What Constitutes Patent Infringement: Claims and Defenses--A Practical Guide

A Cease-and-Desist Letter Lands on the Desk

Picture an internet business—call it Sekure Kart—that has stumbled onto a cheap, fast way to generate the cryptographic keys behind end-to-end encryption. It is signing customers daily: small retailers, scrappy startups, individuals who move money online and would rather not pay the established providers' prices. Growth is the only problem the founders think they have. Then the CFO walks in looking pale and sets a letter on the desk.

It is a cease-and-desist letter. It is not from a patent troll. It is from one of the major payment-processing companies—a competitor that, unable to beat Sekure Kart on price, has decided to compete with a patent infringement lawsuit instead.

This is a frightening moment, because patent litigation is famously expensive and uncertain. A single case routinely runs into the millions of dollars in fees, and the outcome can turn on the construction of a single word. But it is also a moment that rewards understanding. Whether Sekure Kart actually infringes, what defenses it has, what the patentee must prove, and what it all might cost are answerable questions—and answering them well is the difference between panic and a plan.

This guide explains what constitutes patent infringement and how infringement claims and defenses work in U.S. federal court. It is written for three audiences at once: businesses that have been accused, patent owners weighing whether to enforce, and the lawyers who advise both. It takes a neutral, practitioner's view. It builds on our companion patent basics guide and our guide to sending and responding to cease-and-desist letters, and—because infringement turns on intensely fact-specific and technical questions—it is a map of the terrain rather than a substitute for advice from a registered patent attorney about a specific patent or accused product. Readers who want the next chapter, what happens once a case survives summary judgment and heads to trial, should turn to our comprehensive guide to patent infringement litigation, from summary-judgment denial to post-trial.

We will return to Sekure Kart throughout, because its situation illustrates nearly every concept that matters: how to read a patent, what infringement requires, where the battle is actually fought, and how a defense takes shape. (Sekure Kart and its dispute are entirely hypothetical, invented to make the doctrine concrete.)

The Foundation: The Claims Define What Is Infringed

Every infringement analysis begins in the same place—not with the accused product, and not with the patent's drawings or its prose, but with the patent's claims. The reason lies in the patent bargain. A government grants a patent so that an inventor will teach the public how to make and use an invention; in exchange, the inventor receives the right to exclude others from a defined slice of territory for a limited time. The claims mark out that slice. They are the numbered sentences at the very end of a patent, and they define exactly what the inventor owns—and therefore exactly what another party must be doing to infringe.

A patent's claims are not casual descriptions. They are drafted with the precision of a property deed and read with the same care a title examiner gives a metes-and-bounds description of land. A patent typically contains both independent claims, which stand on their own, and dependent claims, which incorporate everything in an earlier claim and then add a further limitation. Because a dependent claim is by definition narrower than the claim it depends from, an accused product that escapes the broad independent claim necessarily escapes the dependent claims beneath it. That is why an accused infringer attacks the broadest independent claims first: knock out claim 1, and a row of dependent claims often falls with it.

The rest of the patent—the specification, with its background, written description, and drawings—does a different job: it teaches the invention and supplies the meaning of the words used in the claims. The specification is where one looks to understand what a claim term means, to find the inventor's own definitions, and to see the worked examples that illustrate what was invented and what was not. But it is the claims that are infringed, not the specification. A useful discipline, for anyone holding a patent that has been asserted against them, is to read the claims first and pin down precisely what the patentee owns before turning to the specification to interpret it. Reverse that order and you will read the marketing pitch in the patent's introduction and convince yourself the patent covers everything in the field—which it almost never does.

From this flows the single most important rule in infringement law: the all-elements rule. Each claim is a list of elements (also called limitations), and to infringe a claim, an accused product or process must include every one of them. If the accused product is missing even a single claimed element—if a claimed machine has one different part, or a claimed process omits one claimed step—then it does not literally infringe that claim. Infringement is judged claim by claim and element by element; there is no infringement by "general similarity," by being "in the same space," or by competing with the patent owner. A product that does everything the patent describes and more still infringes; a product that does ninety percent of what a claim requires infringes nothing at all. Adding features cannot save you; subtracting a required one usually does.

Sekure Kart's founders put this rule to work the night the letter arrives. Reading the asserted patent's first claim, they notice it requires a step of "random number generation," and as they read the specification they see that the patent describes only standard, conventional ways of generating random numbers—pseudorandom algorithms seeded from a system clock, the sort every textbook covers. Sekure Kart, however, uses a wholly different, proprietary method that draws entropy from a physical source. If the claim's "random number generation" step means the conventional methods the patent actually describes, and Sekure Kart does something genuinely different, then Sekure Kart is missing that claimed element—and on the all-elements rule, it does not literally infringe. That is a real and potentially powerful non-infringement position. But, as we will see, it is the opening of the argument, not the end of it, because the patentee will fight over what "random number generation" means and whether Sekure Kart's method is close enough to count anyway.

The Federal Framework

Patent infringement is a creature of federal law, and it lives entirely in federal court. Federal district courts have exclusive subject-matter jurisdiction over patent infringement claims (28 U.S.C. § 1338), so such claims cannot be brought in state court. Any federal district court may hear a case if personal jurisdiction and venue requirements are met, and either party may demand a jury—which means highly technical questions about cryptography or pharmaceutical chemistry are frequently decided by lay jurors after a battle of expert witnesses.

Outcomes vary considerably from district to district, owing to differences in jury pools, judges' experience with patent cases, and time to trial. That variation is not a footnote; it is a major strategic input. It is why certain districts—historically the Eastern District of Texas, the District of Delaware, and the Northern and Central Districts of California, among others—have drawn a disproportionate share of patent suits, and why the choice of forum can be worth as much to a litigant as the strength of the merits. Venue is now constrained, however. In TC Heartland LLC v. Kraft Foods Group Brands LLC, 581 U.S. 258 (2017), the Supreme Court held that, for venue purposes, a domestic corporation "resides" only in its state of incorporation, sharply limiting where many defendants can be sued and shifting a large share of the docket to Delaware, where so many companies incorporate.

All appeals in patent cases—regardless of where the district court sits—go to a single specialized court, the U.S. Court of Appeals for the Federal Circuit, which sits in Washington, D.C. (28 U.S.C. § 1295). This unusual centralization, created in 1982 to bring uniformity to a body of law that had splintered among the regional circuits, means that one court of appeals shapes the doctrine for the entire country. The Federal Circuit's decisions may in turn be reviewed by the Supreme Court, which over the last two decades has taken an unusually active interest in patent law and has repeatedly rewritten the rules discussed throughout this guide—on eligibility, exhaustion, willfulness, fees, and injunctions alike.

The Types of Infringement

There are two broad kinds of infringement—direct and indirect—and indirect infringement comes in two forms, inducement and contribution. A foundational point ties them together: there can be no indirect infringement without an underlying act of direct infringement by someone. Indirect liability is, in effect, liability for helping or encouraging another's direct infringement, so if no one directly infringes, the inducer and the contributor walk free no matter how culpable their conduct.

