The Supreme Court Recalibrates Contributory Copyright Liability in Cox Communications, Inc. v. Sony Music Entertainment

As the justices prepare to take the bench for the new term, a review of last term’s business decisions shows Cox to be among the most important. On March 25, 2026, the Supreme Court issued its decision in Cox Communications, Inc. v. Sony Music Entertainment, reversing a judgment of the United States Court of Appeals for the Fourth Circuit that had held an internet service provider contributorily liable for copyright infringement committed by its subscribers. Cox Commc'ns, Inc. v. Sony Music Ent., 146 S. Ct. 959 (2026) (“Opinion”). The Court held that a provider of a service is contributorily liable for a user’s infringement only if the provider intended its service to be used for infringement, and that the requisite intent may be established in only two ways: by showing that the provider induced the infringement, or by showing that the provided service was tailored to infringement. Because Cox did neither, the Court concluded that it was not contributorily liable. Id. at 964.

The decision resolves a question the Court had not previously addressed—whether an internet service provider’s continued provision of service to subscribers it knows to be associated with infringement is, without more, sufficient to support contributory liability. The result was unanimous as to the judgment but divided as to the reasoning: seven Justices joined the opinion of the Court, and two Justices concurred only in the judgment while objecting that the majority had defined the universe of secondary liability too narrowly. This article summarizes the decision, including the concurrence, and then considers its implications for service providers generally—both for the design of acceptable-use policies and for the litigation that is likely to follow over where the line between lawful, general-purpose service and culpable facilitation should be drawn.

Procedural and Factual Background

Cox Communications is an internet service provider serving approximately six million subscribers, each associated with a unique Internet Protocol address. Id. at 965. As the Court observed, a single IP address may be shared by many individual users—a household, a coffee shop, or a college dormitory ordinarily has one address but multiple users—and a provider can identify the subscriber account associated with an address but cannot, as a general matter, distinguish among the individual users behind it or directly control how the service is used. Id.

Sony Music Entertainment and other major music copyright owners retained a monitoring firm, MarkMonitor, to detect infringement across the internet and trace it to particular IP addresses. Over the roughly two-year period at issue, MarkMonitor sent Cox 163,148 notices identifying subscriber IP addresses associated with alleged infringement. Id. Cox’s subscriber agreement prohibited the use of its service to post, copy, transmit, or disseminate infringing content, and Cox maintained a graduated response system: a warning after a second notice for an account, followed by suspensions and, after a defined number of notices, termination of service. Id. at 965-66.

The parties characterized Cox’s enforcement record differently. Sony emphasized that Cox terminated only a small number of subscribers for alleged infringement during the claim period while terminating far larger numbers for nonpayment and pointed to internal statements reflecting reluctance to act on notices. Id. at 966. Cox emphasized that its warning-and-suspension system curtailed the substantial majority of identified infringement. Id. Sony sued in the Eastern District of Virginia on theories of both contributory and vicarious liability. Id. A jury found for Sony on both theories, found Cox’s infringement willful, and awarded one billion dollars in statutory damages. Id.

The Fourth Circuit affirmed as to contributory liability, reasoning that supplying a product with knowledge that the recipient will use it to infringe is the sort of culpable conduct sufficient for contributory infringement. Id. It reversed as to vicarious liability, concluding that Cox did not receive a direct financial benefit traceable to the infringement, and remanded for a redetermination of damages. Id. The Supreme Court granted certiorari on the contributory-liability question and denied Sony’s cross-petition on vicarious liability. Id. at 967. Notably, the DMCA safe harbor was not available to Cox as a defense, because an earlier decision had foreclosed it for the relevant period. Sony Music Ent. v. Cox Commc'ns, Inc., 93 F.4th 222, 228 (4th Cir. 2024) (citing BMG Rights Mgmt. (US) LLC v. Cox Communications, Inc., 881 F.3d 293, 301-305 (4th Cir. 2018)).

The Opinion of the Court

Justice Thomas, writing for a seven-Justice majority, began from the premise that the Copyright Act does not expressly impose liability on one party for infringement committed by another, and that the Court has historically been reluctant to expand secondary liability beyond the forms its precedents have recognized. Opinion at 967. Ordinarily, the Court reasoned, when Congress intends to impose secondary liability, it does so expressly; absent such language, the Court looks to the specific forms of secondary liability that predated the Copyright Act and declines to enlarge them. Id.

Within that framework, the Court identified two—and only two—paths to the intent required for contributory liability. The first is inducement: a provider induces infringement when it actively encourages infringement through specific acts, as the file-sharing companies did in Metro-Goldwyn-Mayer Studios Inc. v. Grokster, Ltd., 545 U.S. 913 (2005), where the defendants promoted and marketed their software as a tool for infringement and built their business around that use. Id. The second is tailoring: a service is tailored to infringement when it is not capable of substantial or commercially significant noninfringing uses. The Court drew this second path from Sony Corp. of America v. Universal City Studios, Inc., 464 U.S. 417 (1984) (i.e., the Betamax case), in which the sale of a video tape recorder capable of substantial noninfringing uses—such as recording programs for later personal viewing—did not constitute contributory infringement, even though the device could also be used to infringe. Id. at 967-68. The Court noted that these two paths track the structure of patent law’s active-inducement and contributory-infringement provisions. Id. at 968.

