Patent Basics and General Information Concerning Patents: A Practical Guide

Picture a deal struck with the public itself. You walk up to the government and say: here is exactly how my invention works, every part of it, in writing, so that anyone skilled in the field could build it. In return, the government promises that for a set number of years, nobody else may make, use, or sell that invention without your permission. When the clock runs out, your secret belongs to everyone. That bargain—full disclosure now, in exchange for a temporary right to exclude—is the entire engine of the patent system. It is not a reward for cleverness in the abstract; it is a contract to enrich the public domain on a delay.

That framing matters because it explains nearly every rule that follows. Why must you describe the invention so thoroughly? Because disclosure is your half of the bargain. Why does the patent eventually expire? Because the public is owed the full benefit of the teaching. Why can a patent stop a competitor but not necessarily let you practice your own invention? Because the right you receive is a right to exclude, not a license to act. Keep the bargain in mind and the doctrine stops feeling like a thicket of arbitrary statutes and starts looking like the logical machinery of a trade.

This guide is a detailed, cited reference on U.S. patent law, written to be useful to three audiences at once: inventors deciding whether and how to seek protection, lawyers who want a reliable map of the doctrine, and the merely curious who want to understand the boundaries patents draw around the technologies they use every day. It is plain-spoken but does not flinch from the statutes and cases that govern the field. Patents are one of the four great branches of intellectual property, and if you want to see how they sit alongside their cousins, our guide on the differences among copyright, trademark, patent, and trade secret is the place to start. For a gentler on-ramp to this same material, see Patent Basics: A Plain-English Guide and our companion reference, General Information Concerning Patents. What follows is the long form: the framework, the requirements, the process, and the case law, with worked examples along the way and an extensive table of leading decisions at the end.

A word of caution before we begin. Patent law is technical, federal, and relentlessly developed by the courts. This article is a starting point, not a substitute for advice from a registered patent attorney or agent about a specific invention or dispute. With that on the record, let us open the bargain.

The Constitutional and Statutory Framework

Patents are one of the few areas of American law with a direct constitutional pedigree. Article I, Section 8, Clause 8 of the Constitution empowers Congress "to promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries." Read it closely: the purpose clause comes first ("to promote the Progress"), and the means—exclusive rights—come second and are expressly cabined by "limited Times." The Framers did not grant inventors a natural, perpetual property right in their ideas. They authorized a temporary, instrumental monopoly, justified only insofar as it serves the public good. The Supreme Court still treats that instrumental framing as load-bearing; it described the patent grant as a "public franchise" in Oil States Energy Services, LLC v. Greene's Energy Group, LLC, 584 U.S. 325 (2018).

Congress has exercised that authority since the first Patent Act of 1790, signed by George Washington and administered, for a time, by a three-person board that included Thomas Jefferson. The modern foundation is the Patent Act of 1952, codified in Title 35 of the United States Code, which defines what may be patented, sets the conditions of patentability, and structures the U.S. Patent and Trademark Office. The most consequential modern overhaul is the Leahy-Smith America Invents Act (AIA) of 2011, Pub. L. No. 112-29, which did three big things: it moved the United States from a first-to-invent to a first-inventor-to-file system effective March 16, 2013; it rewrote and broadened the definition of prior art; and it created powerful new administrative procedures for challenging issued patents.

That transition date is more than trivia. The AIA's prior-art rules generally apply only to patents and applications that have, or ever had, at least one claim with an effective filing date on or after March 16, 2013. Patents tracing back to earlier filings live under the older, pre-AIA regime, which used a different definition of prior art and resolved priority disputes through "interference" proceedings rather than by filing date. Practitioners therefore still keep both rulebooks on the shelf—but for any new invention, the AIA framework is the one that governs, and it is the one this guide foregrounds.

It helps to know who does what. The USPTO, an agency of the Department of Commerce, examines applications and grants patents. It does not decide infringement suits. Enforcement belongs to the federal courts: district courts have exclusive jurisdiction over patent infringement actions (28 U.S.C. § 1338), the U.S. Court of Appeals for the Federal Circuit hears essentially all patent appeals (28 U.S.C. § 1295), and the Supreme Court sits above the Federal Circuit. Funneling every patent appeal into a single specialized circuit is a defining feature of the American system, designed in 1982 to bring uniformity to a body of law that had splintered when the regional circuits disagreed. The trade-off—a court so specialized that the Supreme Court has, over the past two decades, repeatedly reversed it for being too rigid—is a recurring theme you will meet again in the case tables below.

What a Patent Actually Is: The Right to Exclude

Here is the single most misunderstood feature of a patent, and it trips up sophisticated people constantly: a patent does not give you the right to make or use your own invention. It gives you the right to stop others. In the statute's words, a patentee may exclude others from "making, using, offering for sale, or selling" the invention throughout the United States, or "importing" it into the country (35 U.S.C. § 271(a)). That is a negative right—a no-trespassing sign, not a building permit.

Why does the distinction matter? Consider a clean hypothetical. Suppose Company A holds a patent on a chair. Inventor B later patents a chair with a folding armrest—a genuine, nonobvious improvement. B's patent is valid and lets B stop everyone, including A, from selling a chair with that armrest. But B cannot sell her own folding-armrest chair either, because any such chair is still a chair, and selling it would infringe A's underlying patent. The two patents "block" each other. The usual resolution is a cross-license: A and B each grant the other permission, and the market gets the improvement. The lesson is that owning a patent never inoculates you against everyone else's. This is exactly why companies invest in freedom-to-operate analysis before launching a product—clearing the field of others' blocking rights is a separate exercise from securing your own.

Two more features round out the picture. First, a patent is territorial. A U.S. patent reaches only conduct in the United States and its territories; to stop infringement in Germany or Japan you need German or Japanese patents. Second, enforcement is entirely on you. Once the patent issues, the USPTO washes its hands of policing it. If a competitor copies your invention, no government lawyer rides to the rescue—you sue, in federal court, on your own dime. A patent, in other words, is a sword you must be willing and able to swing. A patent you cannot afford to enforce is, to a determined infringer, a piece of decorative parchment.

The Three Types of Patents

The Patent Act authorizes three distinct kinds of patents, and confusing them is a classic rookie error.