Direct Infringement

A party directly infringes if, without authority, it makes, uses, offers to sell, or sells a patented invention within the United States, or imports it into the United States (35 U.S.C. § 271(a)). Each of those verbs is its own act of infringement: a company that merely uses an infringing machine in its operations infringes just as surely as the company that sold it.

Direct infringement is a strict-liability offense, and this surprises people more than almost anything else in patent law: intent is irrelevant. The accused infringer need not have copied the invention, need not have acted in bad faith, and—crucially—need not even have known the patent existed to be liable. It is enough that the accused product or process falls within a claim. An engineer who independently invents the very same thing, in good faith, with no knowledge of anyone else's patent, still infringes. (Contrast copyright, where independent creation is a complete defense; in patent law it is no defense at all.) That said, knowledge and intent matter a great deal to remedies, as discussed below in connection with willfulness—the difference between paying a reasonable royalty and paying triple. Whether the accused product or process falls within a claim is governed, again, by the all-elements rule—either literally or under the doctrine of equivalents.

Literal Infringement and the Doctrine of Equivalents

Literal infringement exists when every limitation of a claim is found, exactly, in the accused product or process. If a claim recites a chair with four legs attached to a seat and a curved back attached to the seat, a literally infringing chair must have all of those features as described. A three-legged stool does not literally infringe; neither does a four-legged chair with a flat back. Literal infringement is, in this sense, the easy case: you line up the words of the claim against the accused product and check off each box.

But patent law does not let an infringer escape merely by making a trivial change—swapping a screw for a rivet, or a copper wire for an aluminum one. If it did, the right to exclude would be worthless, defeated by any clever copyist willing to make a cosmetic substitution. Under the doctrine of equivalents, an accused product can infringe even when a claim limitation is not literally present, provided the difference between what the claim requires and what the accused product has is insubstantial. Courts often apply the function-way-result test, asking whether the accused element performs substantially the same function, in substantially the same way, to achieve substantially the same result, as the claimed element. The doctrine is applied element by element, not to the invention as a whole, so a plaintiff must show an equivalent for each limitation that is not literally met—it cannot wave at the overall similarity of the two products (Warner-Jenkinson Co. v. Hilton Davis Chemical Co., 520 U.S. 17 (1997)).

The doctrine of equivalents is powerful, so the law fences it in. Its single most important limit is prosecution history estoppel: where an applicant narrowed a claim during prosecution to win allowance—usually to get around prior art the examiner cited—it generally cannot later use the doctrine of equivalents to recapture the very scope it surrendered (Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722 (2002)). Festo established a presumption that a narrowing amendment surrenders everything between the original and amended claim, which the patentee can overcome only by showing the equivalent was unforeseeable, tangential to the reason for the amendment, or otherwise beyond what could fairly be expected. Two related limits round out the picture: the doctrine cannot be used to read a limitation entirely out of a claim (the "all-elements" floor again), and it cannot capture subject matter that is itself disclosed but unclaimed in the patent or covered by the prior art (the Wilson Sporting Goods hypothetical-claim analysis). The practical upshot is that the doctrine of equivalents lives in tension with the patent's notice function: the broader an equivalents theory, the less the public could have predicted the claim's reach from its words.

The doctrine of equivalents is exactly where Sekure Kart's fight will be joined. Sekure Kart's best position is literal non-infringement: its proprietary random-number generator is not the conventional method the claim requires, so a claimed element is simply missing. The founders ask their lead engineer to document the differences, and the result is a meticulous technical report—dozens of pages of flow charts cataloguing many specific differences between Sekure Kart's generator and the methods in the patent. That report is the backbone of the non-infringement defense. But the patentee will respond with the doctrine of equivalents, arguing that however many differences the report lists, they are insubstantial—that Sekure Kart's generator performs the same function (producing the randomness encryption needs) in substantially the same way to reach the same result. Here is the lesson that trips up engineers: the sheer number of differences matters far less than whether even one of them is substantial. A hundred trivial differences add up to nothing; one substantial difference can win the case. And whether a difference is substantial turns, in the end, on what the claim term means. The technical report is necessary but not sufficient; it sets up the two battles that will actually decide the case—claim construction and substantiality.

Indirect Infringement: Inducement and Contributory Infringement

A party that does not itself practice every claim element can still be liable for indirect infringement if it is responsible for someone else's direct infringement. This is how patentees reach the manufacturer whose customers do the infringing, or the supplier of a key component, rather than chasing thousands of end users one at a time.

Induced infringement (35 U.S.C. § 271(b)) reaches a party that actively encourages another to infringe—by, for example, selling a product together with instructions that direct the buyer to use it in an infringing way. Unlike direct infringement, inducement has a knowledge requirement with real teeth. The plaintiff must prove both that the defendant engaged in conduct that encouraged the infringing acts and that the defendant knew those acts constituted infringement of the patent—which entails knowledge of the patent itself. The Supreme Court established that knowledge requirement, and the willful blindness standard that can satisfy it, in Global-Tech Appliances, Inc. v. SEB S.A., 563 U.S. 754 (2011): a defendant who subjectively believes there is a high probability the patent exists and is infringed, and takes deliberate steps to avoid confirming it, cannot escape liability by burying its head in the sand—though mere negligence or even recklessness is not enough.

Two further refinements matter. First, a defendant's good-faith belief that the patent is invalid is not a defense to inducement; the Supreme Court reasoned in Commil USA, LLC v. Cisco Systems, Inc., 575 U.S. 632 (2015), that infringement and validity are distinct issues, and a defendant who thinks a patent is invalid can challenge it directly rather than ignore it. Second, inducement requires an actual completed act of direct infringement by the induced party—you cannot induce infringement that never happens (Limelight Networks, Inc. v. Akamai Technologies, Inc., 572 U.S. 915 (2014)).

Contributory infringement (35 U.S.C. § 271(c)) reaches a party that sells, offers to sell, or imports a component of a patented machine, or material for practicing a patented process, where that component is especially made or adapted for an infringing use and has no substantial non-infringing use. The decisive inquiry is usually that last clause: does the component have a substantial lawful use? If it does—if it is a "staple article of commerce" suitable for legitimate, unpatented purposes—there is no contributory infringement, no matter how the buyer ultimately uses it. Like inducement, contributory infringement carries a knowledge requirement: the seller must know the component is made or adapted for an infringing use, and must know of the patent. And, like inducement, it requires underlying direct infringement by a third party (Aro Manufacturing Co. v. Convertible Top Replacement Co., 365 U.S. 336 (1961), the "Aro I" decision that anchors the modern doctrine).

