Midjourney

Midjourney is an artificial intelligence image generation service and independent research lab headquartered in San Francisco, California, that turns natural language text prompts into images using diffusion models. Founded in August 2021 by David Holz, Midjourney launched its open beta on July 12, 2022, and has grown into one of the most widely used generative AI tools in the world; TIME reported in August 2025 that its image and video products had more than 20 million users, mostly on Midjourney's server on Discord, the most populous on that platform.^[1][2][3] The company is unusual among frontier AI firms: it has never taken outside funding, Holz said in mid-2022 that it was already profitable, and TIME reported that it made about $300 million in 2024.^[2][31][104] Because it has never raised a priced funding round, no valuation for the company has been disclosed. Midjourney is known for a distinctive cinematic, painterly aesthetic that has made it a preferred tool for many concept artists, designers, and hobbyists, and as of October 2026 its default image model is V8.2, released on July 24, 2026.^[4][25][92] In June 2026 the company expanded beyond image and video generation into medical imaging hardware, announcing Midjourney Medical, a venture built around a whole-body ultrasound scanner, and in July 2026 it announced its first acquisition, the astrology app Co-Star.^{[85][86][93][94]}

Company overview

Midjourney, Inc. describes itself as "an independent research lab exploring new mediums of thought and expanding the imaginative powers of the human species." The company was founded by David Holz, who previously co-founded Leap Motion, a sensor technology company that developed optical hand-tracking devices for virtual and augmented reality, and served as its chief technology officer.^[7] Leap Motion was sold to British firm Ultrahaptics (now Ultraleap) in 2019 for a reported $30 million; it had been valued at about $300 million as of its last funding round.^[7][106] Holz grew up in Florida and studied physics and mathematics. He was working toward a PhD in applied mathematics at the University of North Carolina at Chapel Hill, with research stints at NASA's Langley Research Center and the Max Planck Institute, before leaving the program to start Leap Motion.^{[7][104][105]} He told The Verge in 2022 that he ran Leap Motion for 12 years and that "eventually, I was looking for a different environment instead of a big venture-backed company, and I left to start Midjourney."^[105] Holz is an alumnus of the 2006 International Science and Engineering Fair, and in 2026 he joined the Board of Trustees of the Society for Science, which runs the fair.^[7]

After leaving Leap Motion, Holz turned his attention to diffusion models. In a 2022 interview with The Register, he pointed to OpenAI's CLIP (Contrastive Language-Image Pre-training), transformers, and denoising diffusion models as the developments behind the new wave of image generators, and singled out CLIP-guided diffusion, developed by Katherine Crawson, as "the one that really struck my eye personally."^[104][36] Holz assembled a small initial team in San Francisco (he described Midjourney as "about 10 people" in mid-2022), and the team had a working private demo running by September 2021.^[6][104]

Who owns Midjourney and how is it funded?

Midjourney is privately held and was founded and is led by David Holz; it has never raised external venture capital and funds itself through subscription revenue.^[4][11][31] When Meta licensed its technology in August 2025, Holz wrote that Midjourney remained "an independent, community-backed research lab, with no investors," and the June 2026 Midjourney Medical announcement repeated that "Midjourney has no investors."^[86][103] Because the company has never raised a priced round, no valuation has been disclosed. In an interview published by The Register on August 1, 2022, a few weeks after the open beta began, Holz said the company was "already profitable," and TIME wrote in 2025 that Midjourney had become profitable within a year without investors.^[2][104]

Lean operations

Midjourney has kept a small team throughout its growth. Holz described the company as "about 10 people" in mid-2022, and TechCrunch put the team at fewer than 100 people in August 2024.^[104][107] Third-party trackers estimated the 2025 workforce at between 107 and 163 employees, depending on the source.^[9][10] Data provider GetLatka has estimated that the company reached roughly $500 million in revenue and about 100,000 paying customers in 2025; these are estimates rather than company disclosures.^[81]

Most revenue figures for Midjourney are estimates or press reports. GetLatka lists about $50 million for 2022 (a figure it attributes to The Decoder) and an estimate of about $500 million for 2025.^[5][81] For 2023, Winbuzzer, citing an interview Holz gave to The Information, reported that annual revenue "significantly surpassed" $200 million; TIME wrote that the company reportedly made $300 million in 2024.^[2][13] The company's growth has come largely through its community and word of mouth rather than paid marketing.^[11][12] Winbuzzer's March 2026 report, also citing The Information, said that Midjourney's hardware ambitions could require it to seek venture capital for the first time.^[13]

The following table summarizes Midjourney's reported scale and growth.

In August 2025, TIME named David Holz to its TIME100 AI list. The profile described Midjourney's "distinctive image and video generation products," pointed to the 24/7 Midjourney TV stream, and said the company "stands out for its focus on community," with user creations public by default.^[2]

Leadership

Holz is Midjourney's founder and chief executive.^[7] In July 2026, Banu Guler, the founder of the astrology app Co-Star, joined Midjourney as chief design officer after the company acquired Co-Star; Holz wrote that she would help "build a broader organization around Product and Design."^[93][94] Other executives named in coverage include head of hardware Ahmad Abbas, who oversees consumer hardware, and Tom Calloway, head of medical.^{[41][85][107]}

History

Origins and private demo (2021)

David Holz began working on Midjourney in mid-2021, motivated by the rapid advances in text-to-image generation that followed the release of CLIP. CLIP showed that AI systems could effectively assess the alignment between text descriptions and generated images, sparking a wave of research in the field, including OpenAI's DALL-E and the Latent Diffusion Models paper that became the foundation for Stable Diffusion.^[6][14] Holz set up an independent lab in San Francisco with an initial team of about ten people, and the team had a working private demo by September 2021.^[6][15] In 2022 he described Midjourney as "a wide and diverse research lab" with themes such as "reflection, imagination, and coordination," and called its AI image generation work "an engine for the imagination."^[105]