Utility patents are the workhorses—roughly nine of every ten patents granted—and they protect how an invention works or is used. Section 101 allows a utility patent for "any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof." Those four categories—process, machine, manufacture, composition of matter—are deliberately capacious, sweeping in everything from a pharmaceutical molecule to a manufacturing method to a software-driven device. A utility patent runs twenty years from the earliest effective U.S. filing date to which it is entitled (35 U.S.C. § 154(a)(2)), subject to maintenance fees and the term adjustments discussed later. For a deeper treatment aimed squarely at this category, see Utility Patent Basics.

Design patents protect the ornamental appearance of an article of manufacture—its shape, configuration, or surface ornamentation—not how it functions (35 U.S.C. § 171). Think of the silhouette of a classic glass soda bottle or the arrangement of icons on a phone screen. A design patent filed on or after May 13, 2015 lasts fifteen years from grant (fourteen years for earlier-filed designs) (35 U.S.C. § 173), and—a genuine convenience—no maintenance fees are due. Because design patents protect appearance only, a feature that exists for functional reasons cannot be claimed as a design; that functionality line has become surprisingly contested, as we explore in the evolution of design patent functionality. Design patents also overlap with trade dress, a comparison we draw out in design patents vs. trade dress protection for product configurations, and they can run concurrently with utility patents on the same product, raising the double-patenting wrinkles canvassed in the intersection of design and utility patents.

Plant patents protect a new and distinct variety of plant that has been asexually reproduced—propagated by cuttings, grafting, or budding rather than by seed (35 U.S.C. § 161)—other than a tuber-propagated plant or one found in an uncultivated state. The term, like a utility patent's, is twenty years from filing. (Seed-propagated varieties are protected separately, under the Plant Variety Protection Act administered by the Department of Agriculture, not by the USPTO.) Plant patents are a small, specialized corner of practice; unless you breed roses or apples for a living, you will rarely meet one.

Throughout the rest of this guide, "patent" means utility patent unless we say otherwise—the same convention that governs everyday practice.

The Requirements for Patentability

To earn a utility patent, an invention must clear several independent hurdles. It must fall within eligible subject matter and be useful (§ 101), be novel (§ 102), be nonobvious (§ 103), and the application must satisfy the disclosure requirements of § 112. These are not a balancing test; they are a gauntlet. Fail any single one and the patent falls, no matter how brilliantly the others are met. Examiners and litigants attack them in roughly that order, so we will take them in turn.

Eligible Subject Matter and Utility (Section 101)

Section 101 quietly does two jobs. The first is easy: the invention must be useful—it must have a specific, substantial, real-world utility and must actually operate as claimed. This rarely matters, but it is the reason the perennial perpetual-motion machine never gets a patent: a device that cannot work is not useful, and the examiner is entitled to demand a working model when an asserted utility strains credulity.

The second job is where the litigation lives. Section 101 lists the eligible categories—process, machine, manufacture, composition of matter—but the courts have long carved out three judicial exceptions that no clever drafting can patent: laws of nature, natural phenomena, and abstract ideas. The rationale is that these are the basic tools of all scientific work; let one person own E = mc² or the concept of hedging risk, and you choke off the innovation the system exists to promote.

The modern test for applying those exceptions is the two-step framework of Mayo Collaborative Services v. Prometheus Laboratories, Inc., 566 U.S. 66 (2012), and Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014). Step one asks whether the claim is "directed to" a judicial exception—an abstract idea, a law of nature, or a natural phenomenon. If it is not, the claim is eligible and the inquiry ends. If it is, step two asks whether the claim's additional elements supply an "inventive concept"—something "significantly more" than the exception itself, enough to transform it into a patent-eligible application. The cautionary tale is Alice itself: claims to the abstract idea of intermediated settlement, implemented on a "generic computer," added nothing inventive and were ineligible. Reciting "do it on a computer" does not save an abstract idea; the computer is doing only what computers ordinarily do.

A short hypothetical sharpens the line. Suppose an inventor claims "a method of determining whether to extend credit by comparing an applicant's income to a threshold." That is an abstract idea—a mental judgment a loan officer has performed for centuries—and reciting a generic computer to run the comparison adds nothing significant; it fails Alice. Now suppose the inventor instead claims a specific, unconventional data structure that lets a database resolve credit checks in a fraction of the prior time by reorganizing how records are indexed. That claim is plausibly directed to a concrete technological improvement in how computers operate, not to the abstract idea of "extending credit," and stands a real chance under step one. The dividing line—improving a technology versus merely automating a familiar mental task—is the central battleground for software and business-method patents, a fight we trace in detail in patent eligibility after Alice.

The same exceptions reshaped two other fields. In Bilski v. Kappos, 561 U.S. 593 (2010), the Court held that the "machine-or-transformation" test is a useful clue but not the sole test for whether a process is eligible, and struck down claims to the abstract idea of hedging. And in Association for Molecular Pathology v. Myriad Genetics, Inc., 569 U.S. 576 (2013), the Court drew a now-famous line in biotechnology: a naturally occurring DNA segment is a product of nature and not patent-eligible merely because it has been isolated from the body, but synthetic complementary DNA (cDNA), which does not occur in nature, is eligible. Subject-matter eligibility remains the most volatile area of patent law, perennially the subject of failed legislative fixes and Federal Circuit panels that cannot agree among themselves.

Novelty and the Prior Art (Section 102)

An invention must be new. Under § 102, as rewritten by the AIA, you cannot patent a claimed invention that was "patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date," or that was already described in an earlier-filed U.S. patent or published application naming someone else (35 U.S.C. § 102(a)). Novelty is measured against the prior art: the universe of patents, publications, and public activities predating your filing.

The defining rule of novelty is the single-reference rule. A claim is "anticipated"—and thus lacks novelty—only if one prior-art reference, standing alone, discloses every element of the claim arranged as in the claim. You cannot defeat novelty by stitching together pieces of two different references; combining references is the job of obviousness, which we reach next. Anticipation is therefore a demanding, all-or-nothing showing, and a missing limitation in the reference is fatal to the challenge.