A worked example clarifies the line. Suppose a patent claims a water-filtration pitcher with a specially shaped filter cartridge, and a rival sells replacement cartridges that fit only that pitcher and serve no other purpose. Because the cartridge is especially made for the patented combination and has no substantial non-infringing use, selling it is contributory infringement (assuming the seller knows of the patent). Now suppose the rival instead sells ordinary activated charcoal that happens to be one possible filter medium but is sold for dozens of unrelated uses. That charcoal is a staple article with substantial non-infringing uses, so its sale is not contributory infringement—even if some buyers pack it into the patented pitcher. The seller might still face an inducement claim if it advertised the charcoal specifically for use in the patented pitcher, which shows how the two theories interlock: contribution targets the nature of the article, inducement targets the defendant's encouragement.

Divided (Joint) Infringement

A recurring and difficult problem arises when no single entity performs every step of a claimed method, or uses every part of a claimed system—when the steps are divided among multiple actors. This is the bane of method claims written for the networked world, where one step happens on a company's server, another on a partner's server, and a third on the user's phone. For a time, divided infringement was a loophole: a clever defendant could escape a method claim simply by ensuring that its customers, rather than the defendant itself, performed the final step.

The Federal Circuit, sitting en banc, closed much of that loophole in Akamai Technologies, Inc. v. Limelight Networks, Inc., 797 F.3d 1020 (Fed. Cir. 2015). It held that where no one actor performs all the steps, an accused party can still be liable for direct infringement when it directs or controls others' performance—including by conditioning a benefit (such as access to a service) on the performance of a step and establishing the manner or timing of that performance—or when the actors form a joint enterprise. The decision deliberately expanded attribution beyond the old, narrow "agency or contractual" test, recognizing that modern infringement is often a team sport. For system claims, the Federal Circuit has held that a party "uses" a system—and so can infringe—when it puts the system into service by controlling the system as a whole and benefiting from each element of it (Centillion Data Systems, LLC v. Qwest Communications International, Inc., 631 F.3d 1279 (Fed. Cir. 2011); Intellectual Ventures I LLC v. Motorola Mobility LLC, 870 F.3d 1320 (Fed. Cir. 2017)). Divided infringement remains a frequent battleground for internet, telecommunications, and software patents, where claimed steps are naturally spread across servers, service providers, and end users—which is precisely the architecture of a service like Sekure Kart's.

Special Statutory Categories: Sections 271(e), (f), and (g)

Beyond the familiar trio of direct, induced, and contributory infringement, the Patent Act defines several special categories of infringement that catch conduct the general provisions would miss.

Section 271(e) is the engine of pharmaceutical litigation. Its safe harbor permits a generic manufacturer to make and test a patented drug solely to develop the information needed for FDA approval, so that generics can be ready to launch the day the patent expires (a use the Supreme Court read broadly in Merck KGaA v. Integra Lifesciences I, Ltd., 545 U.S. 193 (2005)). But § 271(e)(2) simultaneously makes the very act of filing an Abbreviated New Drug Application (ANDA) seeking to market a generic before a patent expires an "artificial" act of infringement—the trigger for the Hatch-Waxman litigation that dominates the pharmaceutical docket. It is a neat piece of statutory engineering: it lets generics prepare in advance while giving brand owners a way to sue over that preparation.

Section 271(f) addresses an exporter's end-run: supplying from the United States the components of a patented invention for assembly abroad, in a way that would infringe if the assembly happened here. Congress enacted it to overrule Deepsouth Packing Co. v. Laitram Corp., 406 U.S. 518 (1972), which had allowed a defendant to ship a patented shrimp-deveiner abroad in pieces for final assembly and thereby escape liability. The Supreme Court has policed the provision's edges, holding that supplying a single component does not trigger it (Life Technologies Corp. v. Promega Corp., 580 U.S. 140 (2017)).

Section 271(g) reaches in the other direction: it makes it an act of infringement to import into the United States, or to sell or use here, a product made abroad by a process patented in the United States. It lets a U.S. process-patent holder stop foreign manufacturers from doing offshore what they could not do here and then shipping the result back in. Our patent basics guide and our overview of global patent litigation strategies take up the cross-border dimensions of these provisions in more detail.

Claim Construction: The Pivotal Battle

Because infringement and validity both turn on what the claims mean, the interpretation of disputed claim terms—claim construction—is the pivotal event in most patent cases. It is not an exaggeration to say that more patent cases are won and lost at claim construction than anywhere else. A favorable construction can make summary judgment of non-infringement almost automatic; an unfavorable one can hand the case to the other side before a single witness testifies. Lawyers sometimes describe the Markman ruling as the moment the verdict is really decided, with the trial a formality that follows.

The Supreme Court held in Markman v. Westview Instruments, Inc., 517 U.S. 370 (1996), that claim construction is a question for the judge, not the jury. The court resolves the parties' disputes over claim meaning, typically at a dedicated Markman hearing, and issues a Markman order; the resulting constructions then become part of the jury instructions, and the jury applies them in deciding infringement and validity. The division of labor is clean in theory: the judge says what the claim means, the jury decides whether the accused product meets it.

The methodology is by now well settled, and it follows a clear hierarchy of evidence. The court looks first to the intrinsic evidence—the claims themselves, the specification, and the prosecution history—because that public record performs the patent's notice function and is far more reliable than after-the-fact litigation testimony tailored to a desired result (Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576 (Fed. Cir. 1996); Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005) (en banc)). Claim terms are given the ordinary and customary meaning they would have had to a person of ordinary skill in the art at the time of invention, read in the context of the entire patent. Phillips famously called the specification "the single best guide to the meaning of a disputed term," while warning courts not to import limitations from the specification into the claims—a line that consumes a remarkable amount of judicial ink, because the specification both defines terms and describes mere examples, and telling the two apart is the heart of the craft. Two patentee-controlled doctrines bookend the analysis: the inventor may act as his own lexicographer, expressly defining a term to mean something other than its ordinary meaning, and under the disclaimer rule, a meaning the applicant clearly surrendered during prosecution to obtain allowance cannot be reclaimed in litigation.

Extrinsic evidence—expert testimony, dictionaries, treatises, and learned articles—may be consulted, but with caution, and always subordinate to the intrinsic record; a dictionary definition cannot override what the patent itself makes clear. On appeal, the Federal Circuit reviews the ultimate legal construction de novo, giving the district court no deference, but reviews any subsidiary factual findings about extrinsic evidence (for instance, what a skilled artisan understood a technical term to mean) for clear error (Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc., 574 U.S. 318 (2015)). That hybrid standard rewards a litigant who can ground its proposed construction in genuine fact findings the appellate court must respect, rather than in pure legal argument the appellate court is free to disregard.

