itself be an unfair method of competition. In In re Robert Bosch GmbH (2012) and In re Motorola Mobility / Google (2013), the FTC extracted consent decrees restricting SEP holders from seeking injunctions against willing licensees. These cases never produced binding appellate precedent—they settled—but they established a strong agency norm that an injunction against a willing licensee is presumptively abusive.

The Qualcomm saga. The high-water mark, and then the retreat, was FTC v. Qualcomm Inc. The FTC sued in 2017, alleging that Qualcomm leveraged its modem-chip dominance through a "no license, no chips" policy—refusing to sell chips to handset makers unless they also took a separate (and, the FTC said, supracompetitive) patent license, and refusing to license its SEPs to rival chipmakers at all. Judge Lucy Koh of the Northern District of California agreed across the board in a sweeping 2019 decision, finding both a contractual and an antitrust duty to license rivals and ordering Qualcomm to renegotiate its licenses. Then the wheels came off. In an extraordinary public split, the Department of Justice filed a statement of interest siding with Qualcomm against the FTC, and the Ninth Circuit reversed in FTC v. Qualcomm Inc., 969 F.3d 974 (9th Cir. 2020), holding that Qualcomm's conduct was "hypercompetitive, not anticompetitive," that antitrust law generally imposes no duty to deal with rivals, and that breach of a FRAND commitment is a contract matter to be policed by contract law, not Section 2.

FTC v. Qualcomm is the hinge of the whole American story. After it, the antitrust path to disciplining SEP abuse narrowed dramatically; the practical message is that, in the U.S., FRAND breaches are now litigated mostly as contract (the Microsoft v. Motorola track) rather than as monopolization. That contraction is not an accident of a single panel. It tracks a policy whipsaw across administrations: a 2013 joint DOJ–USPTO policy statement that cautioned against injunctions for FRAND-encumbered SEPs; a 2019 DOJ–USPTO–NIST statement (under Assistant Attorney General Makan Delrahim, who reframed hold-out as the bigger danger) declaring that the usual remedies—injunctions, ITC exclusion orders, enhanced damages—remain fully available for SEPs; a withdrawn 2021 draft that tried to swing back; and a 2022 withdrawal of all prior statements, leaving the field, once again, to case-by-case judicial development. The one constant background rule is eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006), which forces any patentee seeking an injunction to prove irreparable harm and the inadequacy of money damages—a test FRAND-encumbered SEP holders find especially hard to satisfy, which is precisely why they prefer venues abroad.

A related and separate dispute, Apple Inc. v. Qualcomm Inc., showed how this plays out between private giants. Apple accused Qualcomm of the same licensing practices the FTC attacked and of charging royalties on the full handset price rather than the chip; Qualcomm countersued for infringement and won an ITC import recommendation and overseas injunctions (including in Germany and China) against certain iPhones. The two companies were litigating on multiple continents until, on the eve of trial in San Diego in April 2019, they settled globally—Apple paid a multi-billion-dollar lump sum, signed a six-year license, and dropped everything. The pattern is by now familiar: even the most consequential SEP fights tend to end not in a definitive ruling but in a negotiated peace, because the cost and uncertainty of fighting on five fronts eventually swamps the value of winning on any one.

The Modern Wars: 2024–2026 Rewrote the Map

For a decade the structural questions—global rates, injunctions, methodology, antitrust—simmered. Between 2024 and 2026 they boiled over, and several jurisdictions handed down decisions that genuinely changed the strategic calculus. We take them in turn.

England Invents the Interim License

The most consequential development of the period came from the English Court of Appeal, and it is subtle, so it rewards a careful read.

Recall that Unwired Planet gave English courts the power to set a global FRAND rate and enjoin a defendant who refuses it. But FRAND trials are slow—often two to three years. In the meantime, the SEP holder can run to faster, injunction-friendly forums abroad (Germany, the Unified Patent Court, Brazil, Colombia, China) and obtain injunctions that pressure the implementer to settle on the holder's terms before the English court ever reaches its measured global rate. The implementer's willingness to be bound by the English determination is rendered worthless if it gets shut down everywhere else first.