When was Midjourney released? Discord launch and open beta (2022)

Midjourney launched a public Discord server in February 2022, and shipped its first model version (V1) the same month. The model produced abstract and painterly images with limited coherence, but it demonstrated the basic capability and built an early community of testers. An invitation-only private beta opened in March 2022. On July 12, 2022, Midjourney announced its open beta, making the tool available to the general public through Discord, with images visible to all users in shared channels.^[15][16][17] The platform attracted a one-million-user Discord server within three months of public launch.^[4][16]

Rapid growth and mainstream attention (2022 to 2023)

The open beta attracted users at an unusual pace. By the end of 2022, Midjourney had released its fourth model version and was iterating rapidly on image quality. The platform's social nature, in which every image was visible in public Discord channels by default, created a built-in viral loop: users could see what others were creating, learn from prompts, and share striking results on Twitter and Reddit.^[4][17]

In September 2022, the platform received mainstream attention when Jason Allen won first place in the digital arts category at the Colorado State Fair with an image titled "Théâtre D'opéra Spatial," which he had produced using Midjourney. The result ignited widespread debate about AI-generated art and its place in creative competitions.^[18][19] In March 2023, a Midjourney-generated image of Pope Francis in a white Balenciaga-style puffer jacket went viral on social media, fooling millions of viewers who believed it was a real photograph. The incident, described as "the first real mass-level AI misinformation case," prompted Midjourney to end its free trial program shortly afterward.^[20][21]

Web platform, video, and partnerships (2024 to 2025)

Throughout 2023 and 2024 Midjourney continued to improve its models and expand beyond Discord. An alpha web interface at midjourney.com opened to high-volume users (the "5,000 Club") in late 2023, and the full web app launched alongside model version 6.1 in August 2024, allowing users to generate and edit images without a Discord account.^[22][23] On April 3, 2025, the company released V7 in alpha; on June 17, 2025, V7 became the default model.^[24][25]

On June 18, 2025, Midjourney released V1 Video, its first video generation model, accessed through an image-to-video workflow.^[26][27] On July 25, 2025, the company rolled out looping clips, custom start and end frames, and Discord access to video generation.^[28] On July 30, 2025, Midjourney launched midjourney.tv, a 24/7 livestream of trending AI-generated videos made by users.^[29][30] In August 2025, Meta announced a partnership to license Midjourney's image and video technology for use across Facebook, Instagram, and WhatsApp; Holz publicly reaffirmed that Midjourney would remain independent and investor-free.^[31][32]

V8 series and expansion beyond images (2026)

On March 17, 2026, Midjourney released the V8 alpha at alpha.midjourney.com, with native 2K HD rendering, faster generation, and better prompt adherence.^[33] V8.1 followed on the alpha site on April 14, 2026, reached midjourney.com and Discord on April 30, and became the default model on June 10, 2026; Midjourney's documentation calls it the company's fastest model so far.^[25][34][77] V8.2, an update focused on aesthetics, image quality, and personalization, became the default on July 24, 2026, and a V8.2 image edit model opened for community testing on August 27, 2026.^[25][92][95]

The company also moved into new kinds of work. On June 17, 2026, Holz unveiled Midjourney Medical and a full-body ultrasound scanner at an event in San Francisco, and in July 2026 Midjourney announced that it had acquired Co-Star; in the Co-Star announcement Holz described Midjourney Medical as the beginning of "the slow blossoming of a very unusual research lab" and said that "half a dozen similarly ambitious projects" would be announced over the following six months.^[93][108] Both are described in more detail below.

How it works

Technical architecture

Midjourney uses closed-source, proprietary software based on latent diffusion model architecture. While the company does not publicly disclose its exact model architecture or training data, the general approach is well understood. The system combines a large language model for text encoding with a diffusion model for image synthesis.^[35]

The image generation process involves several key stages:

1. Text encoding. The user's text prompt is processed by a text encoder, likely a variant of CLIP or a similar transformer-based model. This converts the natural language prompt into dense vector embeddings that capture semantic meaning, artistic style, color, mood, and compositional intent.^[35]
2. Latent diffusion. The text embeddings guide a diffusion process in a compressed latent space. Starting from random Gaussian noise, the model iteratively denoises the latent representation over many steps. A U-Net architecture with cross-attention mechanisms aligns the denoising process with the text embeddings at each step.^[35][36]
3. Decoding. A Variational Autoencoder (VAE) decoder converts the final latent representation back into pixel space, producing the output image.^[35]

Midjourney's models are trained on a large dataset of images sourced from the internet, including images that overlap with LAION's open-source datasets, particularly LAION-5B.^[37][38] The company has also incorporated user feedback data into its training pipeline, using information about which images users upscale, favorite, or select as a reinforcement learning signal to improve aesthetic quality.^[35] This human-in-the-loop signal is one explanation for the platform's pronounced house style.

Major model versions starting with V4 have been built on what Midjourney calls its "AI superclusters," and V4 in particular was an entirely new codebase trained on a new Google TPU cluster.^[39][40] V6 was the third Midjourney model trained from scratch and took roughly nine months to develop.^[39] Holz has framed Midjourney's longer-term goal in terms of real-time, open-world simulation rather than static images, with V1 Video described publicly as a "stepping stone" toward that target.^[26][41]

Interface and access

Midjourney provides two primary interfaces for generating images:

On Discord, users interact with the Midjourney Bot by typing slash commands in text channels. The bot processes the request and returns results directly in the channel. On the web interface, users type prompts into an "Imagine" bar and can adjust settings such as aspect ratio, model version, stylization, and speed through a sidebar panel.^[22]

Commands and features

Midjourney offers a range of commands and features for creating and modifying images. The core workflow involves generating an initial set of images, then refining the results through upscaling, variations, region edits, and outpainting.