For Sekure Kart, claim construction is the whole ballgame. The asserted claim's fate—and Sekure Kart's—depends entirely on what "random number generation" means. If the term is construed to cover only the conventional methods the specification describes, Sekure Kart's proprietary, entropy-based generator falls outside the claim: there is no literal infringement, and the doctrine-of-equivalents argument becomes far harder for the patentee, because a narrow construction tied to specific disclosed methods leaves little room to argue that a different method is "insubstantially" different. If, instead, the term is construed broadly to cover any generation of random numbers for encryption keys, Sekure Kart may be inside the claim after all. The specification's exclusive discussion of standard methods (a strong argument for Sekure Kart that the inventor possessed and described only those), the prosecution history (did the applicant distinguish prior art on the ground that its method was the conventional kind?), and how a skilled cryptographer would understand the term in 2026 will drive the result. This is why the engineer's technical report, while essential, is only half the work. The other half is the legal argument, made at the Markman stage, about what the claim language covers—and that argument is won with the patent's own intrinsic record, not with flow charts. Our companion piece on conducting a freedom-to-operate analysis for new products shows how a company can do this same claim-by-claim reading before launch, when there is still time to design around the claim rather than litigate it.

Key Defenses to Infringement

An accused infringer has a deep bench of defenses. They fall into several families: non-infringement, invalidity, unenforceability, authorization (license and exhaustion), and equitable defenses. A well-run defense develops several of them in parallel, because they fail and succeed independently—a patent can be valid yet not infringed, or infringed yet invalid, or infringed and valid yet unenforceable for misconduct. The patentee must win on every issue; the accused infringer needs to win on only one.

Non-Infringement

The most common and often the strongest defense is simply non-infringement: the accused product or process does not contain one or more claimed limitations, literally or under the doctrine of equivalents, and so does not infringe. It has the great virtue of placing the burden squarely on the patentee, who must prove infringement by a preponderance of the evidence, and—unlike invalidity—it does not require persuading a court to undo the USPTO's grant. This is Sekure Kart's lead defense. A specialized sub-category, already discussed, is the divided-infringement defense: arguing that, because the claimed steps are performed by different actors and no single entity directs or controls all of them, there is no single act of direct infringement to anchor the claim. Where a non-infringement position is strong, an accused infringer will often paper it with a written non-infringement opinion from competent patent counsel—a careful, objective analysis identifying a reasonable claim construction under which the product falls outside the claim. Such an opinion does double duty: it guides business decisions and, as discussed below, can help rebut a later charge of willful infringement.

Invalidity

A second family attacks the patent itself, on the theory that the USPTO should never have granted it. There is an important thumb on the scale here. Because the USPTO examined and granted the patent, every claim is presumed valid (35 U.S.C. § 282), and an accused infringer must prove invalidity by clear and convincing evidence—a meaningfully higher standard than the preponderance standard the patentee bears on infringement. The Supreme Court reaffirmed that demanding standard, even as to prior art the examiner never considered, in Microsoft Corp. v. i4i Limited Partnership, 564 U.S. 91 (2011). (As we will see, the calculus changes at the USPTO's own Patent Trial and Appeal Board, where the petitioner need only prove unpatentability by a preponderance—one of the main reasons accused infringers flock to inter partes review.) Invalidity defenses come in two broad types.

Prior-art invalidity asserts that the claimed invention was not new or not nonobvious. Anticipation (lack of novelty under § 102) requires that a single prior-art reference disclose every claim element, expressly or inherently, arranged as in the claim. It is a "strict identity" mirror image of the all-elements rule: instead of comparing the claim to an accused product, you compare the claim to a reference, and if a single reference contains every limitation, the claim is anticipated and void. The single-reference requirement is strict—you cannot anticipate by stitching two references together. Obviousness (§ 103) is the more common and more flexible attack, and it typically combines multiple references, asking whether the claimed invention as a whole would have been obvious to a person of ordinary skill in the art at the relevant time. Obviousness is analyzed under the factual framework of Graham v. John Deere Co., 383 U.S. 1 (1966)—the scope and content of the prior art, the differences between the prior art and the claims, the level of ordinary skill, and the "secondary considerations" such as commercial success, long-felt need, and the failure of others—as refined by the flexible, common-sense approach of KSR International Co. v. Teleflex Inc., 550 U.S. 398 (2007), which rejected a rigid requirement of an explicit "teaching, suggestion, or motivation" to combine references. Whether prior art is assessed under the pre-AIA (first-to-invent) or the AIA (first-inventor-to-file, for effective filing dates on or after March 16, 2013) regime determines what even counts as prior art; our guide to overcoming obviousness rejections and our patent basics guide explain that distinction in depth.

Non-prior-art invalidity attacks the patent's compliance with the statutory drafting requirements—chiefly § 112, the provision that polices whether the patentee actually disclosed what it claimed. A patent can be invalid for failing the written-description requirement (the specification must show the inventor possessed the full claimed invention, not merely a wish to claim it; Ariad Pharmaceuticals, Inc. v. Eli Lilly & Co., 598 F.3d 1336 (Fed. Cir. 2010) (en banc)); the enablement requirement (the specification must teach a skilled person to make and use the full scope of the claim without undue experimentation—a requirement the Supreme Court recently enforced with vigor in Amgen Inc. v. Sanofi, 598 U.S. 594 (2023), invalidating broad antibody "genus" claims that taught only a fraction of the claimed compounds); or the definiteness requirement (the claims must inform skilled artisans of the invention's scope with reasonable certainty; Nautilus, Inc. v. Biosig Instruments, Inc., 572 U.S. 898 (2014), which replaced the more forgiving "insolubly ambiguous" test). A § 112 attack is the natural complement to claim construction: if the patentee fights for a construction broad enough to cover the accused product, the accused infringer can often turn that breadth against it, arguing the specification never enabled or described anything that wide. (Failure to disclose the best mode of practicing the invention is no longer a ground for invalidity or unenforceability in proceedings governed by the AIA, though the duty to disclose it formally remains on the books.)

A third invalidity ground, patent-ineligible subject matter under § 101, is especially important for software and business-method patents—and therefore sits squarely in Sekure Kart's path. Laws of nature, natural phenomena, and abstract ideas are not patentable, no matter how cleverly claimed. Under the now-ubiquitous two-step framework of Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014), and Mayo Collaborative Services v. Prometheus Laboratories, Inc., 566 U.S. 66 (2012), a court first asks whether a claim is "directed to" a patent-ineligible concept such as an abstract idea, and if so, asks at step two whether the claim's elements, individually and in combination, add "significantly more"—an inventive concept—that transforms the abstract idea into a patent-eligible application, or whether they merely recite the idea implemented on generic, conventional computer components. A defendant facing a software or internet patent will routinely test the claims against Alice, frequently on a motion to dismiss before discovery even begins, because § 101 is a question of law that can sometimes be resolved on the pleadings. An encryption-related claim that, stripped of its jargon, amounts to a familiar mathematical or business concept ("generate a random key and use it to encrypt a transaction") running on an ordinary computer is a prime candidate for an Alice challenge—which is exactly why Sekure Kart's counsel will read the asserted claims with Alice in one hand. Our deeper treatment lives in patent eligibility after Alice.

Inequitable Conduct

A patent—every claim of it—can be rendered unenforceable in its entirety if the patentee engaged in inequitable conduct before the USPTO: breaching the duty of candor and good faith that every applicant owes the examiner, by withholding or affirmatively misrepresenting material information during prosecution. Practitioners call it the "atomic bomb" of patent litigation, because a single act of inequitable conduct can vaporize an entire patent and, in egregious cases, infect related patents in the same family. Because the defense sounds in fraud, it must be pleaded with particularity under Rule 9(b)—the who, what, when, where, and how of the alleged deception.