Central to the majority’s reasoning was its repeated observation that mere knowledge that a service will be used to infringe does not establish the intent required for contributory liability. Id. The Court traced that principle through its earlier decisions, emphasizing that liability does not attach simply because a seller supplies equipment with constructive knowledge that customers may use it to infringe, and that a failure to take affirmative steps to prevent infringement is not, by itself, a basis for contributory liability. Id.

Applying these principles, the Court concluded that Cox neither induced infringement nor provided a service tailored to it. As to inducement, the Court found no evidence that Cox encouraged its subscribers to infringe; to the contrary, Cox discouraged infringement through its subscriber agreement and its system of warnings, suspensions, and terminations, and Sony offered no evidence of promotion or marketing directed at infringing use. Id. As to tailoring, the Court found that Cox’s provision of internet access is plainly capable of substantial noninfringing uses and is used for a wide range of lawful purposes; Cox simply provided general internet access and did not design its service to facilitate infringement. Id. The Fourth Circuit’s contrary rule—that supplying a product with knowledge of infringing use suffices—went beyond the two recognized forms of liability and conflicted with the Court’s admonition that contributory liability cannot rest on knowledge plus inaction alone. Id. at 968-69.

Finally, the Court rejected Sony’s argument that the DMCA safe harbor presupposes that providers serving known infringers can be held liable. The safe harbor, codified at Section 512, conditions certain protections on a provider’s adoption of a policy for terminating repeat infringers in appropriate circumstances. Id. at 969. The Court reasoned that the DMCA does not itself impose liability on providers who serve known infringers; it creates defenses for qualifying providers. Id. The Court further noted that the statute expressly provides that a failure to qualify for the safe harbor does not bear adversely on a provider’s separate defense that its conduct is not infringing. Id. The safe harbor, in the Court’s view, operates as a shield and not as a source of liability.

The Concurrence in the Judgment

Justice Sotomayor, joined by Justice Jackson, concurred in the judgment but disagreed with the majority’s reasoning. She agreed that neither inducement nor tailoring covered Cox’s conduct, but she objected to the majority’s holding that those are the only two routes to contributory liability. In her view, the Court’s prior decisions did not foreclose other common-law theories of secondary liability—most notably aiding and abetting—and the majority’s narrowing of the doctrine was supported by neither precedent nor statute. Id. She read the Court’s earlier cases as defining the scope of secondary copyright liability by reference to common-law principles applied throughout the law, rather than confining it to the two specific circumstances those cases happened to address. Id. at 970-71.

The concurrence further argued that the majority’s rule disturbs the incentive structure Congress established in the DMCA. Id. at 971-72. On this view, the safe harbor was designed to encourage providers to adopt reasonable anti-infringement policies in exchange for protection from liability; if providers face no realistic prospect of secondary liability regardless of what they know or do, the inducement to maintain such policies is diminished.

Despite these objections, Justice Sotomayor concurred in the judgment because, in her view, the record could not establish the intent that an aiding-and-abetting theory would require. Drawing on the Court’s recent decisions in Twitter, Inc. v. Taamneh, 598 U.S. 471 (2023) and Smith & Wesson Brands, Inc. v. Estados Unidos Mexicanos, 605 U.S. 280 (2025), she explained that aiding-and-abetting liability requires conscious and culpable participation in a wrongful act with the intent to help it succeed, and that such intent ordinarily depends on knowledge of specific instances of wrongdoing. Id. at 973-74. Cox, she reasoned, knew only which connections were associated with infringement, not which individuals were responsible; for connections serving households, dormitories, hospitals, or downstream regional providers, Cox could not identify the actual infringers and lacked control over the intermediaries who redirected its service to end users. Id. at 974-75. On that record, she concluded, Cox’s conduct amounted at most to indifference, which is insufficient for aiding-and-abetting liability. Id.

Implications for Service Providers

The holding is narrow in two respects that bear emphasis. First, it addresses only contributory liability, leaving vicarious liability and direct liability for providers who themselves copy, host, or distribute infringing material undisturbed. Second, the Court’s finding of no contributory liability rests on the particular record before the Court. The decision establishes that Cox’s provision of generalized internet access to known infringers did not, in the context of its overall conduct, establish the intent required for contributory liability; it does not establish that knowing provision of internet service to repeat infringers could never, on different facts, support an inference of intent. With that scope in mind, the decision offers reasonably clear guidance for how providers can structure their conduct to reduce contributory-liability exposure. Because liability turns on intent—demonstrated through inducement or tailoring—the decision rewards conduct that affirmatively negates an inference of intent and penalizes conduct that supplies one.