Core commands

Image modification tools

After generating an initial grid of four images, users can refine their results using a set of modification tools:^[42]

Web Editor

The web Editor at midjourney.com provides a single interface for Remix, Vary (Region), Pan, and Zoom Out, with freehand and rectangular selection tools that target the affected area.^[42] In V7 and later, the Editor also supports conversational interaction, allowing users to describe changes in natural language (for example, "replace the cat with an owl" or "change the sky to sunset") inside Draft Mode.^[24][43]

Reference parameters

In addition to text prompts, Midjourney supports image-based reference parameters that lock in specific characters, styles, or moods across multiple generations:

Parameters and prompt engineering

Midjourney supports a variety of parameters that users append to prompts to control the output. Effective use of these parameters is a key part of prompt engineering for visual generation.

Key parameters

Prompt writing tips

Effective Midjourney prompts tend to follow several general principles, as advised by the company and its documentation team:^[49][50]

• Be specific about subject and style. Including details about lighting, composition, medium (watercolor, oil painting, photography), and artistic influences produces more targeted results.
• Reference artists by name. Midjourney can emulate recognizable artistic styles. For example, "a temple by James Gurney" will produce results influenced by Gurney's illustration style. Multiple artist names can be combined.
• Use positive language. Describing what should be present works better than describing what should be absent. Instead of "a hat that is not red," use --no red.
• Specify composition. Terms like "ultrawide shot," "close-up," "bird's-eye view," or "cinematic lighting" help control camera perspective and visual framing.
• Use concrete nouns. Visually well-defined concepts with many reference images on the internet (such as "wizard," "castle," or "mountain") tend to produce better results than abstract or highly conceptual terms.
• For V5 and later, Holz publicly recommended users stop writing keyword lists and instead use complete, descriptive sentences.^[51]
• For V6 and later, prompts can be up to 150 tokens long, double V5's 75-token limit.^[39]

Model versions

Midjourney has released multiple model versions since its initial launch, each representing significant improvements in image quality, prompt understanding, coherence, and speed. The following table summarizes the main image model versions.^{[17][25][33][34][39][52][53][92]}

Niji models

Alongside its main image line, Midjourney offers the Niji series, developed in collaboration with Spellbrush. The Niji models specialize in anime, manga, and illustrative styles influenced by East Asian aesthetics.^[54][55]

Pricing and subscription plans

How much does Midjourney cost?

Midjourney operates on a subscription-based model, with paid plans starting at $10 per month and rising to $120 per month. As of 2025 the platform no longer offers any free access, having discontinued its free trial in April 2023 following the viral Pope Francis image.^[20][56] All plans include commercial usage rights and access to the member gallery.

Fast GPU time is used for priority processing and provides the fastest generation speeds. Once Fast hours are depleted, users on the Standard plan and above can switch to Relax Mode, which queues jobs at lower priority with variable wait times. Unused Fast hours do not roll over between billing cycles.^[56]

Stealth Mode, available only on Pro and Mega plans, prevents generated images from appearing in public galleries on midjourney.com. The feature is important for commercial users who need to keep work confidential.^[56] Turbo Mode is available for all subscribers, generating images at approximately double the GPU time cost. Video generation uses approximately 8 times the GPU time of an image job, with each job producing four 5-second clips.^[26][27]

Video generation

On June 18, 2025, Midjourney released its first video model, designated V1 Video. The feature uses an image-to-video workflow: users first generate or upload a still image, then press an "Animate" button to convert it into a short video clip.^[26][27]

Key details of V1 Video include:

• Duration. Initial clips are 5 seconds long, with the option to extend in roughly 4-second increments up to four times, allowing a maximum length of approximately 21 seconds.^[27]
• Motion control. Users can choose between "high motion" and "low motion" settings. Low motion works well for ambient or contemplative scenes; high motion suits action-oriented content.^[27]
• Prompt modes. An automatic mode generates a motion prompt based on the image content, while a manual mode lets users describe how they want elements to move.^[26]
• Cost. Video generation consumes approximately 8x the GPU time of a standard image job. Each job produces four 5-second video options. Midjourney has framed the per-second cost as "over 25 times cheaper than what the market has shipped before."^[26]
• Availability. Included in existing subscription plans, but draws from the user's GPU time allocation. Available at launch through the web interface; added to Discord on July 25, 2025.^[28]

On July 25, 2025, Midjourney added looping clips (low motion or high motion), custom start and end frames, and an option to extend existing videos by appending new end frames.^[28] On July 30, 2025, the company launched midjourney.tv, a 24/7 livestream that displays trending V1 Video clips with their prompts and creator names visible on hover. The channel is accessible at midjourney.tv and on YouTube.^[29][30]

Competition and market position

How does Midjourney differ from DALL-E, Stable Diffusion, and other tools?

Midjourney operates in an increasingly competitive AI image generation market. As of 2025, Midjourney is estimated to hold around a quarter of the global AI image generator market share.^[3][57] Its primary differentiator is aesthetic quality rather than precise prompt-following or local control: Midjourney is a closed-source subscription service tuned for a cinematic, painterly look, whereas rivals such as Stable Diffusion 3 are open-weight and locally runnable, and DALL-E 3 and gpt-image-1 emphasize conversational generation and text rendering inside ChatGPT. Each major platform has distinct strengths:

Midjourney's competitive advantage lies primarily in aesthetic quality. The platform produces images with a distinctive cinematic, richly detailed look that many users and professionals consider superior for artistic and conceptual work. For tasks requiring precise text rendering, exact spatial control, or photorealistic product photography, some competitors may perform better, although the V8 alpha narrowed those gaps substantially.^[33][58]

The Meta partnership announced in August 2025, in which Meta licensed Midjourney's "aesthetic technology" for integration into Facebook, Instagram, and WhatsApp, is widely understood as Meta's attempt to close the quality gap between its in-house image tools and rivals such as Sora, FLUX, and Veo.^[31][32]

Art, creativity, and cultural impact

What is Midjourney used for?