The AIA reshaped § 102 in two structural ways. First, it pegged priority to the first inventor to file, not the first to invent—so the race is now to the Patent Office, and prompt filing is everything. Second, it globalized prior art: a public use, sale, or disclosure anywhere in the world before your effective filing date now counts, where pre-AIA law limited certain categories to activity in the United States. The catch-all phrase "otherwise available to the public" is broader than people expect. It can sweep in an oral conference presentation, a trade-show demonstration, a YouTube video, a GitHub commit, or a product page that goes live before you file.

There is one mercy: a limited one-year grace period. A disclosure made one year or less before your effective filing date does not count as prior art against you if it was made by you (the inventor) or by someone who got the subject matter from you (35 U.S.C. § 102(b)(1)). But lean on that grace period at your peril. It is personal to the inventor's own disclosures, it lasts only a year, and—crucially—most of the world has no grace period at all. The instant you publish or sell in the United States, your ability to patent in Europe, China, and most other major markets typically evaporates. The safe rule for anyone with global ambitions is brutally simple: file before you disclose.

The on-sale bar deserves its own warning, because it catches founders who think a confidential sale is safe. In Helsinn Healthcare S.A. v. Teva Pharmaceuticals USA, Inc., 139 S. Ct. 628 (2019), the Supreme Court held that a commercial sale can trigger the on-sale bar even when the sale agreement keeps the details of the invention secret. The earlier case of Pfaff v. Wells Electronics, Inc., 525 U.S. 55 (1998), set the two-part trigger: the bar applies once the invention is both (1) the subject of a commercial offer for sale and (2) "ready for patenting." Translated for the boardroom: signing a supply contract for a not-yet-patented product, or accepting a purchase order at a trade show, can start a one-year clock you did not know was running—and in countries without a grace period, can destroy your rights outright.

Nonobviousness (Section 103)

Novelty is necessary but not sufficient. Even an invention no single reference fully discloses cannot be patented if it would have been obvious. Under § 103, a patent is barred if the differences between the claimed invention and the prior art "are such that the claimed invention as a whole would have been obvious before the effective filing date" to a person having ordinary skill in the art (the "POSITA," a hypothetical workaday practitioner, neither genius nor dullard). Unlike anticipation, the obviousness inquiry may combine the teachings of multiple references—which is exactly why it is the single most common ground on which patents are invalidated.

The analytical skeleton comes from Graham v. John Deere Co., 383 U.S. 1 (1966), which set out four inquiries: (1) the scope and content of the prior art; (2) the differences between the prior art and the claims; (3) the level of ordinary skill in the art; and (4) the secondary considerations, also called objective indicia, such as commercial success, long-felt but unmet need, the failure of others, copying by competitors, and industry praise. Those secondary considerations are not afterthoughts; they are real-world evidence that an invention was not as obvious as hindsight makes it appear, and they have rescued many a patent from an obviousness attack.

The modern gloss is KSR International Co. v. Teleflex Inc., 550 U.S. 398 (2007). For years the Federal Circuit had demanded a rigid showing of a "teaching, suggestion, or motivation" (TSM) in the prior art before two references could be combined. KSR loosened that grip, holding that an invention may be obvious when it is no more than "the predictable use of prior art elements according to their established functions," and that an examiner or court may invoke common sense and the ordinary creativity of a skilled artisan. The watchword KSR added to the vocabulary is hindsight bias: because every invention looks obvious once explained, the law insists the analysis be anchored to what the prior art actually taught at the time, not to a roadmap drawn with the patent in hand. Practitioners fighting these rejections will want our deep dive, overcoming obviousness rejections: a guide to Section 103 analysis, and the responding-to-the-office practical guidance in responding to patent office actions.

The Disclosure Requirements (Section 112)

Now we reach the inventor's half of the bargain. In exchange for the right to exclude, you must teach the public how the invention works, and § 112 spells out, in several distinct requirements, how thoroughly. Each is independently capable of sinking a patent.

The written description requirement asks whether the specification shows the inventor was actually "in possession" of the full scope of what the claims cover—did you really invent the whole thing you are claiming, or just a corner of it? The enablement requirement asks whether the specification teaches a skilled artisan how to make and use the full scope of the claimed invention without "undue experimentation." The Supreme Court underscored just how demanding full-scope enablement is in Amgen Inc. v. Sanofi, 598 U.S. 594 (2023). Amgen had claimed an entire genus of antibodies defined by their function—antibodies that bind a particular protein and block it—but disclosed only a relative handful of working examples and a "roadmap" for finding the rest by trial and error. The Court held the claims invalid: if you claim a whole class, you must enable the whole class, and "a roadmap and a research assignment" do not satisfy the statute when the class potentially runs to millions of antibodies. The decision is a sharp reminder that broad functional claims invite a fatal enablement attack, an issue the USPTO confronts daily in the unpredictable arts.

The definiteness requirement (§ 112(b)) demands that the claims "particularly point out and distinctly claim" the invention. In Nautilus, Inc. v. Biosig Instruments, Inc., 572 U.S. 898 (2014), the Court held that a claim is indefinite if, read in light of the specification and prosecution history, it fails to inform skilled artisans about the scope of the invention with "reasonable certainty." Vague, purely subjective claim language ("spaced relationship," "aesthetically pleasing") is where definiteness fights break out. Section 112 also preserves a best mode requirement—the inventor must disclose the best way she knows to carry out the invention—though the AIA stripped away most teeth by making a best-mode failure unavailable as a basis to invalidate or cancel a patent in litigation.

Finally, § 112(f) governs means-plus-function claiming, a drafting style that recites a function ("means for fastening") rather than the structure that performs it. Such a claim is construed to cover only the corresponding structure disclosed in the specification and its equivalents—a narrowing trap for the unwary. In Williamson v. Citrix Online, LLC, 792 F.3d 1339 (Fed. Cir. 2015) (en banc), the Federal Circuit lowered the bar for triggering § 112(f), holding there is no strong presumption against it merely because the word "means" is absent; a functional term like "module" that connotes no definite structure can pull a claim into means-plus-function territory whether the drafter intended it or not.

Claims: The Metes and Bounds of the Right

The claims are the legal heart of the patent. Everything before them—the background, the drawings, the detailed description—is context; the claims are the part that defines what the patent excludes others from doing and the part a court measures an accused product against. A useful analogy: if the specification is the survey describing the land, the claims are the property lines on the deed. Claims come in two flavors. An independent claim stands on its own and states the full combination of elements. A dependent claim refers back to another claim and adds a limitation ("The widget of claim 1, wherein the housing is aluminum"), narrowing the scope but inheriting everything from the claim it depends on—a hedge in case the broader claim is invalidated.