For years the defense was overused, asserted reflexively in nearly every case until it threatened to swallow the system. The Federal Circuit reined it in, en banc, in Therasense, Inc. v. Becton, Dickinson & Co., 649 F.3d 1276 (Fed. Cir. 2011), tightening both elements. The accused infringer must now prove, by clear and convincing evidence, (1) a specific intent to deceive the USPTO—meaning the applicant knew of the reference, knew it was material, and made a deliberate decision to withhold it; mere negligence, or even gross negligence, will not do, and deceptive intent must be the "single most reasonable inference" from the evidence—and (2) but-for materiality, meaning the USPTO would not have allowed the claim had it known the truth. Therasense preserved a narrow exception for affirmative egregious misconduct (such as filing a fabricated declaration), where materiality is presumed. A patent owner worried about a prior-art reference that should have been disclosed can sometimes cure the problem in advance through supplemental examination under § 257, asking the USPTO to consider, reconsider, or correct information; if the office concludes the information raises no substantial new question of patentability (or addresses it on reexamination), that information generally cannot later be used to hold the patent unenforceable. Our dedicated guides to inequitable conduct and to finding evidence of inequitable conduct in patent prosecution work through the doctrine and the discovery that feeds it.

License and Patent Exhaustion

If the accused infringer has the patentee's permission, it has a complete defense—infringement requires acting without authority, so authority is the end of the inquiry. Permission may take the form of an express license (a written agreement specifying the licensee, the licensed products, and the licensed patents, interpreted under ordinary state-law contract principles) or an implied license (inferred from conduct—for example, where a patentee sells an article that has no use except to practice the patent, under circumstances implying a right to use it; Zenith Electronics Co. v. PDI Communications Systems, Inc., 522 F.3d 1348 (Fed. Cir. 2008)). A company evaluating an infringement threat should always trace its supply chain: a component bought from a licensed source may carry an implied license that defeats the claim. For the patentee's side of this equation—how to structure, value, and price the very licenses that create these defenses—see our guide on how to license your patent, from valuation to term sheet.

Closely related, but importantly different, is patent exhaustion (the "first-sale" doctrine). Exhaustion provides that once a patentee makes an authorized first sale of a patented article, it has received its reward and exhausts its patent rights in that particular article; it cannot use patent law to control or extract further royalties on the downstream use or resale of that item. The Supreme Court held in Quanta Computer, Inc. v. LG Electronics, Inc., 553 U.S. 617 (2008), that an authorized sale of an article that substantially embodies a patent exhausts the patentee's rights—even where the article is a component sold under a license, and even as to method claims the component substantially embodies. Then, in the blockbuster Impression Products, Inc. v. Lexmark International, Inc., 581 U.S. 360 (2017), the Court swept aside decades of Federal Circuit law (notably Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700 (Fed. Cir. 1992)) and held two things: a sale exhausts patent rights regardless of any post-sale restriction the patentee tries to attach (so Lexmark could not use patent law to stop refurbishers from refilling its toner cartridges, however clearly its single-use restriction was "communicated"), and even an authorized sale abroad exhausts U.S. patent rights, overturning Jazz Photo Corp. v. International Trade Commission, 264 F.3d 1094 (Fed. Cir. 2001). The Court grounded the rule in the common law's ancient hostility to restraints on the alienation of personal property—you bought it, you own it.

Exhaustion has limits worth knowing. It does not authorize making new copies of a patented article: a farmer who buys patented seed may grow and sell a crop, but may not save and replant the next generation of patented seeds, because that is making, not merely using (Bowman v. Monsanto Co., 569 U.S. 278 (2013)). And exhaustion reaches only the article actually sold and patents it embodies, not separate, unembodied claims. One contrast is essential to keep straight: an implied license can be negated by clear contractual language disclaiming any downstream license, but exhaustion cannot—once there is an authorized sale, no boilerplate "no license granted" recital revives the patentee's exhausted rights. A company in Sekure Kart's position should always check whether the components, libraries, or cryptographic modules it incorporates came from a licensed or authorized source, because an exhaustion or license defense, if available, can end the case without ever reaching claim construction.

Equitable Estoppel, Laches, and the Time Limits

Finally, equitable estoppel can bar a claim where the patentee, by its affirmative assurances or misleading conduct, led the accused infringer to believe it would not enforce the patent, and the accused infringer relied on that belief to its detriment (for example, by investing in a product line). The classic A.C. Aukerman formulation requires misleading conduct, reliance, and material prejudice. The defense surfaces often in standards-setting contexts, where a patent owner that committed to a standards body to license its standard-essential patents on fair, reasonable, and non-discriminatory (FRAND) terms—think the patents baked into Wi-Fi, 5G, or video codecs—may be estopped from later asserting them as though no such commitment existed. We explore that world in our guide to standard-essential patents and FRAND licensing, and the broader equitable toolkit in understanding equitable defenses: laches, acquiescence, waiver, and equitable estoppel.

A word on timing, because it has changed. Laches—unreasonable delay in suing—is no longer a defense to a damages claim brought within the Patent Act's six-year window, after the Supreme Court's decision in SCA Hygiene Products Aktiebolag v. First Quality Baby Products, LLC, 580 U.S. 328 (2017), which reasoned that courts may not use laches to override a limitations period Congress has set. That ruling does not abolish the statutory limit itself: under § 286, a patentee may recover damages only for infringement occurring within six years before the complaint (or counterclaim) is filed—a hard cap on the look-back period, not a deadline to sue. Equitable estoppel, which turns on the defendant's reliance rather than mere delay, survives SCA Hygiene intact.

Remedies

A patentee that proves infringement of a valid, enforceable claim may obtain monetary damages and, in appropriate cases, an injunction. The remedies inquiry is its own full-blown contest, typically waged through dueling damages experts, and it is where many cases are ultimately settled, because the size of the exposure—not just the fact of liability—drives the deal.

Monetary Damages

The Patent Act guarantees damages "adequate to compensate for the infringement," but in no event less than a reasonable royalty for the infringer's use of the invention (35 U.S.C. § 284). A reasonable royalty is the floor, and it is the most common measure of damages. The standard tool for computing it is the hypothetical negotiation: courts reconstruct the royalty a willing patentee and a willing infringer would have agreed to in a negotiation taking place just before infringement began (assuming both knew the patent was valid and infringed), assessed through the fifteen-factor framework of Georgia-Pacific Corp. v. United States Plywood Corp., 318 F. Supp. 1116 (S.D.N.Y. 1970). The factors consider established royalties for the patent, rates for comparable licenses, the commercial relationship between the parties, the advantages of the invention over the prior art, and much more. A royalty can be structured as a lump sum paid up front or as a running royalty (a rate multiplied by a royalty base), and evidence about which form fits often comes from the parties' own licensing histories and from custom in the industry.