In Panasonic Holdings Corp. v. Xiaomi Technology UK Ltd., [2024] EWCA Civ 1143, the Court of Appeal cut this knot. Both sides had agreed to be bound by the English court's eventual global FRAND rate, yet Panasonic kept pursuing injunctions in Germany and at the Unified Patent Court. The Court of Appeal held that a willing licensor, acting in good faith under ETSI Clause 6.1, would grant the implementer an interim license—a temporary, pendente lite license covering the period until the final rate is set—and declared that a willing licensor in Panasonic's position would do exactly that. Lord Justice Arnold's reasoning reframed the FRAND commitment as imposing ongoing good-faith obligations, not merely an obligation to license once a rate is fixed. The interim-license declaration is a soft remedy with hard consequences: it brands the holder's parallel foreign injunction campaign as inconsistent with its FRAND duty and gives the implementer cover to resist settlement extortion.

The doctrine matured months later in Lenovo (United States), Inc. v. Telefonaktiebolaget LM Ericsson, [2025] EWCA Civ 182 (Feb. 28, 2025). Lenovo had undertaken to take a license on whatever terms the English Patents Court found FRAND; Ericsson nonetheless pressed for injunctions in Brazil, Colombia, the U.S. ITC, and elsewhere. The Court of Appeal held that Ericsson had breached its Clause 6.1 good-faith obligation by pursuing those foreign injunctions, declared that a willing licensor would enter an interim license, and—when Ericsson refused the interim terms the court specified (a figure set roughly midway between the parties' offers)—issued a follow-on declaration that Ericsson was an unwilling licensor in breach of FRAND. That is a remarkable inversion. For most of SEP history, "unwilling licensee" was the epithet courts hurled at implementers; the English court turned it around and pinned "unwilling licensor" on one of the world's premier SEP holders.

The strategic upshot, much remarked in the trade press, is that the United Kingdom—long viewed as a patentee-friendly venue because of Unwired Planet—has, through the interim-license doctrine, become attractive to implementers who want a neutral global rate without being bombed into submission abroad while they wait. The chessboard tilted.

The Tesla Footnote: There Is No FRAND Right to a Pool License

The interim-license idea has limits, and Tesla found one. In Tesla, Inc. v. InterDigital, Inc., [2025] EWCA Civ 193 (Mar. 6, 2025), Tesla asked the English court to set a FRAND rate not for one company's portfolio but for the entire Avanci 5G patent pool—the aggregated SEPs of dozens of holders licensed through a single platform (more on Avanci below). The Court of Appeal, splitting, dismissed the claim: the FRAND commitment runs from each individual SEP holder to the SSO, and Tesla could not identify a contractual right entitling it to a court-set rate for a pool as such. An implementer who wants a pool license must negotiate with the pool, or litigate the underlying patents holder-by-holder. The decision drew a useful boundary: English courts will set global rates for a licensor's portfolio, but they will not appoint themselves rate regulators for multi-party licensing platforms. (Arnold LJ pointedly observed that a non-FRAND declaration against a single pool member might nonetheless pressure the pool to revisit its terms—so the door is ajar even where Tesla's particular key did not fit.)

China Sets the World's Rate—Then the WTO Pushes Back

While London perfected the interim license, China asserted itself as a global rate-setter. In Oppo v. Nokia, the Chongqing No. 1 Intermediate People's Court issued the first Chinese judgment setting a global FRAND rate for a cellular portfolio—covering Nokia's 2G through 5G SEPs—at figures on the order of $0.707 per 5G handset for much of the world (with a higher rate for certain regions), and the Supreme People's Court confirmed that Chinese courts have jurisdiction to set worldwide rates. China's State Administration for Market Regulation (SAMR) reinforced the trend by finalizing its Anti-Monopoly Guidelines for Standard-Essential Patents on November 4, 2024, codifying disclosure duties, good-faith negotiation expectations, and constraints on injunctive relief. The message was unmistakable: China would no longer accept rates set in London or Mannheim as the global benchmark for products largely manufactured and sold within its borders.

The most explosive Chinese tactic, however, was the anti-suit injunction, and it is here that 2025 delivered a genuine landmark. Beginning around 2020, Chinese courts issued sweeping ASIs—orders forbidding a party from litigating SEP claims anywhere else in the world—backed by punishing daily fines. In Xiaomi v. InterDigital, the Wuhan court barred InterDigital from pressing parallel proceedings, with penalties around RMB 1 million per day; in Huawei v. Conversant and Oppo v. Sharp, Chinese courts issued ASIs reaching into Germany and Japan. The European Union responded by filing a WTO complaint, DS611, alleging that China's ASI practice unlawfully restricted the enforcement of patent rights in violation of the TRIPS Agreement.