Interpreting that claim language is claim construction, and it is so central that courts hold dedicated Markman hearings to resolve it before trial. Claim construction is a question of law for the judge, though it can rest on subsidiary factual findings (for instance, what a technical term meant to skilled artisans), and the Supreme Court held in Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc., 574 U.S. 318 (2015), that those subsidiary factual findings are reviewed only for clear error rather than de novo. A related doctrine, double patenting, stops an inventor from extending a patent's life by obtaining a second patent on the same or a patentably indistinct invention—usually cured, when possible, by filing a "terminal disclaimer" that ties the later patent's expiration to the earlier one's.

Ownership and Inventorship

Patent rights vest initially in the inventor—not the inventor's employer, not the company that paid for the lab, but the human being or beings who conceived the invention. Inventorship is a legal determination keyed to conception of the claimed invention, and it is narrower than people assume. Under § 116, two or more people who jointly conceive an invention are joint inventors; but someone who merely supplied money, ran an experiment exactly as told, or contributed only ordinary skill in reducing another's idea to practice is not an inventor. Getting inventorship wrong is dangerous, though usually fixable: § 256 allows correction of an erroneously named or omitted inventor.

One modern question deserves a flag. Can an artificial-intelligence system be named as an inventor? The Federal Circuit answered no in Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022), cert. denied: the Patent Act's repeated use of "individual" and "himself or herself" means an inventor must be a natural person. The USPTO has since issued guidance on inventions made with AI assistance, requiring that a natural person have made a "significant contribution." We unpack that fast-moving area in AI-generated inventions: who owns what the machine creates.

A patent is personal property that can be bought, sold, mortgaged, and licensed (35 U.S.C. § 261), and assignments must be in writing. Two default rules surprise nearly everyone and make written agreements indispensable. First, joint owners: absent an agreement otherwise, each co-owner may independently make, use, and license the invention without the consent of, and without owing any accounting to, the other owners (35 U.S.C. § 262). That means a single co-owner can unilaterally license a competitor and pocket the proceeds—and can also single-handedly torpedo an infringement suit, because all co-owners generally must join as plaintiffs. Co-ownership without a written agreement is a recipe for paralysis.

Second, employee inventions. Contrary to widespread assumption, an employee who is not specifically hired to invent generally owns the inventions she creates—even on company time—unless she has signed an assignment agreement. Where there is no assignment, an employer who supplied resources may at most hold a non-exclusive, non-transferable "shop right" to use the invention. Because the default cuts against employers, well-drafted assignment provisions are standard in every serious technology employment and contractor relationship; for the drafting nuances, see employee invention assignment agreements. Recording an assignment with the USPTO provides constructive notice and priority against later purchasers (§ 261)—the patent-world equivalent of recording a deed.

The Patent Application and Prosecution

Getting a patent—the process called prosecution—is a negotiation with the government that typically takes well over two years and often longer. The work demands both legal and technical fluency, which is why a registered patent attorney or agent almost always prepares and prosecutes the application; a poorly drafted claim can permanently narrow or forfeit the very protection you are paying for. Before drafting begins, prudent applicants usually commission a patentability search to assess what the prior art shows and whether the invention is worth the investment—a step we cover in our materials on invention disclosures, including how to prepare an invention disclosure for your patent attorney.

Provisional and Nonprovisional Applications

A utility applicant chooses between two doors. A provisional application (§ 111(b)) is a lower-cost placeholder. It requires a written description and any necessary drawings, but no claims and no oath or declaration; it is never examined; and it goes automatically abandoned after twelve months. Its value is twofold: it locks in an early effective filing date, and it lets the inventor honestly mark products "Patent Pending" while developing the invention—provided a full nonprovisional claiming its benefit is filed within that twelve-month window. Miss the window and the provisional vanishes as if it never existed.

The provisional is a strategic instrument, not a formality, and it has a hidden hazard: it only secures a priority date for subject matter it actually describes well enough to satisfy § 112. A thin, rushed provisional that gestures at an invention without enabling it provides no real priority for the parts it failed to disclose. File a provisional thoroughly, or it is a security blanket with no fabric.

A nonprovisional application (§ 111(a)) is the real thing—the application the USPTO examines and may grant. It must include a specification with at least one claim (§ 112), drawings where needed to understand the invention (§ 113), an inventor's oath or declaration (§ 115), and the required fees (§ 41). The specification follows a conventional architecture—title, cross-references, background, summary, brief description of the drawings, detailed description, claims, and abstract—and must describe the invention in "full, clear, concise, and exact terms." The claims, at the end, set the boundaries of the right.

Filing, Publication, and Examination

Applications are filed electronically through the USPTO's Patent Center, which in late 2023 replaced the legacy EFS-Web and Private PAIR systems and is now the single platform for filing and managing applications. The Office charges filing, search, examination, issue, and maintenance fees, with substantial discounts for small entities (roughly half off) and micro entities (roughly 75% off); the fee schedule changes periodically, so the current USPTO figures should always be consulted rather than memorized. The Office also imposes format requirements, including filing in DOCX format with a surcharge for noncompliance.

Most nonprovisional utility applications are published about eighteen months after the earliest effective filing date (§ 122(b)), at which point the file becomes public. An applicant may opt out of publication only by certifying that no corresponding foreign application has been or will be filed—a narrow option, since most serious applicants do file abroad. Publication carries a bonus: provisional rights under § 154(d). If the published claims and the issued claims are substantially identical and the patentee gave actual notice, the patentee may collect a reasonable royalty for infringement that occurred between publication and grant—a rare bridge across the otherwise-unprotected pendency period.