Acceptable-Use Policies and Anti-Infringement Measures

The Court treated Cox’s contractual prohibition on infringing use, together with its graduated system of warnings, suspensions, and terminations, as evidence that Cox discouraged rather than encouraged infringement. For service providers generally, this suggests that several categories of documentation and practice are useful in establishing the absence of infringing intent. A clear acceptable-use policy that prohibits infringing conduct, incorporated into the terms of service, supplies contemporaneous evidence that the provider does not intend its service to be used for infringement. A functioning notice-handling process—one that receives, records, and acts upon infringement notices—demonstrates responsiveness rather than indifference. Graduated enforcement measures, including warnings and the suspension or termination of service in appropriate circumstances, show corrective effort even where they do not eliminate all infringement.

Two cautions temper this guidance. First, the decision does not hold that any particular enforcement threshold is adequate or that any particular failure to terminate is safe; it holds that Cox’s conduct, considered as a whole, did not establish intent. A provider whose marketing, product design, internal communications, or enforcement choices suggest an affirmative desire to facilitate infringement, rather than mere reluctance or imperfect execution, may present a materially different case. Second, the value of these measures is evidentiary rather than absolute. They help rebut an inference of intent; they do not create a categorical immunity.

Providers should also bear in mind that the DMCA safe harbor remains a separate and independently valuable protection: the reasoning of this decision concerns when liability attaches in the first instance, not whether a provider has qualified for the statutory safe harbor, and the conditions of that safe harbor—including the reasonable implementation of a repeat-infringer policy—continue to apply on their own terms.

Future Litigation: Where Is the Line?

Although the decision provides a workable framework, it leaves substantial room for dispute about its boundaries, and that uncertainty is where future litigation is likely to concentrate. The decision’s logic extends naturally to other providers of general-purpose, broadly lawful functionality—cloud computing platforms, content-sharing and hosting services, and similar intermediaries—because each supplies a service capable of substantial noninfringing uses and ordinarily lacks the kind of specific, individualized knowledge that the Court found absent in Cox’s case. To that extent, the decision is favorable to such providers.

The harder questions concern the two recognized paths themselves. On inducement, future plaintiffs will likely argue that a provider’s documentation, marketing, product design, or internal communications evidence an intent to encourage infringing use, while providers will respond that promoting general capabilities, or merely tolerating known misuse, is not the active encouragement that Grokster requires. The line between marketing a lawful capability and inducing an unlawful one will be contested on specific facts, and providers can expect that their public messaging and internal records will be scrutinized for statements that could be read as endorsing or facilitating infringement. On tailoring, plaintiffs may argue that a particular service—or a particular configuration or feature of it—is not capable of substantial noninfringing uses, an inquiry that becomes more fact-intensive as services become more specialized.

These questions arise with particular salience for emerging technologies, including artificial intelligence systems, which are general-purpose tools that can be used for both lawful and infringing ends and whose training data and outputs are already the subject of separate litigation. The decision says nothing about artificial intelligence directly, and any extension of its reasoning to that context remains untested; it is best understood as one example, alongside cloud and content-sharing platforms, of a general-purpose technology to which the framework may eventually be applied rather than as a question the Court has resolved. What the decision does establish across all of these settings is that generalized knowledge of misuse, without more, is not enough; what will remain disputed is the kind and degree of additional evidence—inducing conduct, tailoring of the service, or documentation suggesting intent—that crosses the line.

A further source of future litigation lies in the doctrinal disagreement between the majority and the concurrence. The majority’s position that inducement and tailoring are the exclusive routes to contributory liability, if read broadly, may foreclose common-law theories such as aiding and abetting; the concurrence’s view would preserve them. Litigants seeking to expand or contract secondary liability will press these competing readings, and lower courts will have to determine how far the majority’s framing forecloses alternative theories.

Drawing Parallels Between Cox and Hikma

In laying out the standard for secondary liability, the Court noted that patent law has a “historic kinship” with copyright law and suggested that the two types of contributory copyright infringement “track patent law,” specifically infringement under 35 U.S.C. § 271(b) and 35 U.S.C. § 271(c). Then, three months later, the Court issued an opinion on the pleading standard for § 271(b) patent infringement, which repeatedly cited to Cox and other cases about copyright infringement. Hikma Pharm. USA Inc. v. Amarin Pharma, Inc., 146 S. Ct. 1391, 1397, 1399-400 (2026). The Court appears to see alignment in the secondary liability standards for patent infringement and copyright infringement.

Given that, it is worth considering whether Hikma’s statements on the pleading standard for induced patent infringement should also be read to apply to pleading induced copyright infringement. We discuss Hikma at length in our Noted With Interest article, including a discussion of its impact on the pleading standard, and what that means practically for potential litigants.

Conclusion

Cox establishes that a service provider’s contributory liability for copyright infringement turns on intent, demonstrated through inducement or tailoring, and not on a generalized awareness that the service will be used by some to infringe. The decision provides service providers with meaningful guidance on conduct that tends to negate infringing intent, while leaving open—by its own terms and through the disagreement between the majority and the concurrence—significant questions about its application to specific services and to emerging technologies. As courts apply the framework to new facts, the practical task for providers and copyright owners alike will be to understand not only what the decision holds, but where its boundaries lie.

//////
Adam Burstain, Associate
Nicola Felice, Partner
nicolafelice@quinnemanuel.com