DS611 resolved in 2025, and it resolved against China. A WTO panel issued its report in April 2025; after the EU invoked the agreed appeal-arbitration procedure, the appeal arbitrator's award of July 21, 2025 found that China's unwritten anti-suit-injunction policy was inconsistent with China's TRIPS obligations. China then announced at the WTO that it had withdrawn the policy, pointing to a notice issued by its Supreme People's Court on September 23, 2025, and the European Commission acknowledged the withdrawal in early 2026. This is the first time an international trade tribunal has constrained a jurisdiction's use of SEP anti-suit injunctions, and it materially cools the "race to the courthouse" dynamic that ASIs had super-charged. It does not end jurisdictional competition—courts will still compete to set the global rate—but the most aggressive weapon in that competition has, at least formally, been sheathed.

Anti-Suit Injunctions, Generally, and the Federal Circuit's Cautious Entry

Step back for the general mechanics, because the vocabulary is delightfully absurd. An anti-suit injunction (ASI) orders a party not to sue (or to stop suing) abroad. Predictably, the targeted forum hits back with an anti-anti-suit injunction (AASI)—German courts pioneered these, forbidding a party from even seeking an ASI elsewhere. And the escalation does not stop: courts have issued anti-anti-anti-suit injunctions to protect their AASIs. The whole edifice is a study in how rational actors, each protecting its own jurisdiction's authority, can collectively produce chaos. (U.S. anti-suit doctrine, for the curious, applies a "comity" analysis that is generally hostile to enjoining foreign litigation except where it threatens the domestic court's own jurisdiction—a posture our library's practice note on anti-suit injunctions in U.S. courts unpacks at length.)

U.S. courts have historically been cautious, granting ASIs only where the foreign suit would frustrate the domestic court's ability to do justice. That changed at the margins in Ericsson Inc. v. Lenovo (United States), Inc., No. 24-1515 (Fed. Cir. Oct. 24, 2024). Ericsson had won injunctions against Lenovo's 5G phones in Brazil and Colombia; Lenovo sought a U.S. ASI to neutralize them pending the U.S. FRAND determination. The Federal Circuit reversed the district court's denial, holding that the lower court applied the wrong legal standard and that a FRAND commitment can support an ASI where pressing the foreign injunction is itself inconsistent with the holder's FRAND duty. The decision signaled that American courts may be more willing than previously thought to police FRAND through ASIs—though, as it happened, the Ericsson/Lenovo war never ran to a U.S. judgment: in April 2025 the parties settled globally, entering a cross-license and agreeing to arbitrate the remaining terms. That ending is itself instructive—even the marquee cases usually end in negotiated peace.

The Unified Patent Court Enters the Arena

A new and powerful forum arrived in mid-2023: the Unified Patent Court (UPC), a single court with jurisdiction over the new European patent with unitary effect, capable of issuing injunctions spanning most of the EU in one stroke. SEP holders watched eagerly to see whether the UPC would be fast and injunction-friendly (good for licensors) or rigorous about FRAND defenses (good for implementers). The first substantive answer came in Panasonic v. Oppo, where the Mannheim Local Division (Nov. 22, 2024) issued the UPC's first FRAND decision. Applying the Huawei v. ZTE framework, the court granted Panasonic an injunction effective across the UPCA states where the patent was in force, finding that Oppo had not behaved as a willing licensee—pointedly, Oppo had failed to disclose its own sales data, which the court treated as essential information a good-faith licensee would provide. The UPC thereby planted a flag as a venue where a patient and well-behaved SEP holder can obtain broad, fast injunctive relief—adding yet another front to the global board and intensifying the very pressure that doctrines like the English interim license are designed to relieve. For implementers, the UPC's arrival is one more reason that venue selection has become as decisive as the merits, a theme our global patent litigation strategies guide develops across the wider patent-enforcement landscape.

A Word on U.S. Juries

For all the doctrinal sophistication abroad, the United States still produces the eye-watering damages numbers, because it lets juries decide patent damages. In January 2024 an Eastern District of Texas jury awarded $67.5 million to G+ Communications for Samsung's infringement of 5G SEPs. In 2025, Headwater Research won a jury verdict of roughly $278.8 million against Samsung and $175 million against Verizon over wireless patents (the latter later settled). These figures dwarf the per-unit royalties courts in London or Chongqing compute, which is precisely why patentees prize U.S. juries and implementers fear them—and why so much SEP strategy is, at root, about where the fight happens. For the broader U.S. patent-litigation playbook, see our comprehensive guide to patent infringement litigation and the case-study lessons in global patent wars.