Midjourney has had an outsized impact on the creative landscape since its launch. It is used to generate concept art, illustrations, mood boards, marketing visuals, album and magazine covers, and short videos, and the platform has been adopted by a wide range of users, from amateur hobbyists exploring their imagination to professional architects, graphic designers, concept artists, illustrators, and advertising agencies using it to prototype visual ideas.^[4][59]

Professional and commercial use

Architects and interior designers use Midjourney to visualize design concepts before committing to detailed plans. Graphic designers use it to generate initial mood boards and explore creative directions. Game studios and film production companies have used it for concept art during pre-production. Advertising agencies have adopted the tool for rapid ideation, generating dozens of visual concepts in the time it would traditionally take to produce a single sketch.^[4][59]

Creative community

Midjourney's Discord-based community has fostered a culture of collaborative experimentation. Public channels function as a live gallery where users observe each other's prompts and results, learn techniques, and share discoveries. This social dimension has been a primary factor in the platform's growth and in the broader popularization of AI-generated art.^[17] TIME's 2025 profile of Holz singled out this focus on community, noting that user creations are public by default so that others can remix and iterate on them.^[2]

Music and entertainment

The platform has been used in music videos and television production. Hungarian DJ and animator Shane 54 created an entire music video from Midjourney-generated frames. Several magazines, including The Atlantic and The Economist, have used AI-generated artwork on their covers. Midjourney TV, launched in July 2025, surfaces and continuously plays user-generated short video clips, giving a window into the volume and diversity of work being produced on the platform.^[29][30]

Impact on artists

The rise of Midjourney and similar tools has sparked significant debate within the art community. Some artists have embraced the technology as a powerful new creative tool. Others have voiced deep concern about its potential to reduce demand for human illustrators and designers. Cartoonist Matt Bors captured this tension in 2022, stating: "To developers and technically minded people, it's this cool thing, but to illustrators it's very upsetting because it feels like you've eliminated the need to hire the illustrator."^[59] In January 2024, an alleged "Midjourney Style List" containing the names of more than 16,000 artists circulated online and later surfaced in court filings, reigniting the debate over consent and training data.^[60]

David Holz's philosophy

Holz has consistently framed Midjourney as a tool for human imagination rather than a replacement for human creativity. In a 2022 interview he said, "We like to say we're trying to expand the imaginative powers of the human species. The goal is to make humans more imaginative, not make imaginative machines, which I think is an important distinction." In the same era he described Midjourney as "an engine for the imagination," and he has framed Midjourney's longer-term goal in terms of real-time open-world simulation: "In 10 years, you'll be able to buy an Xbox with a giant AI processor, and all the games are dreams."^[8][61][105]

Controversies and legal issues

Colorado State Fair incident

In August 2022, Jason Allen entered an image titled "Théâtre D'opéra Spatial" in the digital arts category at the Colorado State Fair's fine art competition. The image, which depicted classical figures in an ornate hall looking through a circular viewport, won first place in the "emerging artist" (non-professional) division on August 29 and a $300 prize. Allen had used Midjourney to generate the image, employing at least 624 text prompt iterations before editing the result in Adobe Photoshop and upscaling it with Gigapixel AI.^[18][19]

The award drew widespread controversy when it became public that the winning entry was AI-generated. Many traditional artists accused Allen of cheating and argued that AI-generated art should not compete alongside human-created work. The two judges later said they did not know Midjourney used AI but would have awarded Allen the top prize regardless. Allen maintained that he was using Midjourney as a creative tool, similar to how other artists use digital software.^[18][19] Starting in 2023, the Colorado State Fair required participants to disclose whether they used AI.^[18]

The controversy deepened when Allen applied for copyright protection for the image. On September 5, 2023, the U.S. Copyright Office Review Board issued a final determination that "Théâtre D'opéra Spatial" could not receive copyright registration, concluding that Allen's contribution (typing text prompts) was insufficient to establish human authorship.^[62] In 2024, Allen filed an appeal in U.S. District Court in Colorado, challenging the Copyright Office decision.^[62]

Pope Francis deepfake

In March 2023, a Midjourney-generated image of Pope Francis wearing a large white Balenciaga-style puffer jacket went viral across social media platforms, fooling millions of viewers. The image was created by Pablo Xavier, a 31-year-old construction worker in Chicago, and posted to a Facebook group called AI Art Universe before spreading to Reddit and other platforms.^[20][21]

Newsletter writer Ryan Broderick called it "the first real mass-level AI misinformation case." While the image contained telltale signs of AI generation (such as distorted hands and overly sharp skin textures), the overall quality was convincing enough to deceive casual viewers. Shortly after the incident, Midjourney discontinued its free trial program. The company also implemented content restrictions; for example, the word "pope" was temporarily restricted, and a mid-2023 algorithm update further enhanced the platform's content filters around real public figures.^[20][21]

Training data and copyright lawsuits

Is Midjourney being sued for copyright?