Then comes the heart of prosecution: examination. A USPTO examiner, assigned by technology, searches the prior art and reviews the application against every statutory requirement, communicating through a document called an office action. Brace yourself: it is entirely routine for some or all claims to be rejected in the first office action—often over § 102 anticipation, § 103 obviousness, or § 112 indefiniteness. A first rejection is the opening move in a negotiation, not a verdict. The applicant must respond to every ground, addressing the examiner's reasoning point by point, and may amend the claims—so long as the amendments add no "new matter" beyond the original disclosure (you cannot retroactively pad a thin specification). A second office action is frequently made final, which constrains the applicant's options but does not end the road. Throughout, applicants and their representatives owe the Office a duty of candor and good faith, including a duty to disclose known material prior art via an Information Disclosure Statement (37 C.F.R. § 1.56). Breach that duty and the consequence is severe: inequitable conduct can render the entire patent unenforceable, under the demanding "but-for materiality" plus specific-intent standard of Therasense, Inc. v. Becton, Dickinson & Co., 649 F.3d 1276 (Fed. Cir. 2011) (en banc), which deliberately raised the bar to curb what the court called an overused "atomic bomb" of patent litigation. We treat that doctrine at length in finding evidence of inequitable conduct in patent prosecution.

Appeals, Continuations, and Patent Term Adjustment

If the examiner digs in, the applicant has options. She may appeal to the Patent Trial and Appeal Board (PTAB) (§ 134) and, from an adverse PTAB decision, to the Federal Circuit or by civil action in district court (§§ 141, 145). Alternatively, to keep examination going, she may file a request for continued examination (RCE) to reopen prosecution after a final rejection, or a continuation application to pursue different or broader claims from the same disclosure while the parent is still pending. Continuation practice is a powerful and sometimes-criticized strategic tool: it lets a patentee keep a "child" application alive for years, watching the market and later drafting claims tailored to a competitor's product.

Because all this back-and-forth can devour the effective term of a patent—remember, the twenty years runs from filing, not grant—Congress built in compensation. Patent term adjustment (PTA) (§ 154(b)) adds days to the term to make up for certain USPTO delays (for example, taking too long to issue a first office action). Patent term extension (PTE) (§ 156) is a separate, narrower mechanism that restores a portion of the term lost to regulatory review, most importantly the FDA's approval of a new drug or medical device—a lifeline in the pharmaceutical world.

Allowance, Issuance, and Maintenance

When the examiner is finally satisfied, the USPTO issues a Notice of Allowance and Fee(s) Due. The applicant must pay the issue fee within the set period or the application goes abandoned—a heartbreaking way to lose a patent at the one-yard line. Pay it, and the patent issues: it receives a number, its file becomes public, and the right to exclude springs into existence. (Not before. "Patent Pending" confers no enforceable rights; it is a warning shot, not a weapon.)

A patent does not coast to the end of its term on autopilot. For utility patents from applications filed on or after December 12, 1980, maintenance fees are due at 3.5, 7.5, and 11.5 years after grant. Miss a payment—even after a six-month grace period with surcharge—and the patent expires prematurely (§ 41(b)). These fees escalate, by design, so that owners must periodically decide whether a given patent still earns its keep; many are deliberately allowed to lapse, returning their inventions to the public early. Design and plant patents, by contrast, require no maintenance fees.

Marking is the unglamorous step that quietly determines how much money you can recover. A patentee who makes or sells a patented article should mark it with the patent number—or use virtual marking, a "Patent" notice plus a freely accessible website listing the relevant numbers (a 2011 AIA innovation that lets companies update the list without re-tooling). The stakes are concrete: under § 287, failure to mark generally bars recovery of damages for infringement occurring before the infringer received actual notice of the patent. In a hypothetical where a company sells an unmarked patented gadget for three years and only then sends a cease-and-desist letter, those three years of pre-notice sales may yield zero damages. The flip side is false marking: labeling an unpatented article as patented, with intent to deceive, is prohibited and can expose the marker to liability. We cover the mechanics in depth in understanding patent marking requirements and its broader sibling, understanding patent and trade dress marking requirements.

After issuance, a patent is mostly beyond the USPTO's reach, but several correction mechanisms survive. A certificate of correction fixes minor clerical or typographical errors. A reissue application (§ 251) corrects a patent that is defective or that claims more or less than the patentee had a right to claim—without adding new matter, and with a two-year window for broadening reissues. Ex parte reexamination (§ 302) lets anyone, including the patentee, ask the Office to reexamine the patent over prior-art patents and publications that raise a "substantial new question of patentability." And supplemental examination (§ 257), an AIA creation, lets a patentee proactively put potentially problematic information before the Office, immunizing the patent against later inequitable-conduct attacks based on that information—a quiet but valuable cleanup tool.

Post-Grant Proceedings Before the PTAB

The AIA's most disruptive contribution was a suite of administrative challenges that let third parties attack issued patents before the PTAB, often faster and cheaper than district-court litigation. These proceedings have become a central front in nearly every significant patent dispute, and an accused infringer's first instinct is frequently to file one.

Inter partes review (IPR) (§ 311) lets a challenger attack a patent's validity on anticipation or obviousness grounds, and only on the basis of patents and printed publications. It is available for any patent, old or new, and a petitioner must generally file within one year of being served with an infringement complaint. The PTAB decides validity by a preponderance of the evidence—a lower bar than the "clear and convincing" standard a defendant faces in district court—which is a large part of why IPRs are so popular with accused infringers. The proceeding runs on a statutory timetable, with a final written decision normally due within one year of institution.

Post-grant review (PGR) (§ 321) is broader but briefer: a challenger may raise essentially any invalidity ground—including § 101 eligibility and § 112 disclosure failures, not just prior art—but must file within nine months of issuance, and PGR is available only for AIA (post-March-16-2013) patents. Ex parte reexamination (§ 302) remains available as a quieter, cheaper alternative based on patents and publications. And under first-inventor-to-file, derivation proceedings (§ 135) resolve whether an earlier-filing applicant actually derived the invention from a later applicant—the AIA's slimmed-down replacement for the old interference contest.

The constitutionality of all this was hotly contested—was the executive branch unconstitutionally stripping a property right that belonged in court?—and the Supreme Court settled it in Oil States Energy Services, LLC v. Greene's Energy Group, LLC, 584 U.S. 325 (2018), upholding IPR on the theory that a patent is a "public franchise" Congress may permit an agency to reconsider. Two companion cases shaped the practice: Cuozzo Speed Technologies, LLC v. Lee, 579 U.S. 261 (2016), held the PTAB's decision to institute review generally unreviewable; and SAS Institute Inc. v. Iancu, 584 U.S. 357 (2018), held that if the Board institutes, it must decide the patentability of all challenged claims, not a subset. Coordinating these parallel tracks—an IPR at the PTAB running alongside a district-court case—is now a core litigation skill, with stays, estoppel under § 315(e), and timing all in play.