The Federal Circuit has policed royalty theories with growing rigor, insisting that the number reflect what was actually taken. It rejected mechanical shortcuts like the so-called "25% rule of thumb" as fundamentally untethered to the facts of any case (Uniloc USA, Inc. v. Microsoft Corp., 632 F.3d 1292 (Fed. Cir. 2011)). And, most importantly, it requires careful apportionment: where the patented feature is one component of a larger multi-component product, damages must be tied to the value of the patented feature alone, not to the revenue of the whole product, unless the patentee can prove under the strict entire-market-value rule that the patented feature is what drives consumer demand for the entire product (VirnetX, Inc. v. Cisco Systems, Inc., 767 F.3d 1308 (Fed. Cir. 2014)). Apportionment is the central battleground of modern patent damages—the difference between a royalty on a $5 chip and a royalty on a $1,000 phone—and the patentee bears the burden of isolating the incremental value its invention contributed (Prism Technologies LLC v. Sprint Spectrum L.P., 849 F.3d 1360 (Fed. Cir. 2017)). Damages also cannot sweep in activity that is not infringement, such as foreign sales the U.S. patent does not reach (Enplas Display Device Corp. v. Seoul Semiconductor Co., 909 F.3d 398 (Fed. Cir. 2018)). The same apportionment instinct—pay only for what the protected feature is worth—runs through IP damages generally; our companion piece on damages apportionment in trademark cases shows the parallel principles at work in a neighboring field.

A patentee that competes with the infringer in the same market may aim higher than a royalty and seek lost profits: the profits it actually would have earned but for the infringement, which can dwarf a reasonable royalty. The classic test asks whether the patentee can show (1) demand for the patented product, (2) the absence of acceptable non-infringing substitutes, (3) its manufacturing and marketing capacity to meet the demand, and (4) the profit it would have made (Panduit Corp. v. Stahlin Bros. Fibre Works, Inc., 575 F.2d 1152 (6th Cir. 1978)). The patentee must establish but-for causation; once it does, the burden shifts to the infringer to show that customers would have bought a non-infringing alternative instead. The second Panduit factor—non-infringing substitutes—is usually the fight, because a robust alternative on the market means the patentee would have lost many of those sales anyway, even without infringement.

In egregious cases, the court may enhance damages up to treble the compensatory award for willful infringement (§ 284). For years a rigid two-part test made willfulness hard to prove, but the Supreme Court swept it away in Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U.S. 93 (2016), restoring broad discretion to district courts to punish "egregious cases of misconduct beyond typical infringement"—the "wanton and malicious pirate" the statute has targeted since the founding era. After Halo, the analysis proceeds in two steps: the jury first decides whether the infringer acted willfully, judged by the infringer's knowledge and culpability at the time of the challenged conduct (a defense cooked up later for litigation does not retroactively cleanse earlier willful behavior, WBIP, LLC v. Kohler Co., 829 F.3d 1317 (Fed. Cir. 2016)); the court then decides, in its discretion and weighing case-specific factors, whether and how much to enhance (Eko Brands, LLC v. Adrian Rivera Maynez Enterprises, Inc., 946 F.3d 1367 (Fed. Cir. 2020); SRI International, Inc. v. Cisco Systems, Inc., 14 F.4th 1323 (Fed. Cir. 2021)). Two practical points loom large here. First, willfulness usually requires that the infringer knew of the patent—which is exactly why a cease-and-desist letter, by supplying that knowledge, can start the willfulness clock and raise the stakes of continuing the accused activity. Second, by statute an accused infringer's failure to obtain an opinion of counsel may not be used to prove willfulness or inducement (§ 298)—though many companies still obtain a competent non-infringement or invalidity opinion as affirmative evidence of good faith. We unpack that dynamic in the shield of good faith: how clearance searches, attorney opinions, and USPTO approval can protect against willful-infringement claims.

The court may also award attorney's fees to the prevailing party in exceptional cases—those that "stand out from others" in the substantive weakness of a party's litigating position or the unreasonable manner in which the case was litigated, judged by the totality of the circumstances and reviewed deferentially for abuse of discretion (Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S. 545 (2014); Highmark Inc. v. Allcare Health Management System, Inc., 572 U.S. 559 (2014)). Octane lowered the bar that had made fee awards rare, and it cuts both ways—an overreaching patentee that brings a baseless suit is as exposed as a defendant who litigates abusively.

Finally, two statutory limits bound any damages award. As noted, recovery reaches back only six years before suit (§ 286). And under the marking statute (§ 287), a patentee that sells unmarked patented products generally cannot collect damages for the period before the infringer received actual notice of the patent (such as a specific charge of infringement); a cease-and-desist letter is one classic way to supply that notice and start the damages clock, which is yet another reason these letters carry so much strategic weight. Our guide to understanding patent marking requirements explains how to comply and what marking (including "virtual" marking) does and does not buy.

Injunctive Relief

The most powerful remedy is an injunction ordering the infringer to stop—because money can be paid and absorbed, but an injunction can shut a product line down. Yet an injunction is no longer automatic, as it once effectively was. Under eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006), a patentee that wins must still satisfy the traditional four-factor test of equity: (1) that it has suffered irreparable harm; (2) that money damages are inadequate to compensate for that harm; (3) that the balance of hardships favors an injunction; and (4) that the public interest would not be disserved by one. The decision famously rejected the notion that a patent's right-to-exclude entitles every winning patentee to an injunction as a matter of course.

The practical consequences are stark and depend on who the patentee is. A patentee that competes with the infringer in the same market—losing sales, market share, and goodwill it cannot fully recover in dollars—is far more likely to show irreparable harm and win an injunction. A patentee that does not practice the patent, does not compete, and is content to license for royalties (the paradigm of the patent-assertion entity) will find irreparable harm hard to show, because money is precisely what it wanted all along. This dynamic shapes settlement leverage enormously. The payment processor threatening Sekure Kart is an operating competitor, so it may credibly seek not just a royalty but an order that stops Sekure Kart's business—an existential threat that raises the stakes far above a damages number and changes the entire posture of any settlement talk.

Procedure and Strategy

Pre-Suit Investigation and the Accused Infringer's Playbook

Before filing, a patentee must conduct a reasonable pre-filing investigation sufficient to form a good-faith belief in infringement; Rule 11 requires it, and courts have sanctioned patentees who sued first and analyzed the claims later. A responsible patentee maps the asserted claims against the accused product element by element, searches for invalidating prior art that a defendant will surely find, and confirms it owns the patent and has standing.