Yes. Midjourney's use of copyrighted images as training data has become the subject of multiple high-profile lawsuits, including a landmark June 2025 suit brought by Disney and Universal:

• Andersen v. Stability AI (January 2023). A class-action lawsuit filed in the Northern District of California on January 13, 2023, on behalf of artists Sarah Andersen, Kelly McKernan, and Karla Ortiz alleged that Midjourney, Stability AI, and DeviantArt violated copyright by training their models on billions of images scraped from the internet, particularly the LAION-5B dataset.^[63][64] In August 2024, U.S. District Judge William Orrick allowed the case to proceed on its core copyright claims against Midjourney, Stability AI, DeviantArt, and Runway, denying motions to dismiss for those claims.^[65]
• Disney and Universal v. Midjourney (June 2025). On June 11, 2025, Walt Disney Studios, NBCUniversal, and DreamWorks Animation filed a 110-page complaint in the U.S. District Court for the Central District of California (case 2:25-cv-05275) alleging direct and secondary copyright infringement, the first major lawsuit by Hollywood studios against a generative AI image company.^[66][67][82] The lawsuit included visual evidence showing that Midjourney could generate highly recognizable depictions of copyrighted characters from franchises including Marvel, Star Wars, Shrek, The Simpsons, and Despicable Me, even from generic prompts. The plaintiffs described Midjourney as a service that "functions as a virtual vending machine, generating endless unauthorized copies" of their copyrighted works. The complaint identified 199 specific infringing images.^[66][67][82]
• Warner Bros. Discovery v. Midjourney (September 2025). On September 4, 2025, Warner Bros. Entertainment Inc., DC Comics, Turner Entertainment Co., Hanna-Barbera Productions, and Cartoon Network filed a separate 87-page copyright lawsuit in the Central District of California, alleging "systematic" and "willful" infringement. The complaint highlighted characters including Superman, Batman, Wonder Woman, Bugs Bunny, Scooby-Doo, Tom and Jerry, the Powerpuff Girls, and Rick and Morty, and seeks $150,000 per infringed work. The complaint also pointed to "Midjourney TV" as evidence the company was publicly performing infringing works.^[68][69]
• Consolidation and discovery (late 2025 to 2026). In November 2025, the Warner Bros. case (2:25-cv-08376) was consolidated for all purposes, including trial, with the earlier-filed Disney case (2:25-cv-05275), which serves as the lead case.^[70] In scheduling proposals filed that month, Disney sought a non-expert discovery cutoff of August 17, 2026 and a December 21, 2026 deadline for motions, while Midjourney asked for dates about three months later.^[83] The pretrial schedule the court set, as reported in January 2026, called for a report on the outcome of mediation by August 21, 2026, the close of non-expert discovery on September 21, 2026, and the close of expert discovery on November 9, 2026.^[102] On June 15, 2026, Magistrate Judge Joel Richlin denied Midjourney's request for broad discovery into the studios' own use of AI, limiting it to "consumer-facing" AI applications; in early July, Midjourney asked Judge John Kronstadt to overturn that ruling, arguing that such evidence "goes to the heart of Midjourney's fair use and unclean hands defenses."^[101] As of October 2026, Deadline described the lawsuit as ongoing.^[100]

Midjourney has argued that its training falls within fair use parameters. In a 43-page response to the Disney and Universal complaint, the company's lawyers wrote that "copyright law does not confer absolute control over the use of copyrighted works" and that "fair use safeguards public interests in the free flow of ideas and information."^[70][82] These lawsuits are part of a broader wave of more than 70 copyright cases filed against AI companies as of 2025, and their outcomes may set significant precedent for the entire generative AI industry.^[66] See Disney & Universal v. Midjourney for a fuller account of the studio litigation.

In October 2026, Deadline reported on a study by VN.ai, a startup that tests generative AI tools for copyright infringement, which compared 15 popular AI models. The study, conducted in September 2026, found Midjourney V8.1 to be the worst of the group at reproducing images of copyrighted characters, and found that it produced characters in ways that violated brand-safety measures (for example, in scenes of violence or drug use) in 37 percent of text prompts. Midjourney did not respond to Deadline's requests for comment.^[100]

Deepfake and misinformation concerns

Beyond the Pope Francis incident, Midjourney's improving photorealism has raised ongoing concerns about its potential for deepfakes and visual misinformation. A mid-2023 algorithm update further sharpened the platform's depiction of real public figures, including a series of fake images of a Pentagon attack that briefly moved markets.^[21][71] Midjourney maintains content moderation policies and employs a small team of moderators, and the platform prohibits adult content, gore, sexualized images of real people, and offensive or inflammatory depictions of public figures. Some prompts (such as the names of named politicians during election periods, and sensitive religious terms) are blocked or filtered automatically.^[72]

2025 to 2026 developments

V7 release

Midjourney V7 launched in alpha on April 3, 2025, and became the default model on June 17, 2025.^[24][25] In its official announcement, Midjourney called V7 "an amazing model," adding that it is "much smarter with text prompts" and that "bodies, hands and objects of all kinds have significantly better coherence on all details."^[24] V7 was a complete rebuild rather than an incremental update, and its main innovations were:

• Draft Mode, which the company describes as "half the cost and renders images at 10 times the speed." On the web, Draft Mode runs in a conversational prompt bar where users can iterate with short follow-ups or voice commands.^[24][43]
• Default-on personalization, the first Midjourney model to ship with personalization turned on by default. Midjourney's launch post states, "V7 is the first model to have model personalization turned on by default. You must unlock your personalization to use it. This takes ~5 minutes." Once set, that profile subtly tunes every subsequent generation.^[24][73][74]
• Omni Reference (--oref), a unified character and object reference system controlled with --ow (omni weight, range 1 to 1000, default 100). Unlike the V6 --cref parameter it replaces, Omni Reference handles external photos well and works across poses, outfits, and environments.^[46][47]
• Conversational editing on the web Editor, accessed during Draft Mode, where users can describe changes in natural language.^[43]
• Two launch speed modes, Turbo and Relax, with upscaling, retexturing, and certain edits still routed to V6 models at launch.^[24]