Patent Enforcement, Defenses, and Remedies

A patent is worth only what its owner can enforce. A patentee may sue for infringement in federal district court, or—where the infringement involves imports—before the U.S. International Trade Commission (ITC) under Section 337 of the Tariff Act (19 U.S.C. § 1337), which moves fast and can issue exclusion orders directing U.S. Customs to bar infringing products at the border. The ITC cannot award damages, but its speed and the threat of an import ban give it real leverage, especially against foreign manufacturers. What follows is a compressed tour of enforcement; for the full treatment, see our companion guide, what constitutes patent infringement: claims and defenses, along with what constitutes patent infringement and the litigation-focused comprehensive guide to patent infringement litigation.

Types of Infringement

Direct infringement (§ 271(a)) is a strict-liability wrong: anyone who makes, uses, offers to sell, sells, or imports a patented invention without authorization infringes—regardless of intent or knowledge. The accidental, good-faith copyist is liable just the same as the deliberate one. Infringement is judged claim by claim, and the patentee must show that the accused product or process contains every element of an asserted claim—the "all-elements rule." That match can be literal (the accused product reads on the claim word for word) or under the doctrine of equivalents, which reaches insubstantial variations so that an infringer cannot escape by trivial change. The doctrine of equivalents has a major limit: prosecution history estoppel bars a patentee from recapturing, through equivalents, subject matter she surrendered by amendment during prosecution. Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722 (2002), held that a narrowing amendment creates a presumption of surrender, rebuttable only in narrow circumstances—a powerful illustration of how concessions made to the examiner come back to haunt the patentee in court.

Method claims add a wrinkle. Direct infringement of a method ordinarily requires a single entity to perform every step. What about a scheme where one company performs some steps and its customers perform the rest? The Federal Circuit addressed this divided infringement problem in Akamai Technologies, Inc. v. Limelight Networks, Inc., 797 F.3d 1020 (Fed. Cir. 2015) (en banc), holding that the acts of multiple actors can be attributed to one when that actor directs or controls the others' performance, or when they act as a joint enterprise.

Indirect infringement captures those who orchestrate or enable another's direct infringement. Induced infringement (§ 271(b)) reaches one who knowingly aids and abets another's direct infringement—and knowledge matters here. In Commil USA, LLC v. Cisco Systems, Inc., 575 U.S. 632 (2015), the Court held that a defendant's good-faith belief that the patent is invalid is not a defense to inducement (invalidity and infringement are separate questions); and in Limelight Networks, Inc. v. Akamai Technologies, Inc., 572 U.S. 915 (2014), the Court held there can be no inducement without an underlying act of direct infringement by someone. Contributory infringement (§ 271(c)) targets the seller of a component especially made for an infringing use and lacking any substantial noninfringing use. Special statutory regimes round out the picture: § 271(e) makes the mere filing of certain generic-drug applications an artificial "act of infringement" that launches Hatch-Waxman litigation; § 271(f) addresses supplying components for assembly abroad (the Court held in Life Technologies Corp. v. Promega Corp., 580 U.S. 140 (2017), that supplying a single commodity component is not enough); and § 271(g) reaches importing products made abroad by a patented process.

Defenses

The accused infringer has a well-stocked arsenal. Noninfringement—a contention that the accused product or process simply lacks at least one claim limitation—is the cleanest defense and is litigated claim by claim, usually downstream of claim construction. Invalidity is the other great defense, but it carries a heavy burden: an issued patent enjoys a statutory presumption of validity (§ 282), and a challenger in court must prove invalidity by clear and convincing evidence, a markedly higher bar than the preponderance standard at the PTAB. Invalidity can rest on ineligible subject matter (§ 101), anticipation (§ 102), obviousness (§ 103), or a § 112 failure.

Unenforceability through inequitable conduct (Therasense, above) taints the entire patent if the patentee deceived the Office. Beyond these, the defendant can invoke patent exhaustion: the first authorized sale of a patented article exhausts the patentee's rights in that article, so the patentee cannot use patent law to control its downstream resale or use. Impression Products, Inc. v. Lexmark International, Inc., 581 U.S. 360 (2017), pushed exhaustion hard, holding that exhaustion applies even to authorized sales made abroad and even where the patentee purported to impose post-sale resale restrictions. Exhaustion has limits, though: it permits using and reselling the article bought, not making new copies of it—so a farmer could not plant and replant patented self-replicating soybeans across successive seasons without authorization (Bowman v. Monsanto Co., 569 U.S. 278 (2013)).

Other defenses include assignor estoppel—which the Court narrowed in Minerva Surgical, Inc. v. Hologic, Inc., 594 U.S. 234 (2021), confining it to cases where the assignor's invalidity argument contradicts representations made when assigning the patent—and a clutch of equitable defenses. One myth worth dispelling: in SCA Hygiene Products Aktiebolag v. First Quality Baby Products, LLC, 580 U.S. 328 (2017), the Court held that laches is not a defense to a damages claim brought within the six-year statutory damages window of § 286. Delay alone, within that window, does not bar recovery.

Remedies

Win, and several remedies open up. Injunctions (§ 283) are available but, since eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006), no longer automatic. A patentee seeking a permanent injunction must satisfy the traditional four-factor equitable test (irreparable harm, inadequacy of money damages, balance of hardships, and the public interest). eBay quietly transformed enforcement, particularly for patent-assertion entities that do not practice their patents and therefore struggle to show irreparable harm—a shift that pushed many disputes toward damages rather than market exclusion.