The IoT and Automotive Frontier: Who Takes the License?

The smartphone wars were fought among parties who all understood the rules. The expansion of cellular standards into cars, appliances, meters, and machines has dragged whole new industries—often unwilling and unprepared—into the SEP arena, and it has surfaced a structural question the phone industry never had to answer: at what point in the supply chain should the license be taken?

A smartphone maker like Apple or Samsung designs the whole device and licenses SEPs directly. A car is different. An automaker buys a telematics module from a Tier-1 supplier (say, Continental or Bosch); the Tier-1 buys a cellular modem from a Tier-2 supplier; the modem contains a baseband chip from a chipmaker like Qualcomm or MediaTek. The 5G functionality lives in a $15 chip buried three layers down inside a $40,000 vehicle. So when an SEP holder demands a royalty, who pays, and on what base?

SEP holders overwhelmingly prefer device-level (end-product) licensing: license the automaker, and base the royalty on the connectivity's value to the car. Their argument is partly economic (the connectivity enables valuable features—navigation, over-the-air updates, emergency calling—whose worth is realized at the vehicle level) and partly practical (there are a handful of automakers but thousands of component suppliers, so device-level licensing is administratively simpler). Component makers and many automakers counter with license-to-all (or "access for all"): they argue the FRAND commitment obliges the SEP holder to grant a license to any willing applicant in the chain, including the chipmaker, and that licensing at the chip level both improves efficiency and ties the royalty to the smallest unit that actually practices the patent. (The IEEE's 2015 policy, recall, leaned toward the license-to-all camp by requiring FRAND availability regardless of production stage—one reason the choice of SSO and its IPR policy can quietly decide the outcome.)

This "holy war"—the phrase is the industry's own—nearly produced a definitive answer. In Nokia v. Daimler, the Düsseldorf Regional Court referred pointed questions on component-level licensing to the CJEU (Case C-182/21), asking whether an SEP holder may insist on licensing the carmaker rather than the supplier. The European IP world braced for a landmark ruling. It never came: Nokia and Daimler settled in June 2022, Daimler took a vehicle-level license, and the referral was withdrawn before the CJEU could rule. The structural question therefore remains formally unresolved in Europe as of 2026, governed by an uneasy détente and the antitrust authorities' general (and contested) view that the FRAND commitment does not strictly compel chip-level licensing so long as the chain ends up licensed. It is the most important open question in IoT patent law—and it is open precisely because the party that was losing kept settling before a court could rule against it.

Patent Pools: The Avanci Solution

If forty SEP holders each demanded a separate negotiation with each automaker, the transaction costs would be staggering. The market's answer is the patent pool—a platform that aggregates many holders' SEPs and offers a single license at a single rate. The dominant example in connected vehicles is Avanci. An automaker signs one Avanci agreement and obtains rights to the pooled cellular SEPs of dozens of contributors, paying a flat per-vehicle fee. Avanci's 4G Vehicle program historically charged about $20 per car (raised in recent years), and its 5G Vehicle program, launched in 2023, set a rate of roughly $32 per vehicle (with an early-licensee discount), folding in 2G–5G and C-V2X technologies. By the mid-2020s, well over 200 million vehicles had been licensed through Avanci, and the licensor roster kept growing—Lenovo and the IoT module-maker Quectel, among others, joined as contributors. Avanci has also expanded beyond cars into adjacent programs for IoT and other connected devices.

Pools solve the transaction-cost problem elegantly, but they also raise their own antitrust questions, because a pool is, by definition, an agreement among horizontal competitors to license collectively. U.S. enforcement guidance—built up through DOJ business review letters going back to the MPEG-2 and DVD pools of the late 1990s—asks a few recurring questions to separate a procompetitive pool from a cartel. Does the pool license only patents that are genuinely essential and complementary (so the pool integrates technologies that must be used together) rather than substitute patents that would otherwise compete on price? Is an independent expert used to vet essentiality? Are licensees free to take individual licenses outside the pool and to challenge the validity of pooled patents? Does the pool refrain from sweeping in downstream price terms or foreclosing rival technologies? A pool that clears these hurdles is generally treated as legitimate; one that bundles substitutes or fixes downstream prices courts per se condemnation. (Our practice library's note on antitrust issues in patent pools and cross-licenses, and the broader treatment in our legal-protection-of-software overview, develop these factors further.)