Video generation launch and Midjourney TV

Midjourney entered the video space on June 18, 2025, with V1 Video. The image-to-video workflow integrates with the existing image pipeline, allowing users to animate stills into 5-second clips that can be extended to roughly 21 seconds total. Each job produces four clips at approximately 8x the GPU cost of an image job, which Holz framed as roughly the cost of one image per second of video and "over 25 times cheaper than what the market has shipped before."^[26][27]

On July 25, 2025, looping clips, custom start and end frames, and Discord-side video generation rolled out.^[28] On July 30, 2025, Midjourney launched midjourney.tv, a 24/7 livestream of trending V1 Video clips along with personalized feeds.^[29][30]

Meta partnership

On August 22, 2025, Meta Chief AI Officer Alexandr Wang announced a partnership with Midjourney to license its image and video generation technology. The deal brings Midjourney's "aesthetic technology" into Meta's ecosystem, with plans to integrate AI-generated tools into Facebook, Instagram, and WhatsApp. Financial terms were not disclosed.^[31][32] Midjourney confirmed that it would continue operating independently. Holz wrote on X the same day, "We remain an independent, community-backed research lab, with no investors."^[31][103] TechCrunch also cited a report by Upstarts Media that Meta had at one point talked with Midjourney about acquiring the company.^[31]

V8 alpha and V8.1

On March 17, 2026, Midjourney released the V8 alpha at alpha.midjourney.com. V8 introduces native 2K HD rendering via --hd (eliminating the need for separate upscaling), a new --q 4 quality mode, generation speeds roughly 5x faster than V7, and substantially improved prompt adherence, especially for text inside quotation marks. In its announcement, Midjourney wrote that "image generation is ~5x faster than before" and that "text rendering works better than ever (when specified in 'quotes')." The company also noted that --hd, --q 4, --sref, and Moodboard jobs each cost roughly 4x the GPU time of a regular job, so combining --hd with --q 4 costs about 16x. The alpha also brought a new grid view, sidebar settings, an improved conversation mode, and updated style reference and moodboard algorithms (--sv 7).^[33][58][84]

V8.1 arrived on alpha.midjourney.com on April 14, 2026. Midjourney said it had "a consistent and familiar aesthetic in the spirit of V7," made HD mode 3x faster and 3x cheaper (and therefore the default for V8.1), and made standard resolution 50% faster and 25% cheaper. V8.1 also brought back image prompts and image weights, added a Prompt Shortener that activates when a prompt exceeds the length limit, and updated Describe to match V8's prompting style.^[34] On April 30, 2026, V8.1 became available on midjourney.com and Discord.^[77] Midjourney's documentation describes V8.1 as its fastest model so far, with standard jobs rendering about 4 to 5 times faster than earlier versions, notes that users may need to unlock their Global V7/V8 personalization profile to use it, and records that it was the default model from June 10 to July 23, 2026.^[25][75]

Spring 2026 rollout and web updates

After V8.1 reached midjourney.com and Discord on April 30, Midjourney spent May refining the model and the web experience rather than shipping a new version. The V8.1 update post noted that standard-definition mode had been made a temporary default during a server transition, with users able to switch back to HD through the settings panel or by adding --hd, and confirmed that community rating feedback was feeding the next model, V8.2.^[77]

Earlier in the V8 cycle the company also widened access and gathered ranking data. On March 21, 2026, Relax mode was turned on for V8 across Standard, Pro, and Mega subscribers, supporting most commands.^[78] On April 27, 2026, ahead of what it called "major aesthetic updates," Midjourney ran a ranking session at full 2K resolution (the first time it had ranked raw images at that resolution) to tune V8.1 and V8.2 toward more natively HD output, inviting subscribers to compare image pairs at midjourney.com/rank-v8-1.^[79]

On May 27, 2026, Midjourney pushed a batch of web platform updates. Conversational mode, including voice input, gained access to a session's Image Prompts, Style References, sidebar settings, and recent jobs, and tray images now persist across voice submissions. A new "Rerun as HD" button lets users re-render any V8.1 image generated in standard definition at HD. The update also added a hidden-item count to the Create and Organize folder views, grouped Profiles, Moodboards, and Liked Styles in the mobile settings menu, and enabled search for signed-in members without an active subscription.^[80] V8.1 became the default model on June 10, 2026, replacing V7.^[25]

Hardware exploration

In February 2024, Midjourney hired Ahmad Abbas, who had helped engineer Apple's Vision Pro headset and had previously worked with Holz at Leap Motion. On August 28, 2024, the company announced on X that it was "officially getting into hardware," with a new team based in San Francisco.^[76][107] The Decoder reported that the effort would be led by Abbas, as head of hardware, and Holz, and that the project was informally called "Midjourney Orb."^[41] In March 2026, Winbuzzer, citing an interview with The Information, reported that Midjourney's hardware plans could require it to seek venture capital for the first time.^[13] On June 16, 2026, Midjourney said it would unveil "our first secret hardware project," which it called "ambitious, (physically!) big, and unexpected," in a livestream the next day.^[113] The project turned out to be a medical scanner.