A defendant that receives a demand letter—Sekure Kart's situation—should move methodically rather than emotionally. The playbook: (1) preserve documents immediately and instruct the team to stop deleting anything potentially relevant, because spoliation can be more damaging than the underlying claim; (2) build a non-infringement position, claim by claim and element by element, ideally memorialized in a written opinion of counsel; (3) build an invalidity position in parallel, commissioning a prior-art search and—for software or business-method claims—a § 101/Alice analysis; (4) assess damages exposure, including whether the patentee marked its products and whether the letter has started the willfulness and notice clocks; (5) consider challenging the patent at the USPTO through inter partes review; and (6) evaluate whether the company holds counter-patents the patentee might infringe, which can transform a one-sided demand into a mutual standoff that settles. The detailed technical report Sekure Kart commissioned is exactly the kind of work product that anchors a non-infringement position—provided, as always, that it is paired with the legal claim-construction argument that gives it force.

Declaratory Judgment and Notice Letters

A potential defendant need not sit and wait to be sued on the patentee's timetable. It may file a declaratory-judgment action asking a court to declare the patent not infringed, invalid, or unenforceable, thereby seizing the choice of forum and timing—a meaningful advantage given how much outcomes vary by district. The only prerequisite is a justiciable case or controversy. The Supreme Court broadened that gateway in MedImmune, Inc. v. Genentech, Inc., 549 U.S. 118 (2007), replacing the old, rigid "reasonable apprehension of suit" test with an all-the-circumstances inquiry into whether there is a real and immediate dispute.

The practical consequence is one every patent owner must reckon with before writing a demand letter: an aggressive cease-and-desist letter can itself manufacture declaratory-judgment jurisdiction, handing the recipient the power to sue first in a forum of its own choosing. This is the central tension in our cease-and-desist guide—the letter that is too soft accomplishes nothing, while the letter that is too aggressive may surrender the choice of forum. Note one more wrinkle that protects accused parties: even in a declaratory-judgment action brought by a licensee, the patentee still bears the burden of proving infringement (Medtronic, Inc. v. Mirowski Family Ventures, LLC, 571 U.S. 191 (2014)). Flipping the caption does not flip the burden.

Venue, Joinder, and Local Patent Rules

Beyond the TC Heartland venue limits, the America Invents Act tightened the rules on joinder. A patentee generally may not sue many unrelated defendants in a single action merely because each independently infringes the same patent; joinder now requires that the claims arise from the same transaction or occurrence and share a common question of fact (§ 299). This provision was aimed squarely at the practice—common before the AIA—of naming dozens of unrelated companies in one sprawling complaint, and it has reshaped how assertion campaigns are structured. Separately, many of the busiest patent districts have adopted local patent rules that impose an early, disciplined disclosure schedule: the patentee must serve infringement contentions (a claim chart mapping each asserted claim to the accused product) and the accused infringer must serve invalidity contentions (identifying the prior art and theories) well before claim construction, forcing both sides to commit to positions early and crystallizing the disputed claim terms that the Markman process will resolve.

Forums Beyond the District Court

District-court litigation is not the only arena, and choosing the right forum—or fighting on several fronts at once—is a strategy in itself.

The International Trade Commission (ITC), acting under Section 337 of the Tariff Act of 1930 (19 U.S.C. § 1337), hears patent disputes involving imported articles. It cannot award damages, but it offers something often more valuable: an exclusion order directing U.S. Customs to bar infringing imports at the border, plus cease-and-desist orders against domestic sale of already-imported goods. ITC cases move fast (typically to a final determination in well under two years) and the bite of an exclusion order gives patentees enormous leverage, which is why the ITC has become a favored venue for patentees who can satisfy its "domestic industry" requirement—frequently litigating the same patents in parallel with a district-court damages suit.

The USPTO offers a suite of validity challenges that increasingly run alongside, and sometimes derail, district-court litigation. Ex parte reexamination (§ 302) lets anyone ask the office to reexamine an issued patent over prior-art patents and printed publications. The AIA added the heavyweight post-grant proceedings before the Patent Trial and Appeal Board (PTAB): inter partes review (IPR) (§§ 311–319), a trial-like adjudication that challenges claims as anticipated or obvious over patents and printed publications, decided under a preponderance standard (lower than the clear-and-convincing standard that governs in district court); and post-grant review (PGR) (§§ 321–329), a broader challenge—reaching § 101 and § 112 grounds, not just prior art—available only within nine months of issuance for AIA patents. IPR has become the accused infringer's weapon of choice, because the lower burden of proof and the PTAB's technical expertise make it a more hospitable forum than a lay jury for an invalidity attack.

But these proceedings carry consequences that demand early strategic decisions. They trigger estoppel: an IPR petitioner who receives a final written decision generally cannot later relitigate, in court or at the ITC, any ground it raised or reasonably could have raised in the IPR (§ 315(e)). So an accused infringer must decide, early, whether to commit its best prior art to the PTAB (and risk estoppel on grounds it does not raise) or hold it for the district court. It must also weigh whether to seek a stay of the litigation pending the PTAB's decision—stays are common when an IPR is instituted before trial, and a successful IPR can end the case entirely. Finally, parties frequently resolve patent disputes through arbitration or mediation, which the Patent Act expressly authorizes for validity and infringement (§ 294) and which license and settlement agreements routinely require; we weigh the tradeoffs in our guide to arbitration, mediation, and choosing a dispute-resolution forum.

Back to Sekure Kart

Return to the encryption-key startup with the cease-and-desist letter on the desk. What should it actually do? The analysis in this guide gives a clear, sequenced plan.

First, read the claims and apply the all-elements rule. The founders have already done the most important thing right: they identified that the asserted claim requires a "random number generation" step and that Sekure Kart's proprietary generator appears to be something the patent never describes. That is a genuine literal non-infringement position, and the engineer's technical report documenting the differences is the right foundation—provided counsel converts it into a written non-infringement opinion grounded in a reasonable claim construction.

Second, anticipate the doctrine of equivalents. The patentee will argue that Sekure Kart's differences, however numerous, are insubstantial—same job, same way, same result. Sekure Kart must be ready to show that at least one difference is substantial, and it must recognize that this fight and the literal-infringement fight both collapse into a single question: what does "random number generation" mean? Sekure Kart's strongest argument is that the specification describes only conventional methods, so the term should be construed narrowly to exclude its physical-entropy approach—an argument to be won at the Markman hearing, on the intrinsic record, with the engineer as a supporting expert rather than the star.

Third, build defenses in parallel. Sekure Kart should commission a prior-art search to test the patent's validity for anticipation (§ 102) and obviousness (§ 103), and—because this is encryption software—should seriously evaluate a § 101 / Alice challenge attacking the claims as a patent-ineligible abstract idea (a cryptographic algorithm) implemented on a generic computer. It should trace its supply chain for any license or exhaustion defense. And it should weigh petitioning for inter partes review at the PTAB—where the lower burden of proof favors the challenger—while watching the estoppel and stay consequences, and consider whether the aggressive letter has handed it declaratory-judgment jurisdiction under MedImmune to choose its own forum and sue first.