Pools also create commercial controversies that antitrust does not reach. Holdouts who refuse to join can still demand separate royalties, undermining the pool's promise of a single clean payment. Implementers may dispute whether the pool's flat rate is itself FRAND—precisely the argument Tesla tried and failed to litigate in Tesla v. InterDigital, where (as discussed) the English Court of Appeal held there is no contractual right to a court-set pool rate. And allocating the pooled royalty fairly among contributors, who do not all hold equally valuable patents, is a perennial source of internal friction. Pools are a genuine improvement over chaos; they are not a cure.

The IoT frontier sharpens an old tension. Many IoT devices—a $3 smart sensor, a connected water meter—sell for a fraction of a phone's price, yet implement the same cellular standard. A royalty that is trivial on a $400 phone can swamp the margin of a $3 sensor. Whether FRAND rates should scale to the device's value or to the connectivity's value (which is arguably the same regardless of host device) is unresolved and economically momentous as cellular spreads into ever-cheaper objects. Companies entering connected-device markets should treat SEP exposure as a first-order design and budgeting question, not an afterthought—much as they would conduct a freedom-to-operate analysis for any new product before committing to a launch, and much as a software company would map its open-source and patent exposure before shipping (see our overview of protecting a mobile app with a comprehensive IP strategy).

Policy in Flux: The EU Regulation's Collapse and the U.S. Vacuum

For a few years it looked as though Europe might legislate its way out of the FRAND morass. In April 2023 the European Commission proposed a sweeping SEP Regulation that would have created an EU register of SEPs, mandatory essentiality checks by independent evaluators, expert determinations of aggregate royalties, and a compulsory FRAND-conciliation step before litigation—all administered through a new competence center at the EU Intellectual Property Office. Implementers broadly welcomed the transparency; SEP holders attacked it as bureaucratic, costly, and tilted against innovators. The European Parliament adopted its position in early 2024, and then the proposal stalled.

In a striking reversal, the European Commission withdrew the SEP Regulation proposal—announcing the intention in its February 2025 work programme and confirming the withdrawal on October 6, 2025—citing the absence of any foreseeable agreement. The decision did not pass quietly: the European Parliament voted to challenge the withdrawal before the Court of Justice, seeking annulment of the Commission's decision, so the procedural fight over whether the proposal can be revived was itself live into 2026. Substantively, the upshot is that Europe will continue to resolve SEP disputes through courts—the UPC, the German regional courts, and the English courts—rather than through a harmonized regulatory regime. The grand experiment in administrative FRAND was, for now, abandoned.

The United States, as the antitrust history above shows, has no comparable framework and has whipsawed with each administration. The 2013 joint DOJ–USPTO statement, the 2019 DOJ–USPTO–NIST statement, the withdrawn 2021 draft, and the 2022 withdrawal of all prior statements left U.S. SEP policy to case-by-case judicial development—a posture that has persisted, with the DOJ and FTC not always aligned (their public split in Qualcomm being the vivid example). The practical consequence is that, in the U.S., SEP rules come from courts (and juries), not from agencies, and they come dressed mostly as contract law after FTC v. Qualcomm foreclosed the antitrust path. eBay remains the quiet background rule that makes U.S. injunctions hard for SEP holders to get—exactly why they prefer Germany, the UPC, and the ITC.

Internationally, the World Intellectual Property Organization (WIPO) has positioned its Arbitration and Mediation Center as a neutral forum for FRAND disputes, administering scores of mediation requests, and arbitration is a rising trend (the Ericsson/Lenovo parties chose to arbitrate their final terms). Confidential, expert, single-forum, and globally enforceable under the New York Convention, arbitration sidesteps the jurisdictional circus entirely—if the parties will agree to it. For the mechanics and limits of arbitral resolution, see our guides to arbitration as alternative dispute resolution and AAA commercial arbitration.

A Practical Framework for FRAND Negotiations

Doctrine is interesting; what a client actually wants is a playbook. The Huawei v. ZTE dance maps neatly onto a phased negotiation that disciplines both sides and builds the record each will need if talks fail. Whether you hold the patents or build the products, the following structure is sound across jurisdictions.