Damages (§ 284) are floored at a reasonable royalty and may rise to the patentee's lost profits where the patentee can prove it would have made the infringer's sales. A central, hard-fought issue is apportionment: damages must be tied to the value attributable to the patented feature, not the entire product, when the patented feature is only one component of a multi-feature device. For willful infringement, the court may award enhanced damages up to treble the award; Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U.S. 93 (2016), swept away the Federal Circuit's rigid two-part test in favor of a flexible standard committed to the district court's discretion, reserving enhancement for egregious, willful misconduct. Attorney's fees are available in "exceptional" cases (§ 285); Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S. 545 (2014), relaxed the "exceptional case" standard to one that simply stands out from others, and its companion Highmark Inc. v. Allcare Health Management System, Inc., 572 U.S. 559 (2014), held that the trial court's fee determination is reviewed deferentially for abuse of discretion. Damages can even reach lost foreign profits flowing from domestic infringement in the right posture (WesternGeco LLC v. ION Geophysical Corp., 585 U.S. 407 (2018)).

A final practical constraint: venue. In TC Heartland LLC v. Kraft Foods Group Brands LLC, 581 U.S. 258 (2017), the Court held that a domestic corporation "resides," for patent-venue purposes, only in its state of incorporation. The decision sharply curtailed the concentration of cases in plaintiff-friendly districts and reshaped the geography of patent litigation overnight.

Exploiting Patent Rights: Assignment and Licensing

A patent is an asset, and suing infringers is only one way to monetize it. Because a patent is personal property, its owner can assign (sell) it outright in writing (§ 261)—and the buyer should promptly record the assignment with the USPTO to secure priority against later purchasers. More commonly, owners license the patent, granting permission—exclusively or non-exclusively, in defined fields of use, territories, and on negotiated royalty terms—to practice the invention. At bottom, a patent license is simply the patentee's enforceable promise not to sue the licensee for the licensed activity. Structuring that promise well, valuing the patent, and negotiating royalties are arts unto themselves; our guide how to license your patent: from valuation to term sheet walks through the deal mechanics, and in standards-heavy industries the analysis bends toward standard-essential patents and FRAND licensing.

The Supreme Court has policed the outer edges of licensing. In Kimble v. Marvel Entertainment, LLC, 576 U.S. 446 (2015)—the Spider-Man web-shooter case—the Court reaffirmed the old rule that a patentee may not collect royalties for use of an invention after the patent expires, even though several Justices doubted the rule's economic wisdom; stare decisis carried the day, and the lesson is to structure post-expiration payments carefully (for instance, as deferred payment for pre-expiration use). A patent may also serve as collateral for financing, and patents routinely feature in the IP toolkits of startups raising capital, a topic we connect in popular legal documents for startups.

International Patent Protection

Because patent rights are territorial, an inventor who wants protection abroad must obtain patents under the laws of each country or region—there is no such thing as a "world patent." Several treaties soften the logistics and cost. The Paris Convention for the Protection of Industrial Property guarantees national treatment and, crucially, a right of priority: an applicant who first files in one member country may file in other member countries within twelve months (for patents) and claim the original filing date, so that disclosures occurring in the interim do not defeat the later applications. The Patent Cooperation Treaty (PCT), with more than 150 contracting states, lets an applicant file a single international application that preserves the option to seek protection in many countries, provides an international search and (optionally) a preliminary opinion on patentability, and defers the substantial cost of "national stage" filings by up to thirty months from the priority date—buying time to assess commercial prospects. Regional systems such as the European Patent Office (EPO) let one application mature into protection across many member states.

Two practical warnings. First, recall that most countries have no grace period: a public disclosure or sale before filing, which a U.S. inventor might survive under the one-year grace period, can irrevocably destroy foreign rights. File first, disclose later. Second, U.S. law requires inventors of inventions made in the United States to obtain a foreign filing license before filing abroad in certain situations (§ 184)—a national-security screen ordinarily satisfied as a matter of course by filing first in the United States. Plan the international strategy early; foreign rights are easily and permanently lost through delay or premature disclosure.

A Worked Walkthrough: From Idea to Issued Patent

To tie the pieces together, follow a single hypothetical from conception to enforcement. (This scenario is illustrative only.)

An engineer, Dana, devises a genuinely clever battery-cooling manifold for electric vehicles. Step one: keep quiet and file. Before pitching investors or demonstrating a prototype, Dana files a thorough provisional application describing the manifold in enough detail to satisfy § 112—locking in a priority date and earning "Patent Pending" status. Because Dana hopes to sell in Europe and Asia, she resists the urge to demo publicly until the provisional is on file, preserving foreign rights that have no grace period. Step two: the nonprovisional. Within twelve months, Dana's patent attorney files a nonprovisional with carefully layered claims—broad independent claims for maximum scope, narrower dependent claims as fallbacks—claiming the provisional's benefit.

Step three: prosecution. Eighteen months in, the application publishes. The examiner's first office action rejects the broad claims as obvious under § 103, combining two prior-art references under KSR. Dana's attorney responds, amending the independent claims to add a specific feature absent from the prior art and submitting evidence of commercial success as a secondary consideration under Graham. The examiner allows the amended claims—but note the cost: that narrowing amendment may later create prosecution history estoppel under Festo, limiting Dana's reach under the doctrine of equivalents. Step four: issuance and marking. Dana pays the issue fee, the patent issues, and Dana marks her products (virtually, via a website) to preserve full damages under § 287.

Step five: enforcement. A competitor launches a suspiciously similar manifold. Dana sues in her state of incorporation (mindful of TC Heartland venue limits). The competitor counterattacks on two fronts: it files an IPR at the PTAB arguing obviousness over additional prior art (preponderance standard), and in district court it argues noninfringement and invalidity (clear-and-convincing standard). If Dana's patent survives, she may seek a reasonable royalty or lost profits, with damages apportioned to the value her manifold adds, and—if the copying was egregious—enhanced damages under Halo. Every doctrine in this guide shows up somewhere in that arc. The throughline is that early discipline (file before disclosing, draft claims carefully, mark diligently) pays off years later when the patent is tested.

Frequently Asked Questions

How long does a patent last, and does it ever renew? A utility or plant patent lasts twenty years from its earliest effective U.S. filing date; a design patent lasts fifteen years from grant (for designs filed on or after May 13, 2015). Patents do not renew. Utility patents do require maintenance fees at 3.5, 7.5, and 11.5 years to stay in force, but paying them merely keeps the existing term alive—nothing extends a patent past its statutory expiration except the narrow term-adjustment and regulatory-extension mechanisms (PTA and PTE).

