Last reviewed
Jun 8, 2026
Sources
23 citations
Review status
Source-backed
Revision
v1 ยท 1,956 words
Add missing citations, update stale details, or suggest a clearer explanation.
Disney and Universal v. Midjourney is a copyright infringement lawsuit filed on June 11, 2025, in which The Walt Disney Company and Comcast's NBCUniversal jointly sued the generative AI image service Midjourney in the U.S. District Court for the Central District of California. It is widely described as the first major copyright suit brought by Hollywood studios against a generative AI company, distinguishing it from earlier AI image cases such as Getty Images v. Stability AI and Andersen v. Stability AI, which were brought by a stock-photo company and individual artists respectively.[1][2][3]
The 110-page complaint alleges that Midjourney both copied the studios' works to train its models and operates a service that, on demand, produces near-identical images of protected characters such as Darth Vader, Yoda, the Minions, Shrek, Homer Simpson and Elsa. The studios famously labeled Midjourney "the quintessential copyright free-rider and a bottomless pit of plagiarism."[1][4] Midjourney denies wrongdoing, arguing that training on publicly available images is transformative fair use.[5] In September 2025, Warner Bros. Discovery filed its own parallel suit against Midjourney, and by mid-2026 the Disney and Universal case had entered private mediation rather than reaching a verdict.[6][7]
The case is captioned Disney Enterprises, Inc. et al. v. Midjourney, Inc., Case No. 2:25-cv-05275, filed on June 11, 2025, in the U.S. District Court for the Central District of California.[1][8] The named plaintiffs are subsidiaries of the two media conglomerates:
| Plaintiff entity | Parent | Representative franchises asserted |
|---|---|---|
| Disney Enterprises, Inc. | The Walt Disney Company | Mickey Mouse, Frozen (Elsa), The Little Mermaid (Ariel), WALL-E |
| Marvel Characters, Inc. | The Walt Disney Company | The Avengers, Spider-Man, Iron Man |
| Lucasfilm Ltd. LLC | The Walt Disney Company | Star Wars (Darth Vader, Yoda) |
| Twentieth Century Fox Film Corporation | The Walt Disney Company | The Simpsons (Homer, Bart) |
| Universal City Studios Productions LLLP | NBCUniversal (Comcast) | Despicable Me / Minions |
| DreamWorks Animation L.L.C. | NBCUniversal (Comcast) | Shrek, Kung Fu Panda, How to Train Your Dragon |
The sole defendant is Midjourney, Inc., the San Francisco-based developer of the Midjourney image generator and its successive models, including Midjourney V7. The company was founded by and is led by CEO David Holz, and it operates without outside venture capital; press reports placed its annual revenue at roughly US$300 million in 2024, rising toward US$500 million in 2025.[2][9]
The complaint advances both direct and secondary copyright infringement theories. On direct infringement, the studios allege that Midjourney made unauthorized copies of "countless" copyrighted works when it scraped images from across the internet to assemble its training data, and that the service then reproduces those works as output. They describe Midjourney as "a virtual vending machine, generating endless unauthorized copies" of the plaintiffs' characters.[1][3]
To illustrate the output claims, the complaint contains dozens of side-by-side comparisons pairing stills from the studios' films with images generated by Midjourney from simple text prompts. According to the filing, prompts can readily produce recognizable renderings of Darth Vader, Yoda, the Minions, Shrek, Homer and Bart Simpson, Ariel, Elsa and WALL-E, among others.[1][4] The studios assert that Midjourney's own website "displays hundreds, if not thousands, of images generated by its Image Service at the request of its subscribers that infringe Plaintiffs' Copyrighted Works."[10]
On secondary infringement, the plaintiffs contend that Midjourney is liable for the infringing images its subscribers generate and distribute, framing the conduct as "calculated and willful."[4] The complaint also alleges that the infringement continued after the studios raised the issue: they say they contacted Midjourney and asked it to implement technical measures to stop generating their characters (the kind of content controls that some rival services use), but that Midjourney declined to adopt such measures and instead released newer, more capable versions of its product.[1][3]
The studios paired the complaint with pointed public statements. Horacio Gutierrez, Disney's senior executive vice president and chief legal and compliance officer, said that "piracy is piracy, and the fact that it's done by an AI company does not make it any less infringing." Kim Harris, NBCUniversal's executive vice president and general counsel, said "Theft is theft regardless of the technology used, and this action involves blatant infringement of our copyrights."[2][11]
The plaintiffs seek injunctive relief and damages. The requested injunction would bar Midjourney from offering its service without copyright protections; commentators noted that such relief could force at least a temporary shutdown of the product. The studios also seek statutory damages for willful infringement, which can reach US$150,000 per infringed work.[3][12]
