UMG v. Suno
Last reviewed
May 19, 2026
Sources
No citations yet
Review status
Needs citations
Revision
v1 · 3,697 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
May 19, 2026
Sources
No citations yet
Review status
Needs citations
Revision
v1 · 3,697 words
Add missing citations, update stale details, or suggest a clearer explanation.
UMG Recordings, Inc., et al. v. Suno, Inc. is a landmark copyright infringement lawsuit filed on June 24, 2024 by a coalition of major record labels, coordinated by the Recording Industry Association of America (RIAA), against the generative AI music startup Suno in the U.S. District Court for the District of Massachusetts. The case, docketed 1:24-cv-11611, accuses Suno of copying millions of copyrighted sound recordings without authorization to train its AI music generation models. A companion case, UMG Recordings, Inc. v. Uncharted Labs, Inc., was filed against the rival AI music service Udio on the same day in the U.S. District Court for the Southern District of New York.[^1][^2]
Together, the two suits represented the first major copyright actions brought by the recorded music industry against AI music generation companies, and they have been described in legal commentary as the most consequential test cases yet for how United States copyright law applies to AI training on sound recordings.[^3]
Suno, Inc. is a generative AI company headquartered in Cambridge, Massachusetts. It was founded in 2022 by Mikey Shulman, Georg Kucsko, Martin Camacho, and Keenan Freyberg, four former colleagues from the financial-analytics startup Kensho Technologies. The company developed a text-to-music platform that allows users to generate complete songs, including vocals and instrumentation, from natural-language prompts.[^4]
By the time the lawsuit was filed in June 2024, Suno had become one of the most prominent AI music services. The company had raised a $125 million Series B financing round in May 2024 led by Lightspeed Venture Partners, with participation from Nat Friedman, Daniel Gross, Matrix, and Founder Collective.[^5] Subsequent rounds brought the company to a $2.45 billion post-money valuation in a Series C announced on November 19, 2025, led by Menlo Ventures with participation from NVentures (NVIDIA's venture arm), Hallwood Media, Lightspeed, and Matrix.[^6] By early 2026, Suno reported having two million paid subscribers and roughly $300 million in annual recurring revenue.[^7]
In a March 2024 Rolling Stone feature on Suno, Matrix Partners venture capitalist Antonio Rodriguez, an early backer of the company, was quoted candidly on the company's licensing position. Rodriguez told the magazine: "Honestly, if we had deals with labels when this company got started, I probably wouldn't have invested in it. I think that they needed to make this product without the constraints." Rodriguez also acknowledged that the firm understood Suno could face lawsuits from record labels and music publishers, characterizing this as "the risk we had to underwrite when we invested in the company."[^8]
The quotation was widely cited by music-industry observers as evidence that Suno's investors knew the company had not obtained licenses for its training data, and the RIAA's complaints later quoted publicly available statements describing Suno's training process.[^1]
The plaintiffs in the action are three of the world's largest recorded-music groups, working in coordination through the RIAA. They are Universal Music Group's UMG Recordings and Capitol Records subsidiaries, Sony Music Entertainment, and Warner Music Group's Atlantic and Warner Records labels along with several affiliates. Together these groups control the rights to a large fraction of the recorded music released in the United States since the early twentieth century.[^9]
The complaint was filed on June 24, 2024 in the U.S. District Court for the District of Massachusetts and assigned to Chief Judge F. Dennis Saylor IV.[^9][^10] The action is captioned UMG Recordings, Inc., Capitol Records, LLC, Sony Music Entertainment, Atlantic Recording Corporation, Atlantic Records Group LLC, Rhino Entertainment LLC, The All Blacks U.S.A., Inc., Warner Music International Services Limited, and Warner Records Inc. v. Suno, Inc. (Case No. 1:24-cv-11611).[^9][^10]
The nine named plaintiffs are:
The complaint alleges that Suno engaged in "mass-scale" copyright infringement by copying "decades worth of the world's most popular sound recordings" into its training corpus without obtaining licenses from the rightsholders. The plaintiffs contend that this unauthorized copying constituted infringement at the input stage (in copying the recordings to assemble training data) and at the output stage (in generating songs that compete commercially with the originals).[^1][^11]
RIAA Chief Legal Officer Ken Doroshow described the cases at the time of filing as "straightforward cases of copyright infringement involving unlicensed copying of sound recordings on a massive scale." RIAA Chairman and CEO Mitch Glazier added that "unlicensed services like Suno and Udio that claim it's 'fair' to copy an artist's life's work and exploit it without consent or pay" undermined the prospects of "responsible, consumer-friendly AI in which we can all thrive."[^1][^11]
The complaint asks the court for three forms of relief:
A significant portion of the complaint is devoted to documenting examples in which Suno's service, when given prompts and lyrics from specific copyrighted songs, generated outputs that allegedly mimicked the distinctive features of the original recordings. The complaint enumerates dozens of such examples.