Midjourney Medical

On June 17, 2026, at an event in San Francisco that was also livestreamed on Discord and X, Holz announced a health division, Midjourney Medical, together with a blog post titled "A New Era of Midjourney," marking the company's first move beyond image and video generation and into AI in healthcare hardware.^{[86][108][109][113]} Engadget reported that the project is led by Tom Calloway, PhD, Midjourney's head of medical; it had first named head of hardware Ahmad Abbas as the lead, then corrected its story to say that Abbas oversees consumer hardware.^[85]

The product, which Midjourney calls the Midjourney Scanner, is a whole-body imaging machine. The company's website describes it as "full-body Ultrasonic CT," and The Register reported that the accompanying technical video calls the method "Fullbody Ultrasound Computational Tomography."^[86][110] Despite the "CT" name, the device uses no X-rays and emits no ionizing radiation; it is an ultrasound system. A person stands on a platform that lowers them into a shallow pool of water at about 2 inches (5 centimeters) per second, past a ring of ultrasonic elements that send sound waves through the body from many angles and record how the waves change (a process the company likens to the echolocation used by dolphins), and a large compute cluster reconstructs the signals into cross-sectional images.^[86] Midjourney describes the ring as "half a million tiny squares"; The Register reported that the prototype's ring of 40 modules contains 358,000 ultrasonic elements.^[86][110] Midjourney says its goal is for a scan to take no more than 60 seconds, compared with the 60 to 90 minutes a full-body MRI typically takes, and claims the result is a three-dimensional map of the body "down to a fraction of a millimeter, that looks a lot like today's MRIs but at nearly a hundred times the speed."^[85][86] Midjourney plans to deliver the scan to consumers through a venue it calls the "Midjourney Spa," with hot tubs, saunas, and cold plunges.^[86][88]

The imaging hardware does not come from Midjourney's generative models but from a partnership with Butterfly Network, Inc. (NYSE: BFLY), the medical-device maker behind a semiconductor "Ultrasound-on-Chip" technology. The current prototype system incorporates 40 Butterfly Ultrasound-on-Chip modules, and Butterfly has said future generations are expected to use substantially more.^[87] According to Butterfly's Form 8-K, on November 17, 2025 a Butterfly subsidiary entered into a Co-Development and Licensing Agreement that grants Midjourney an exclusive, non-transferable license, within a specified field of use, to Butterfly's ultrasound-on-chip technology, software, and backend technology. Midjourney agreed to pay a one-time fee of $15 million and a $10 million annual license fee, payable quarterly, over a five-year term, plus up to $9 million in milestone payments, revenue-sharing payments on hardware that incorporates Butterfly chips, and payments for any chip purchases.^[89] Butterfly's June 18, 2026 commentary put the expected payments at up to $74 million over the five years.^[87] Butterfly chief executive Joseph DeVivo said the company was "proud to support Midjourney's mission to democratize access to personal imaging data," and Butterfly shares rose about 33 percent in trading on June 18.^[87][111] In its second-quarter 2026 results, Butterfly attributed a 57 percent rise in U.S. revenue mainly to its Butterfly Embedded partnerships, including the co-development partnership with Midjourney, and to higher probe sales.^[112]

The following table summarizes the disclosed and reported terms of the Butterfly Network agreement.

Midjourney has laid out an ambitious roadmap. Over the year after the announcement it planned to refine its algorithms and hardware, run research trials, move toward a second-generation design, and build a first "research spa." It aims to open its first Midjourney Spa in San Francisco around the end of 2027, to follow in 2028 with expansion to more cities and a third-generation machine built on custom silicon, and to reach a goal of more than 50,000 scanners worldwide with capacity for a billion scans a month by 2031.^[85][86][88] "Our goal is to build a fleet of 50,000 of these scanners," Holz said at the launch event, declining to say what the product will cost.^[108]

Independent observers urged caution about Midjourney's claims, which at the time of the announcement were self-reported and had not been independently validated. The headline figures are presented by the company as goals: Midjourney's own post says "the goal is for this process to take no more than 60 seconds," and The Register wrote that it was not clear how fast the prototype unit actually was.^[86][110] Holz said, "We're not even using any AI in this yet, just really cool hardware and software"; The Next Web noted that the company uses AI to segment and label what a scan shows, while the imaging itself relies on ultrasound and signal processing.^[108][109] Commentary published after the launch reported that the first-generation prototype took roughly 20 minutes per scan, had been used on about a dozen people, and did not yet use a neural network in its reconstruction.^[90][91] Commentators also questioned the marketing language on physics grounds, noting that an ultrasound device is not a CT scanner and that comparisons with MRI had not been independently validated.^[88][91] The regulatory path is likewise unsettled: the scanner has no FDA clearance, and Midjourney has said it will start by providing "detailed body composition maps" and will submit test results to the FDA for increased capabilities.^[86][109] One healthcare analysis compared this to the general-wellness positioning used by full-body screening firms such as Prenuvo and Ezra and raised concerns about overdiagnosis and incidental findings at the scale Midjourney envisions.^[91] Butterfly's own commentary cautioned that actual results could differ because of development, regulatory, manufacturing, and market risks.^[87]

V8.2, the Edit Model, and Thinking Mode

On July 24, 2026, Midjourney launched V8.2 and made it the default model. The announcement said the update "focuses on aesthetics and image quality and personalization," that images should be "more creative, bold, sophisticated, edgy and fresh," and that personalization would better reflect a user's taste, especially for profiles with many ratings.^[25][92]

On August 27, 2026, Midjourney opened public testing of its first V8.2 image edit model. It supports editing images with written instructions, generating images from up to four image references at once (replacing Omni Reference), inpainting and outpainting, and use of personalization, moodboards, and style references; on Discord it is invoked with --edit followed by an image URL.^[95] Midjourney's documentation says the Edit Model replaces Omni Reference, Character Reference, and the Retexture tool.^[25]

From August to October 2026 the company used alpha.midjourney.com to test a redesigned web interface, publishing changelogs that added project folders and groups, saved default parameters, live style previews generated by what it called "fast models," Korean-language support (extended to midjourney.com in September), and shared folders in which invited collaborators can generate images into the same folder.^[96][97][99] In early October 2026 it began testing "Thinking Mode" on the alpha site, a rerun option for V8.2 jobs in which, according to the changelog, Midjourney "looks at your image, finds where it missed your prompt (wrong objects, layout, anatomy, text) and generates again." The company said its tests showed gains in prompt accuracy, typography, and coherence.^[97][98]