Midjourney has contested the suit and signaled it will defend on fair use grounds. In its response, filed in August 2025, the company argued that the way it trains its models is fundamentally different from copying and reselling another company's content, characterizing the use of publicly available images as transformative. It moved to have the complaint dismissed with prejudice and sought recovery of attorneys' fees.[5][13]
A recurring theme in Midjourney's defense is that the studios are themselves participants in the generative-AI ecosystem. The company argued that "Plaintiffs cannot have it both ways, seeking to profit, through their use of Midjourney and other generative AI tools, from industry-standard AI training practices."[5][6] CEO David Holz has long likened model training to human learning, asking whether a person may "look at somebody else's picture and learn from it and make a similar picture," and arguing that "to the extent that AIs are learning like people, it's sort of the same thing."[14] The studios have used Holz's own earlier words against the company: in a 2022 interview he described his approach to data as employees who "grab everything they can, they dump it in a huge file, and they kind of set it on fire to train some huge thing," a quote later cited in Warner Bros. Discovery's complaint.[15]
The suit is significant primarily because of who brought it. Prior high-profile AI copyright actions in the image space were led by a stock-photography licensor (Getty Images v. Stability AI) or by individual visual artists (Andersen v. Stability AI), while text-focused cases were brought by publishers and authors, including New York Times v. OpenAI, Authors Guild v. OpenAI and Kadrey v. Meta. Disney and NBCUniversal are among the most litigious and well-resourced rights holders in the world, and their entry marked the first time major Hollywood studios directly challenged a generative-AI developer.[2][3]
Legal observers also highlighted the unusually vivid evidence. Because the asserted works are globally recognized fictional characters rather than generic stock photographs, the side-by-side output comparisons are visually striking and may resonate strongly with a jury, a contrast frequently drawn with the Getty litigation.[16] Beyond infringement, the case raises questions about whether the future of AI and entertainment lies in licensing deals between studios and AI developers rather than uncompensated scraping, and it has become a closely watched test of how the fair use doctrine applies to training data and to character-level outputs.[3][17]
On September 4, 2025, Warner Bros. Discovery filed its own copyright infringement suit against Midjourney in the Central District of California, becoming the second Hollywood conglomerate to do so. Its complaint alleges that Midjourney "brazenly dispenses Warner Bros. Discovery's intellectual property," producing recognizable versions of DC characters such as Superman, Batman, Wonder Woman and the Joker, along with Looney Tunes (Bugs Bunny, Tweety), Tom and Jerry, Scooby-Doo and Rick and Morty, even when prompts do not name them. Warner Bros. Discovery described the alleged infringement as "systematic, ongoing, and willful" and, like the other studios, sought Midjourney's profits or statutory damages of up to US$150,000 per work. Disney said it was "pleased to be joined by Warner Bros. Discovery."[6][7][18]
In the Disney and Universal case itself, the parties moved toward settlement discussions rather than an early ruling. After a late-2025 status conference, the court directed the parties toward mediation, and in December 2025 they selected the Hon. Suzanne H. Segal, a retired magistrate judge of the Central District of California, as a private neutral.[19][20] As of mid-2026, no judgment or settlement had been announced; the schedule set a post-mediation status conference for August 31, 2026, with a mediation-outcome report due August 21, 2026, fact discovery closing September 21, 2026, and a motions deadline of November 23, 2026.[8][19]
The Midjourney suits are part of a broad wave of AI copyright litigation working through courts in the United States and abroad, with mixed early signals. In Andersen v. Stability AI, the U.S. District Court for the Northern District of California allowed key infringement theories to proceed into discovery, with trial scheduled for September 2026, but it had not ruled on whether training is fair use. In New York Times v. OpenAI, a court declined to dismiss the core direct-infringement claims and ordered extensive discovery. By contrast, in Getty Images v. Stability AI, the High Court of England and Wales handed down a judgment on November 4, 2025 that rejected Getty's central copyright claim, holding that an AI model's trained parameters are not a stored "copy" of the training images under UK law, while finding only limited trademark liability. These divergent outcomes underscore that the questions at the heart of the Midjourney cases, namely whether training on and reproducing copyrighted works is infringement or fair use, remain unsettled.[21][22][23]