The works cited as the basis for these claims include Chuck Berry's "Johnny B. Goode," Bill Haley & His Comets' "Rock Around the Clock," B.B. King's "The Thrill Is Gone," The Temptations' "My Girl," The Beach Boys' "I Get Around," ABBA's "Dancing Queen," Mariah Carey's "All I Want for Christmas Is You," and Green Day's "American Idiot," among others.[^11][^12]
For Chuck Berry's "Johnny B. Goode" in particular, the complaint states that Suno's service generated "29 different outputs" reflecting the style of the original recording when prompted with the song's lyrics combined with descriptive cues such as "1950s rock and roll, rhythm & blues, 12 bar blues, rockabilly, energetic male vocalist, singer guitarist." One output, titled in the complaint as "Deep Down in Louisiana Close to New Orle," allegedly replicated "the highly distinctive rhythm of the original's chorus" and used "the same melodic shape on the phrases 'go Johnny, go, go.'"[^12][^13]
The plaintiffs presented side-by-side musical transcriptions to illustrate similarities between original recordings and AI-generated outputs, and they argued that these output examples were probative evidence that the original recordings had been used as training data.[^11][^12]
On August 1, 2024, Suno filed its formal answer to the complaint. The answer admitted, in a striking concession, that Suno's training data included copyrighted recordings. The filing stated that "it is no secret that the tens of millions of recordings that Suno's model was trained on presumably included recordings whose rights are owned by the Plaintiffs."[^14][^15]
Suno raised fair use as a primary affirmative defense, asserting that the use of copyrighted works to train a generative model is transformative and protected by Section 107 of the Copyright Act. The company analogized AI training to the digitization of books at issue in Authors Guild v. Google, in which courts found Google's mass scanning of books to build a searchable index to be fair use.[^14][^15]
In response to the "Johnny B. Goode" allegations, Suno argued that the existence of outputs similar to a famous recording does not prove that the specific copyrighted sound recording was in the training data. Suno's counsel emphasized that the compositional copyright in "Johnny B. Goode" was owned by music publishers (BMG Rights Management and Music Services) that were not parties to the suit, and that "no one owns musical styles." Suno contended that copyright protections apply to specific recordings and compositions, not to the broader genre or style they exemplify.[^13][^16]
The same day that Suno filed its answer, CEO and co-founder Mikey Shulman published a blog post defending the company's training practices. Shulman wrote: "We train our models on medium- and high-quality music we can find on the open internet… Much of the open internet indeed contains copyrighted materials, and some of it is owned by major record labels." He drew an analogy to human musical learning, comparing the training process to "a kid writing their own rock songs after listening to the genre," and concluded: "Learning is not infringing. It never has been, and it is not now."[^14][^17]
The RIAA responded to the filing the same day, calling Suno's admission "a major concession of facts they spent months trying to hide and acknowledged only when forced by a lawsuit," and rejecting the fair-use framing as "industrial-scale infringement" that "does not qualify as 'fair use.'"[^14][^17]
After the August 2024 answer, the case moved into discovery. By early 2025 the parties were litigating questions about the form in which the plaintiffs would produce the copyright registration deposit copies for the thousands of recordings at issue. In April 2025, the parties filed a joint stipulation allowing UMG and the other plaintiffs to provide digital deposit copies rather than physical media, with provision for certified deposit copies from the U.S. Copyright Office on request.[^18]
On September 19, 2025, UMG and the other plaintiffs moved to amend their complaint to add new claims under the anti-circumvention provisions of the Digital Millennium Copyright Act (DMCA). The proposed amended complaint alleged that Suno had obtained "many (if not all)" of the copyrighted recordings in its training data by "stream-ripping" them from YouTube, that is, by circumventing YouTube's "rolling cipher" encryption to download audio from the platform.[^19][^20]
