The U.S. Supreme Court Reverses a $1 Billion Copyright Verdict and, Arguably, Renders the DMCA Safe Harbor Obsolete

On March 25, 2026, the Supreme Court unanimously reversed a $1 billion copyright infringement verdict against Cox Communications, one of the nation’s largest internet service providers (“ISPs”). Cox Communications, Inc. v. Sony Music Entertainment, 607 U.S. ___, No. 24–171 (2026). The decision is the Court’s first major ruling on secondary copyright liability since Metro-Goldwyn-Mayer Studios Inc. v. Grokster, Ltd., 545 U.S. 913 (2005), more than twenty years ago.

Justice Thomas’s majority opinion holds that an ISP cannot be contributorily liable for its subscribers’ copyright infringement merely because it continues to provide internet service with knowledge that some subscribers will use it to infringe. The required intent can be established only in two ways: by showing that the defendant induced the infringement or that it provided a service tailored to infringement. Cox did neither.

The result was 9–0, but the reasoning was not. Justice Sotomayor, joined by Justice Jackson, concurred only in the judgment. Her concurrence argues that the majority unnecessarily forecloses other common-law theories of secondary liability and, in doing so, may have rendered the Digital Millennium Copyright Act’s (“DMCA”) safe harbor provision a dead letter.

The Facts

Cox Communications serves approximately six million subscribers. Each subscriber account is associated with a unique Internet Protocol (“IP”) address, but multiple users may share a single address; e.g., a household, a coffee shop, a university dormitory. Cox has limited visibility into who among those users is doing what. It knows which IP address corresponds to which account, but it cannot identify individual users or control how they use the service.

Sony Music Entertainment and other major music copyright owners retained MarkMonitor to track online infringement. Over a roughly two-year period, MarkMonitor sent Cox 163,148 notices identifying subscriber IP addresses associated with copyright infringement. Cox maintained a thirteen-strike escalation system: after initial notices, subscribers received warnings; after further notices, Cox suspended service temporarily; after thirteen notices, Cox terminated the account. During the relevant period, Cox terminated thirty-two subscribers for copyright infringement, while terminating hundreds of thousands for nonpayment. Slip op. at 4–5.

Cox also contractually prohibited subscribers from using their connections to “post, copy, transmit, or disseminate any content that infringes the patents, copyrights … or proprietary rights of any party.” Slip op. at 4.

Procedural History

Sony sued Cox in the Eastern District of Virginia, asserting two theories of secondary copyright liability: contributory infringement (Cox materially contributed to its subscribers’ infringement by continuing to serve known infringers) and vicarious infringement (Cox profited from and had the ability to supervise the infringing activity). A jury found for Sony on both theories, found Cox’s infringement willful, and awarded $1 billion in statutory damages. 464 F. Supp. 3d 795, 807–808 (E.D. Va. 2020).

The Fourth Circuit affirmed on contributory liability but reversed on vicarious liability, finding that Cox did not receive a direct financial benefit from its subscribers’ infringement. 93 F.4th 222 (4th Cir. 2024). The court vacated the damages award and remanded. On contributory liability, the Fourth Circuit applied its own circuit precedent, reasoning that “supplying a product with knowledge that the recipient will use it to infringe copyrights is exactly the sort of culpable conduct sufficient for contributory infringement.” Id. at 236 (citing BMG Rights Mgmt. (US) LLC v. Cox Communications, Inc., 881 F.3d 293, 308 (4th Cir. 2018)).

The Supreme Court granted Cox’s petition for certiorari on the contributory liability question. It denied Sony’s cross-petition on vicarious liability.

The Majority Opinion: Two Forms of Contributory Liability, and Only Two

Justice Thomas’s opinion begins from a structural premise: the Copyright Act “does not expressly render anyone liable for infringement committed by another.” Sony Corp. of America v. Universal City Studios, Inc., 464 U.S. 417, 434 (1984). The Court has recognized secondary liability only where it predated the statute, and it is “loath to expand such liability beyond [its] precedents.” Slip op. at 6–7.

Those precedents, Thomas writes, support exactly two forms of contributory copyright liability. The first is inducement: actively encouraging infringement through specific acts, as in Grokster, where file-sharing companies “promoted and marketed their software as a tool to infringe copyrights.” 545 U.S. at 926. The second is providing a service “tailored to infringement,” meaning the service is “not capable of ‘substantial’ or ‘commercially significant’ noninfringing uses.” Id. at 942 (Ginsburg, J., concurring). This is the Sony test: the Betamax could be used to record television programs for later personal viewing (a noninfringing use), so selling it to the general public did not constitute contributory infringement. 464 U.S. at 456.

The majority draws an explicit parallel to patent law, where 35 U.S.C. § 271(b) addresses inducement and § 271(c) addresses the sale of products with no substantial noninfringing use. The Court “repeatedly made clear” that mere knowledge a service will be used to infringe is insufficient. Slip op. at 8–9 (collecting Kalem Co. v. Harper Brothers, 222 U.S. 55 (1911); Sony; Grokster).