Phase 1 — Preparation. The implementer's first job is technical and unglamorous: map which standards each product actually implements and which declared SEPs plausibly read on those features. Most declared SEPs are not essential; a sober essentiality screen of the holder's portfolio is the single highest-leverage thing an implementer can do, because it converts a demand for "10,000 essential patents" into a defensible position about the few hundred that matter. The implementer should also pull the holder's public licensing history and any prior rate determinations. The SEP holder, meanwhile, should clean up its declaration records (over-declaration corrodes credibility and invites essentiality challenges), build a defensible valuation grounded in comparables and a top-down cross-check, and learn the target's business—product lines, unit volumes, geographic mix—so its offer is tailored rather than boilerplate.

Phase 2 — Notice and Engagement. Under Huawei v. ZTE, the holder must give specific notice: which patents, why essential, how infringed—ideally with claim charts mapping patent claims onto the standard's specification. A vague "you infringe our portfolio, please pay" may not trigger the implementer's duty to respond and can sink a later injunction request, as the UPC's Panasonic v. Oppo and the English Lenovo v. Ericsson decisions both underscore in their attention to who behaved well. The implementer must answer promptly and substantively, signaling genuine willingness to license on FRAND terms without conceding infringement or essentiality, and requesting the information (claim charts, rate justification, comparable-license summaries) it needs to evaluate the demand. Silence or stonewalling is the fastest route to "unwilling licensee" status and an injunction.

Phase 3 — Offer and Counter-Offer. The holder's opening offer should be genuinely FRAND, not an aspirational anchor to be haggled down—Microsoft v. Motorola teaches that a wildly inflated opening demand can itself breach the FRAND commitment, and the FTC's N-Data and Bosch matters teach that seeking an injunction against a willing licensee can draw antitrust scrutiny even after Qualcomm narrowed the theory. The offer must specify rate, royalty base, territory, term, and payment structure. The implementer cannot simply say "no"; it must make a FRAND counter-offer, supported by analysis (comparables, top-down, or modified Georgia-Pacific). The discipline cuts both ways: an implementer who rejects without countering risks being branded unwilling, while a holder who refuses a reasonable counter risks being branded—as Ericsson was in London—an unwilling licensor.

Phase 4 — Security and Resolution. If talks deadlock, the implementer should provide security for ongoing use (an escrow deposit or bank guarantee sized to its proposed rate). Doing so preserves its "willing licensee" status, blunts injunction risk, and demonstrates good faith. From here the paths diverge: a global FRAND determination in a willing court (with, increasingly, a request for an English-style interim license to neutralize foreign injunction pressure during the wait); arbitration before WIPO or another body, prized for confidentiality and a single global forum; or, where a pool license fits, simply taking the Avanci or comparable platform license and avoiding bilateral combat altogether. The choice of forum is often more decisive than the merits, so sequencing and venue selection deserve as much strategic attention as the rate itself.

Throughout, both sides should remember the lesson of the marquee cases: they almost all settle. Apple v. Qualcomm settled on the courthouse steps; Ericsson/Lenovo settled globally; Nokia v. Daimler settled before the CJEU could rule. The litigation is leverage-building, not an end in itself. The rational goal is a license at a defensible rate, reached before the combined cost of fighting in London, Mannheim, Chongqing, and Texas exceeds whatever was in dispute.

Key Takeaways

Standard-essential patents sit at a genuine fault line in the law: they protect innovations valuable enough to be written into the technologies the whole world uses, yet that very essentiality strips away the competitive discipline that keeps ordinary patent prices in check. The FRAND commitment is the patch the standards system applied to that wound—a promise to license fairly, reasonably, and without discrimination—and its enduring weakness is that no one ever defined what those words mean.

The law has tried three different tools to enforce that promise. Contract (the Microsoft v. Motorola line) treats FRAND as an enforceable bargain a court can value. Competition law (the European Huawei v. ZTE line) folds it into the test for abuse of dominance and turns the standoff into a procedure. And antitrust (the Rambus / Broadcom / Qualcomm line) once promised to condemn deception and refusal-to-deal as monopolization—until FTC v. Qualcomm narrowed that path to a sliver, leaving the United States to police FRAND mostly through contract and juries.