Fourth, assess exposure and leverage. Because the patentee is an operating competitor, it may seek an injunction under eBay to stop Sekure Kart rather than merely collect a royalty—a far graver threat that shapes everything. The letter has also likely supplied the knowledge that starts the willfulness clock under Halo and the notice that starts the damages clock under § 287, so continuing the accused activity carries real risk. Patent litigation is expensive and uncertain for both sides, which is why the overwhelming majority of patent disputes settle. Sekure Kart's goal is not to win a moral victory; it is to make its non-infringement and invalidity positions strong enough that settlement—or a design-around—comes on terms the company can live with. If the payment processor was looking for a fight, Sekure Kart can give it one. But it should bring a strategy, not just a stack of flow charts.

Frequently Asked Questions

Do I infringe a patent if I never knew it existed? Yes, for purposes of direct infringement, which is a strict-liability tort: independent invention and ignorance of the patent are no defense (35 U.S.C. § 271(a)). Knowledge matters in two other places, though. It is a required element of indirect infringement (inducement and contribution), and it is central to willfulness, which can triple your damages. So while not knowing about a patent will not save you from liability, learning about one (for instance, through a demand letter) and pressing on can make things much worse.

Is "patent pending" enough to sue me? No. A pending application confers no enforceable exclusive rights; only an issued patent can be infringed. A published application can, however, support a claim for provisional rights—a reasonable royalty—but only after the patent actually issues and only if the infringer had actual notice of the published application and the claims issued in substantially the same form. Until issuance, "patent pending" is a warning, not a weapon.

The accused product is better than what the patent describes. Doesn't that mean I don't infringe? No, and this is one of the most common and costly misconceptions. Infringement asks whether your product contains every element of a claim, not whether it is similar overall or even improved. Adding features, or doing the claimed thing better, does not avoid infringement; a product can be a brilliant improvement and still practice every limitation of an older patent's claim. The only way to escape a claim is to lack at least one of its elements (literally and under the doctrine of equivalents)—or to invalidate the claim.

How much does patent litigation cost? A lot, and the uncertainty is part of the cost. Fees through trial commonly reach into the millions of dollars, driven by claim construction, expert discovery, and the technical depth involved. That expense is the single biggest reason most cases settle, and it is why an accused company should price out its exposure and its defense costs early—the economically rational outcome is frequently a negotiated license or a design-around, not a jury verdict.

Can I just challenge the patent at the Patent Office instead of going to court? Often, yes, and it is frequently the better forum for an invalidity attack. Inter partes review before the PTAB lets you challenge claims as anticipated or obvious over patents and printed publications under a preponderance standard—easier than the clear-and-convincing standard a court applies. But IPR is limited to certain prior-art grounds, carries estoppel that can bar you from re-raising those grounds later, and runs on its own timeline. Whether to file IPR, when, and whether to seek a stay of any parallel lawsuit are early strategic decisions to make with counsel.

What is the difference between an invalidity defense and an exhaustion or license defense? An invalidity defense says the patent should never have issued (it claims something old, obvious, or improperly disclosed) and asks the court—or the USPTO—to undo the grant against the high clear-and-convincing standard. A license or exhaustion defense concedes the patent may be perfectly valid but says you had authority to do what you did: either because you (or your supplier) were licensed, or because the patentee already made an authorized first sale of the very article and exhausted its rights in it. Exhaustion and license defenses are powerful because they can end a case quickly, often without ever reaching the costly battle over claim construction.

Leading Authorities at a Glance

Practical Takeaways

For an accused business, the discipline is to start with the claims and the all-elements rule. Identify every limitation of every asserted claim and test, element by element, whether your product or process includes it—literally and under the doctrine of equivalents. A genuine missing element is your strongest and cheapest defense, because it puts the burden on the patentee and does not require unwinding the USPTO's grant. But never forget that the dispute will turn on claim construction, so pair your technical analysis with a legal argument about what the claim terms mean, drawn from the patent's own specification and prosecution history. Build defenses in parallel—non-infringement, invalidity (anticipation, obviousness, § 112, and, for software, § 101 eligibility), and any license or exhaustion defense lurking in your supply chain. Consider inter partes review and a possible declaratory-judgment action. Above all, preserve documents, mind the willfulness and marking-notice consequences of continuing after you receive notice, and bring in experienced patent litigation counsel early, while you still have options.

For a patent owner weighing enforcement, the same map shows where your position is strong or weak. Can you prove every element is present in the accused product? Will your claim construction hold up at Markman? Can your patent survive an invalidity and § 101 challenge under the lower PTAB burden? And can you show the damages or irreparable harm that justify your demand? Remember the traps the modern doctrine has set for the overconfident plaintiff: an aggressive demand letter can hand the recipient a declaratory-judgment action under MedImmune; an injunction is not automatic after eBay, and depends heavily on whether you compete with the infringer; your damages will be tested rigorously for apportionment under VirnetX; and a baseless suit can expose you to attorney's fees under Octane. Investigate before you assert (Rule 11), mark your products to preserve damages, and calibrate your strategy—and your tone—to the genuine strength of your position. If licensing rather than litigation is the goal, our guide to licensing your patent, from valuation to term sheet shows how to turn the same rights into revenue.

For everyone, the unifying lesson is that patent infringement is a precise, claim-driven inquiry, not a contest of general resemblance or who got there first. What constitutes infringement is the presence of every claimed element in the accused product or process; what defeats it is the absence of even one, a successful attack on the patent's validity or enforceability, or proof that the activity was authorized. Patent litigation is expensive and uncertain, which is precisely why most disputes settle—and why the party that understands the claims, the defenses, and the remedies, and acts on that understanding early, is almost always the party that fares best.

Related Articles

• Comprehensive Guide to Patent Infringement Litigation: From Summary-Judgment Denial to Post-Trial — what happens once a case clears summary judgment and heads to trial and appeal.
• Conducting Freedom-to-Operate Analysis for New Products — clearing your product against existing patents before launch, when designing around is still cheap.
• Patent Basics and General Information Concerning Patents — the framework, patentability, and the patent lifecycle.
• How to License Your Patent: From Valuation to Term Sheet — turning the right to exclude into licensing revenue.
• Patent Eligibility After Alice — the § 101 defense for software and business-method claims.
• Overcoming Obviousness Rejections: A Comprehensive Guide to Section 103 Analysis — the most common invalidity attack, up close.
• Inequitable Conduct in Patent Prosecution: Navigating the Atomic Bomb of Patent Law — the unenforceability defense after Therasense.
• Standard-Essential Patents and FRAND Licensing in 5G and IoT — estoppel and licensing duties in the standards world.
• Trademark Cease-and-Desist Letters: Sending and Responding — the demand-letter dynamics that also govern patent notice letters.
• Arbitration, Mediation, and Choosing a Dispute-Resolution Forum — resolving patent disputes outside court under § 294.

This article is provided for general informational purposes and does not constitute legal advice. Patent infringement and validity are intensely fact-specific and technical, and the governing law continues to develop. Sekure Kart and its dispute are hypothetical, used only to illustrate doctrine. Consult a registered patent attorney about any specific patent, accused product, demand letter, or dispute.