Applying these principles, the Court found that Cox neither induced infringement nor provided a service tailored to it. Cox did not promote, market, or encourage infringement; in fact, it “repeatedly discouraged copyright infringement by sending warnings, suspending services, and terminating accounts.” Slip op. at 9. And internet access is self-evidently capable of substantial noninfringing uses. “Cox simply provided Internet access, which is used for many purposes other than copyright infringement.” Id.

The Fourth Circuit’s rule—that supplying a service with knowledge of infringement is “sufficient for contributory infringement”—was a third theory of liability that the Court’s precedents do not support. The majority reversed.

The Sotomayor Concurrence

Justice Sotomayor’s concurring opinion, joined by Justice Jackson, agreed that Cox is not liable but rejected the majority’s reasoning on two grounds.

First, the concurrence argued that the majority artificially limits secondary liability to two forms when Grokster itself held that Sony “neither ‘displace[d] other theories of secondary liability’ nor ‘foreclose[d] rules of fault-based liability derived from the common law.’” Sotomayor concurrence at 4 (quoting Grokster, 545 U.S. at 934–935). The majority, in Justice Sotomayor’s opinion, “does not even mention” this language. Id. Instead, the majority invokes a general principle that Congress ordinarily imposes secondary liability expressly. But Sotomayor argued that principle is irrelevant because the Court held over 40 years ago in Sony that the Copyright Act impliedly provides for secondary liability. “Stare decisis requires this Court to apply that holding fairly, not ignore or artificially constrain it.” Id. at 5.

Second, Sotomayor’s concurrence argued that Cox should lose under common-law aiding-and-abetting principles, but only because the facts do not support the requisite intent. Drawing on Twitter, Inc. v. Taamneh, 598 U.S. 471 (2023), and Smith & Wesson Brands, Inc. v. Estados Unidos Mexicanos, 605 U.S. 280 (2025), the concurrence identified the “conceptual core” of aiding and abetting as conscious, culpable participation in wrongful conduct with the intent to help it succeed.

The problem for Sony, Sotomayor explained, is informational. When Cox receives a copyright-violation notice, it learns only that a particular IP address was associated with infringement, not who among the users of that connection committed it. A household may have multiple family members and neighbors with the Wi-Fi password. A regional ISP that purchases connectivity from Cox may serve thousands of end users. Without knowledge of who the actual infringers are, Cox cannot be found to have intended to aid specific instances of infringement. And Sony did not show the “pervasive, systemic, and culpable assistance” that Taamneh requires for a more generalized theory. At most, Sony showed Cox was “indifferent,” but indifference is not enough. Sotomayor concurrence at 11–13.

The DMCA Safe Harbor Problem

The most consequential tension between the majority and the concurrence concerns what the decision does to the DMCA.

In 1998, Congress enacted the DMCA safe harbor, codified at 17 U.S.C. § 512, which shields ISPs from secondary copyright liability if they, among other things, adopt and reasonably implement “a policy that provides for the termination in appropriate circumstances of subscribers and account holders” who “are repeated infringers.” § 512(i)(1)(A). The premise of that bargain is straightforward: ISPs face potential secondary liability for their subscribers’ infringement; in exchange for implementing repeat-infringer policies, they get a defense. The safe harbor thus creates an incentive structure: ISPs are motivated to police their networks because, without doing so, they face liability exposure.

The majority’s holding removes the predicate. If ISPs are not contributorily liable for continuing to serve known infringers (absent inducement or a service tailored to infringement), the safe harbor protects against a liability that, after Cox, largely does not exist. As Sotomayor wrote, “ISPs no longer face any realistic probability of secondary liability for copyright infringement, regardless of whether they take steps to address infringement on their networks and regardless of what they know about their users’ activity.” Sotomayor concurrence at 6. Cox’s own counsel conceded at oral argument that under the majority’s rule, the safe harbor provision will not “d[o] anything at all” going forward. Tr. of Oral Arg. 28–29.

Justice Thomas addressed this argument in two sentences. He noted that the DMCA does not expressly impose liability on ISPs who serve known infringers; it merely creates defenses. And § 512(l) provides that failure to qualify for the safe harbor “shall not bear adversely upon … a defense by the service provider that the service provider’s conduct is not infringing.” Slip op. at 10. That is technically correct as a matter of statutory text. But it sidesteps the structural argument: Congress enacted § 512 against the backdrop of the Court’s secondary liability precedents. If the underlying liability disappears, the incentive structure disappears with it.

Takeaway

Cox eliminates the most expansive theory of contributory copyright liability that had gained traction in the lower courts: that an ISP’s continued service to known infringers, standing alone, is sufficient. Content owners must now show either that the defendant actively induced infringement or that its service was specifically designed for it. For ISPs, the practical consequence is significant: arguably, they no longer face contributory liability merely for declining to disconnect subscribers flagged by automated infringement-detection systems.

But Sotomayor’s concurrence identifies a real problem. The majority’s rule may have eliminated the incentive for ISPs to maintain repeat-infringer policies at all. Before Cox, those policies were the price of the DMCA safe harbor. After Cox, the safe harbor protects against a form of liability that the Court has now held does not exist, at least not in the form Congress appears to have contemplated when it enacted the DMCA in 1998. Whether Congress responds legislatively remains to be seen.