The 2024–2026 period sharpened the picture in several durable ways. England, through Panasonic v. Xiaomi and Lenovo v. Ericsson, invented the interim license and showed it would call out an SEP holder as an unwilling licensor, reorienting the UK toward implementers. China set its own global rates in Oppo v. Nokia and codified SEP antitrust rules, while the WTO's DS611 ruling forced it to withdraw its aggressive anti-suit-injunction policy, cooling the worst of the jurisdictional arms race. The Unified Patent Court opened a fast, pan-European injunction venue with its first FRAND decision in Panasonic v. Oppo. Europe's grand attempt to regulate FRAND collapsed when the Commission withdrew the SEP Regulation, leaving courts in charge. And patent pools like Avanci matured into the default mechanism for licensing connected vehicles, even as the foundational question of who in the supply chain takes the license remains unanswered because the losing party keeps settling first.

The connected future will only widen the field. 6G standards loom for the next decade; cellular is spreading into cars, sensors, meters, and machines that sell for a fraction of a phone; and AI techniques are beginning to enter standards themselves, raising fresh questions our writing on AI-generated inventions and the key legal issues in artificial intelligence begins to map. For any company that builds, sells, or monetizes connected technology, fluency in SEP and FRAND principles is no longer a specialist's luxury. It is, increasingly, a condition of doing business in a connected world.

Frequently Asked Questions

What is a standard-essential patent in plain English? It is a patent on a piece of technology that an industry standard requires, so that you cannot build a product compliant with the standard without using the patented invention. Because there is no way to design around a mandatory feature of a standard, the patent holder gains unusual power—everyone who wants to make a compliant product must license it. That power is what FRAND commitments are designed to restrain.

What does "FRAND" actually mean—is there a formula? FRAND stands for fair, reasonable, and non-discriminatory (some standards bodies say just "RAND"). There is no formula. The phrase is a promise SEP holders make to standards bodies, but the standards committees never defined the words. Courts have improvised three main methods to fill the gap—comparing actual licenses, working "top-down" from a total acceptable industry royalty, and applying a modified version of the Georgia-Pacific patent-damages factors—and they often use all three as cross-checks. Different methods can produce very different numbers for the very same patents.

Is breaching a FRAND promise an antitrust violation in the U.S.? Rarely, anymore. For a time, regulators and courts treated deceiving a standards body (the Rambus "patent ambush") or reneging on a FRAND commitment as potential monopolization under the Sherman Act or an unfair method of competition under the FTC Act (the N-Data and Bosch consent orders). But the D.C. Circuit's Rambus reversal showed how hard causation is to prove, and the Ninth Circuit's 2020 decision in FTC v. Qualcomm held that breach of a FRAND commitment is a contract matter, not an antitrust one, and that there is generally no antitrust duty to license rivals. Today most U.S. FRAND disputes proceed as contract claims under the Microsoft v. Motorola framework.

Can an SEP holder get an injunction to stop sales? Sometimes, but only if it plays by the rules. Under the EU's Huawei v. ZTE framework, the holder must give specific notice, make a real FRAND offer, and give the implementer a fair chance to respond before an injunction is available; a holder that skips those steps abuses its dominant position. In the U.S., the eBay decision makes injunctions hard to obtain at all without proof of irreparable harm—which is why SEP holders favor Germany, the Unified Patent Court, and the ITC. England recently added a twist: an SEP holder that pursues foreign injunctions while a global rate is being set may be ordered to grant an interim license and branded an "unwilling licensor."

What changed in 2024–2026 that practitioners need to know? Four big things. England created the interim-license doctrine (Panasonic v. Xiaomi; Lenovo v. Ericsson). China set global FRAND rates in Oppo v. Nokia and issued SEP antitrust guidelines, but the WTO found its anti-suit-injunction policy violated TRIPS, prompting China to withdraw that policy. The Unified Patent Court issued its first FRAND ruling (Panasonic v. Oppo) and granted a multi-country injunction. And the EU withdrew its proposed SEP Regulation, leaving FRAND disputes to the courts.

Who takes the license in a connected car—the automaker or the chip supplier? Unsettled. SEP holders generally insist on licensing the automaker at the vehicle level; component makers and many automakers argue for "license to all," meaning the patent holder must license any willing party in the chain, including the chipmaker. The CJEU nearly ruled on this in Nokia v. Daimler, but the parties settled and the question was withdrawn before a decision. As of 2026 it remains the most important open question in IoT patent law.