The proposed amended complaint sought up to $150,000 per work in statutory damages for direct copyright infringement and up to $2,500 for each individual act of circumvention of a technological protection measure under 17 U.S.C. § 1201. Commentators noted that the strategy was inspired by Anthropic's $1.5 billion class-action settlement in Bartz v. Anthropic, which had focused attention on how AI developers acquired their training data.[^19][^20]
Suno opposed the motion, arguing in part that YouTube's rolling cipher is a "copy control" rather than an "access control" within the meaning of the DMCA, and therefore that circumventing it does not violate Section 1201. The plaintiffs replied that the cipher functions as an access control because it conditions the user's ability to obtain a decoded audio stream on the use of YouTube's player. Briefing on the motion to amend, including a sur-reply by Suno permitted by the court, continued through late 2025.[^19][^21]
Trial-court scheduling has shifted multiple times. The summary-judgment deadline was at one point set in mid-2026 but was later pushed back. As of a March 2026 docket update, the deadline for summary-judgment motions in UMG Recordings v. Suno was reset to January 8, 2027, with the case falling behind several other AI copyright cases in the queue for fair-use rulings.[^22]
On the same day that UMG and its co-plaintiffs sued Suno, the RIAA coordinated a parallel suit against the rival AI music service Udio, operated by Uncharted Labs, Inc. That action was filed in the U.S. District Court for the Southern District of New York and made substantially similar allegations.[^1][^11]
The Udio case followed a different trajectory. On October 29-30, 2025, Universal Music Group and Udio announced that they had reached a settlement of UMG's claims and signed a strategic partnership. Under the deal, Udio agreed to launch a new platform in 2026 powered by AI models trained on authorized and licensed music. Udio's existing service would remain available to users during a transition period within a "walled garden," with additional fingerprinting and filtering controls. The settlement amount and detailed terms were not publicly disclosed.[^23][^24]
Sony Music and Warner Music did not settle with Udio at that time and continued to litigate separately. UMG's separate copyright case against Suno also remained active despite the Udio settlement.[^23][^24]
Approximately a month after the Udio settlement, on November 25, 2025, Warner Music Group announced that it had settled its claims against Suno and entered a strategic partnership of its own. As part of the agreement, Suno acquired Warner's live-music platform Songkick. The deal also contemplated a transition to licensed AI music models on Suno's platform, expanded artist and rightsholder compensation, and increased restrictions on free-tier downloads.[^25][^26]
Despite the Warner settlement, UMG and Sony continued to litigate against Suno on substantially the same theories raised in the original 2024 complaint. Reporting in late 2025 indicated that Suno sought to keep the Warner settlement terms confidential, including from the remaining plaintiffs.[^27]
The central legal question presented by the lawsuit is whether training a generative AI model on copyrighted sound recordings, without a license, qualifies as fair use under Section 107 of the U.S. Copyright Act. The four statutory fair-use factors are (1) the purpose and character of the use, (2) the nature of the copyrighted work, (3) the amount and substantiality of the portion used, and (4) the effect of the use on the potential market for or value of the copyrighted work.[^15]
Suno argues that the training process is "exceedingly transformative," producing a model whose outputs are new musical works rather than mere reproductions of any single training input. The plaintiffs argue that the use is essentially commercial, that the recordings are copied in full, and that the resulting outputs compete directly in the market for recorded music. They reject Suno's analogy to Authors Guild v. Google, contending that the comparison ignores that Google's index neither functioned as a substitute for the underlying books nor generated derivative works.[^11][^15]