Can I sell my invention or talk about it publicly before filing? You can, but it is risky and often costly. In the United States, a one-year grace period may forgive your own disclosure if you file within a year, but the on-sale and public-use bars are unforgiving in their details (recall Helsinn: even a secret commercial sale can start the clock). And most foreign countries have no grace period at all, so a single public disclosure can destroy your rights everywhere but the U.S. The safe rule: file at least a provisional before any public disclosure, sale, or offer for sale.

What is the difference between a provisional and a nonprovisional application? A provisional is an inexpensive, unexamined twelve-month placeholder that secures a priority date and "Patent Pending" status but never becomes a patent on its own. A nonprovisional is the real application the USPTO examines and may grant. To capture a provisional's early date, you must file a nonprovisional claiming its benefit within twelve months—and the provisional only secures priority for subject matter it actually describes.

Do I need a patent attorney, or can I file myself? You may file pro se, but it is rarely wise for anything you care about. The claims define your rights, and subtle drafting choices determine scope, validity, and enforceability for two decades. A registered patent attorney or agent (agents have the technical and patent-law qualifications but are not general lawyers) brings exactly the skill set that prevents expensive, often irreversible mistakes. The investment usually pays for itself in the strength of the resulting right.

My patent is granted. Why can't I sell my own product? Because a patent is a right to exclude others, not a right to practice your own invention free of everyone else's patents. Your product might still infringe a broader, earlier patent owned by someone else. Confirming you are clear to launch is a separate exercise—a freedom-to-operate analysis—distinct from obtaining your own patent.

Someone is copying my patented product. What do I do? Patent enforcement is private: the USPTO will not help. Consult a patent litigator promptly, both to evaluate infringement and validity and to mind the clock—damages reach back only six years (§ 286), and unmarked products may limit pre-notice recovery under § 287. Options range from a demand or licensing letter to suit in district court or, for imports, a Section 337 action at the ITC. Be aware that asserting a patent often invites a counterattack at the PTAB.

Can an AI be an inventor? No. The Federal Circuit held in Thaler v. Vidal that an inventor under the Patent Act must be a natural person. Inventions developed with AI assistance can still be patented, but a human must have made a significant contribution to the claimed invention; see AI-generated inventions.

Patent Cases by Topic

The following tables collect leading patent decisions by topic, with citations, as a reference map to the case law. The one-line summaries are general; consult the opinions themselves for holdings and reasoning, and remember that this body of law keeps moving.

Subject-Matter Eligibility (35 U.S.C. § 101)

Novelty, Anticipation, and Statutory Bars (35 U.S.C. § 102)

Nonobviousness (35 U.S.C. § 103)

Disclosure: Written Description, Enablement, Definiteness (35 U.S.C. § 112)

Claim Construction and Equivalents

Inventorship

Infringement: Direct, Equivalents, Indirect, Divided

Defenses: Exhaustion, Inequitable Conduct, Assignor Estoppel, Laches

Remedies: Injunctions, Damages, Willfulness, Fees

Declaratory Judgment, Jurisdiction, and Venue

PTAB and USPTO Practice

Practical Guidance and Takeaways

For inventors, a handful of disciplines separate durable patents from cautionary tales. Move fast: the United States awards priority to the first inventor to file, so consider a thorough provisional to lock in an early date and "Patent Pending" status while you develop. Guard your own mouth: public uses, sales, demos, and posts can become prior art, and although a one-year grace period exists for your own U.S. disclosures, the safest course—especially if you have any international ambition—is to file before you disclose, because most of the world grants no grace period at all. Understand what you are buying: a patent lets you exclude others, not practice your own invention free of others' rights, and you must enforce it in court. And hire the right help: the claims define your rights for twenty years, and a registered patent attorney or agent is the cheapest insurance you will ever buy against an irreversible drafting error.

For practitioners and businesses, the doctrine repays attention to the requirements that decide the most cases: subject-matter eligibility under the Alice/Mayo framework (still the wild card for software and diagnostics); obviousness under Graham and KSR (the most common invalidity ground); the § 112 disclosure requirements as sharpened by Amgen and Nautilus (a graveyard for over-broad functional claims); and the contours of infringement and remedies shaped by eBay, Halo, Octane, and the exhaustion and venue cases. The rise of PTAB proceedings—IPR and PGR, with their lower burden of proof and faster timelines—has made post-grant validity challenges a routine parallel front to district-court litigation, and any enforcement strategy must anticipate them. The international dimension—territoriality, Paris priority, the PCT, and the no-grace-period reality abroad—must be planned at the very start, before any disclosure forecloses options.

For everyone, return to where we began: the patent bargain. A limited, powerful right to exclude, granted in exchange for a full public disclosure that ultimately enriches the public domain. The system rewards those who invent something genuinely new and nonobvious, disclose it fully, file promptly, claim it carefully, and enforce it diligently. Master the framework—what can be patented, how patents are obtained and maintained, and how they are enforced and challenged—and an idea becomes a durable, defensible asset. Misunderstand it, and the rights slip away in the gaps: a premature demo, a thin provisional, an unpaid maintenance fee, an unmarked product. The difference between those outcomes is, more often than not, simply knowing the rules in advance.

Related Articles

• Patent Basics: A Plain-English Guide — the same fundamentals at a gentler pace.
• General Information Concerning Patents — a companion reference overview.
• Utility Patent Basics — a focused look at the most common patent type.
• Copyright vs. Trademark vs. Patent vs. Trade Secret — how patents fit among the four branches of IP.
• Patent Eligibility After Alice — the § 101 framework for software and business methods.
• Overcoming Obviousness Rejections: A Guide to Section 103 — beating the most-litigated invalidity ground.
• Responding to Patent Office Actions — prosecution strategy at the USPTO.
• What Constitutes Patent Infringement: Claims and Defenses — enforcement in federal court and at the ITC.
• Conducting Freedom-to-Operate Analysis for New Products — clearing the path before launch.
• How to License Your Patent: From Valuation to Term Sheet — turning the right to exclude into revenue.
• Understanding Patent Marking Requirements — preserving your damages.

This article is provided for general informational purposes and does not constitute legal advice. Patent law is detailed, technical, and continually developed by the courts and the USPTO; statutes, rules, fees, and case law change over time. Consult a registered patent attorney or agent about any specific invention, application, or dispute.