What is a patent pool, and should my company join one? A patent pool aggregates many holders' SEPs and offers a single license at one rate, sparing implementers dozens of separate negotiations. Avanci is the leading pool for connected vehicles, charging a flat per-vehicle fee (about $32 for its 5G program). Pools dramatically reduce transaction costs but raise antitrust concerns if they combine substitute rather than complementary patents or fix downstream prices—which is why well-run pools license only vetted essential patents and let licensees take individual licenses and challenge validity. You generally cannot force a court to set a "FRAND" rate for the pool itself, as Tesla learned in Tesla v. InterDigital. For most implementers in a covered market, taking the pool license is the path of least resistance.

I received an SEP licensing letter. What should I do first? Do not ignore it—silence can make you an "unwilling licensee" and expose you to an injunction. Respond promptly, express genuine willingness to license on FRAND terms (without conceding infringement), and ask for the specifics: which patents, claim charts mapping them to the standard, and the basis for the proposed rate. In parallel, have counsel screen the portfolio for essentiality and validity, because many declared SEPs are neither. Then follow the phased framework above. This is a moment to engage qualified patent counsel, not to negotiate alone.

Related Articles

Selected Authorities and Further Reading

  • Huawei Technologies Co. v. ZTE Corp., Case C-170/13 (CJEU July 16, 2015)
  • Microsoft Corp. v. Motorola, Inc., 795 F.3d 1024 (9th Cir. 2015)
  • Unwired Planet International Ltd. v. Huawei Technologies Co. Ltd., [2020] UKSC 37
  • Panasonic Holdings Corp. v. Xiaomi Technology UK Ltd., [2024] EWCA Civ 1143
  • Lenovo (United States), Inc. v. Telefonaktiebolaget LM Ericsson, [2025] EWCA Civ 182
  • Tesla, Inc. v. InterDigital, Inc., [2025] EWCA Civ 193
  • Ericsson Inc. v. Lenovo (United States), Inc., No. 24-1515 (Fed. Cir. Oct. 24, 2024)
  • TCL Communication Technology Holdings Ltd. v. Telefonaktiebolaget LM Ericsson (C.D. Cal. 2017; vacated in part, Fed. Cir. 2019)
  • Ericsson, Inc. v. D-Link Systems, Inc., 773 F.3d 1201 (Fed. Cir. 2014); CSIRO v. Cisco Systems, Inc., 809 F.3d 1295 (Fed. Cir. 2015)
  • LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51 (Fed. Cir. 2012)
  • FTC v. Qualcomm Inc., 969 F.3d 974 (9th Cir. 2020)
  • Rambus Inc. v. FTC, 522 F.3d 456 (D.C. Cir. 2008)
  • Broadcom Corp. v. Qualcomm Inc., 501 F.3d 297 (3d Cir. 2007)
  • American Society of Mechanical Engineers v. Hydrolevel Corp., 456 U.S. 556 (1982)
  • In re Negotiated Data Solutions LLC, FTC File No. 051-0094 (2008); In re Robert Bosch GmbH, FTC Docket No. C-4377 (2012)
  • eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006)
  • Georgia-Pacific Corp. v. U.S. Plywood Corp., 318 F. Supp. 1116 (S.D.N.Y. 1970)
  • Oppo v. Nokia, Chongqing No. 1 Intermediate People's Court (global FRAND rate determination)
  • Panasonic v. Oppo, Unified Patent Court, Mannheim Local Division (Nov. 22, 2024)
  • Nokia v. Daimler, CJEU Case C-182/21 (referral withdrawn after settlement, 2022)
  • WTO, China — Enforcement of Intellectual Property Rights, DS611, Appeal Arbitration Award (July 21, 2025)
  • ETSI Intellectual Property Rights Policy, Clause 6.1; IEEE-SA Standards Board Bylaws (2015 patent policy update)
  • DOJ–USPTO–NIST, Policy Statement on Remedies for Standards-Essential Patents (2019); DOJ business review letters to VITA (2006), IEEE (2007, 2015), and GSMA (2019)
  • SAMR, Anti-Monopoly Guidelines for Standard-Essential Patents (Nov. 4, 2024)
  • European Commission, Proposal for a Regulation on Standard-Essential Patents, COM(2023) 232 (withdrawn Oct. 6, 2025)

This article provides general information and is not legal advice. SEP and FRAND law is fast-moving and varies significantly by jurisdiction; for advice on a specific licensing matter or dispute, consult qualified intellectual property counsel.