A significant external development in the case is the June 23, 2025 ruling by Senior District Judge William Alsup of the Northern District of California in Bartz v. Anthropic. In that case, Judge Alsup held that Anthropic's use of books for AI training was "exceedingly transformative" and constituted fair use, while separately holding that the company's downloading of pirated copies of books from "shadow libraries" for its permanent internal library was not fair use. The case settled in August 2025 with class counsel describing the outcome as "historic" and beneficial to authors.[^28][^29]
Suno has invoked the Bartz fair-use analysis in its defense. The plaintiffs, by adding stream-ripping allegations to their proposed amended complaint, have echoed the Bartz distinction between lawful training methods and unlawful acquisition of training material.[^19][^28]
Under 17 U.S.C. § 504(c), the plaintiffs may elect statutory damages of up to $150,000 per work for willful infringement. Given that the plaintiffs collectively assert rights in many thousands of recordings allegedly used to train Suno, the potential statutory exposure is substantial. The proposed addition of DMCA Section 1201 claims, which permit statutory damages of up to $2,500 per act of circumvention, could further multiply potential liability if the court permits the amendment and the plaintiffs prevail on the merits.[^19][^20]
As of May 2026, UMG Recordings v. Suno remains pending in the U.S. District Court for the District of Massachusetts before Chief Judge F. Dennis Saylor IV. The case has narrowed but remains broad in legal stakes:
The fair-use motion before Judge Saylor is widely viewed as one of the most consequential pending rulings in U.S. AI copyright litigation. A ruling in favor of Suno on fair use would significantly weaken the leverage of rightsholders in negotiating licenses with AI music developers, while a ruling against Suno would reinforce the licensing template established by the UMG-Udio and Warner-Suno settlements and the parallel Anthropic class action.[^28][^30]
The lawsuit, together with the related Udio action, has reshaped the commercial and legal landscape of AI music generation.
Licensing as the emerging norm. Both the UMG-Udio and Warner-Suno settlements established that future commercial AI music services from those companies will rely on licensed training data and compensate rightsholders. The Copyright Alliance characterized the UMG-Udio deal as evidence that "generative AI products should and can be crafted responsibly and with respect for copyright owners."[^24]
Continuing legal uncertainty. With UMG and Sony still litigating against Suno, the fair-use question remains unresolved at the appellate level for AI music training. A district-court summary-judgment ruling in UMG v. Suno, if it issues, would be one of the first reasoned decisions specifically addressing fair use as applied to AI music generation on sound recordings, as distinct from text or images.[^22][^30]
Cross-influence with other AI copyright cases. The trajectory of Bartz v. Anthropic, including its $1.5 billion class settlement, has directly influenced both Suno's defenses (which invoke Bartz's transformative-use holding) and the plaintiffs' pleading strategy (which now mirrors Bartz's focus on how training data was acquired).[^19][^28]
Commercial repositioning of Suno. During the pendency of the litigation, Suno has continued to grow rapidly, releasing the Suno v5 model in September 2025 with substantially improved audio quality, and reaching a $2.45 billion valuation in November 2025. Industry coverage has framed Suno's strategic decisions, including the Warner settlement and Songkick acquisition, as moves to position the company within rather than outside the licensed music economy, even as it litigates the fair-use question with the remaining plaintiffs.[^6][^25][^31]
The eventual outcome of UMG Recordings v. Suno, whether by judgment or settlement, is expected to be a foundational precedent for how copyright law applies to AI training on sound recordings in the United States.