Acquisition of Co-Star

In a blog post dated July 23, 2026, Holz announced Midjourney's first acquisition: Co-Star, the astrology app founded by Banu Guler. He wrote that he had first met Guler "a little over 5 years ago" and that she and her team would join Midjourney "to help build a broader organization around Product and Design," while Co-Star itself would "remain entirely under Banu's control and direction."^[93] Midjourney told Gizmodo that the acquisition had happened in the spring; financial terms were not disclosed. Guler became Midjourney's chief design officer. Gizmodo, citing Bloomberg, reported that Co-Star had about 4.3 million monthly active users and that Midjourney planned to offer an astrology-focused image generator.^[94]

Related work

Midjourney sits within a broader ecosystem of text-to-image models and AI art tools. Adjacent technologies include the DDIM (Denoising Diffusion Implicit Models) sampler family, SDXL and other open-weight diffusion checkpoints, the Diffusion Transformer (DiT) architecture used by newer rivals, and the MMDiT backbone that underpins Stable Diffusion 3. In video, Midjourney V1 competes with Sora, Veo 3, Kling, Pika, Runway Gen-4, and Runway Aleph. In open-weight image generation, Midjourney is often compared with FLUX.1, FLUX.2, Stable Diffusion 3.5, and SDXL. Among closed services, the most direct comparisons are DALL-E 3, gpt-image-1, Imagen 3, Imagen 4, Adobe Firefly, and Ideogram 3.0.

Continue reading

• Diffusion modelA diffusion model is a generative model that learns to transform samples from a simple reference distribution into samples resembling a data distribution by reversing a gradual corruption process.Mentioned 6 times above
• Stable DiffusionStable Diffusion is a family of generative image models that can synthesize and edit images from text and other conditions.Mentioned 6 times above
• MMidjourney MedicalMidjourney Medical is a healthcare initiative announced in mid-June 2026 by Midjourney, the artificial intelligence research lab best known for its text-to-image generative AI service.Mentioned 6 times above
• DALL-EDALL-E is a family of text-to-image systems developed by OpenAI that generates images from natural-language descriptions.Mentioned 5 times above
• Generative AIGenerative AI is a class of artificial intelligence systems that produces new data instances, such as text, software code, images, audio, video, molecular structures, or other representations, by learning…Mentioned 4 times above
• Imagen (text-to-image model)Imagen is a family of text-to-image diffusion models developed by Google, first introduced in May 2022Mentioned 4 times above

More from See also

• AI art
• AI Image Generation
• Latent diffusion model
• DALL-E 3
• FLUX.1
• Adobe Firefly
• Stability AI
• Black Forest Labs
• Getty Images v. Stability AI
• LAION
• Deepfake
• Text-to-Image Models
• Text-to-video generation
• Midjourney V7
• Disney & Universal v. Midjourney
• AI in Healthcare
• Meta AI
• List of neolabs
• Neolab

Topics:AI CompaniesDiffusion ModelsGenerative AIImage Generation

References

1. ^Midjourney, "About", Midjourney official site, 2025. midjourney.com. Accessed 2026-05-24.
2. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8TIME, "David Holz: The 100 Most Influential People in AI 2025", TIME, 2025-08-27. time.com/...david-holz. Accessed 2026-05-24.
3. ^1 ^2 ^3Statista, "Midjourney: statistics & facts", Statista Research Department, 2025. statista.com/...midjourney. Accessed 2026-05-24.
4. ^1 ^2 ^3 ^4 ^5 ^6 ^7Mitra, S., "Ultralight Startup Midjourney Bootstraps to $200M in Revenue", Sramana Mitra, 2025-03-12. sramanamitra.com/...-bootstraps-to-200m-in-revenue. Accessed 2026-05-24.
5. ^GetLatka, "How Midjourney hit $500M revenue and 100K customers in 2025", GetLatka, 2025. getlatka.com/...midjourney. Accessed 2026-05-24.
6. ^1 ^2 ^3Tan, A., "How did Midjourney start", Over the Anthill (Substack), 2024. overtheanthill.substack.com/...id-midjourney-start. Accessed 2026-05-24.
7. ^1 ^2 ^3 ^4 ^5Society for Science, "David Holz", Society for Science alumni profile, 2024. societyforscience.org/...david-holz. Accessed 2026-05-24.
8. ^Thompson, B., "An Interview with Midjourney Founder David Holz about Generative AI, VR, and Silicon Valley", Stratechery, 2022-11-08. stratechery.com/...rative-ai-vr-and-silicon-valley. Accessed 2026-05-24.
9. ^1 ^2ElectroIQ, "How Many People Work At Midjourney? Statistics By Region And Category (2025)", ElectroIQ, 2025. electroiq.com/...how-many-people-work-at-midjourney. Accessed 2026-05-24.
10. ^1 ^2SEO.ai, "How Many People Work at Midjourney? Statistics & Facts (2025)", SEO.ai, 2025. seo.ai/...how-many-people-work-at-midjourney. Accessed 2026-05-24.
11. ^1 ^2 ^3Sacra, "Midjourney revenue, funding & news", Sacra, 2025-12-14. sacra.com/...midjourney. Accessed 2026-05-24.
12. ^SQ Magazine, "Midjourney Statistics 2026: From Growth Data to Big Impact", SQ Magazine, 2026. sqmagazine.co.uk/midjourney-statistics. Accessed 2026-05-24.
13. ^1 ^2 ^3 ^4Winbuzzer, "Midjourney Revenue Tops $200M as It Eyes Hardware", Winbuzzer, 2026-03-31. winbuzzer.com/...-200-million-hardware-push-xcxwbn. Accessed 2026-05-24.
14. ^Rombach, R. et al., "High-Resolution Image Synthesis with Latent Diffusion Models", arXiv:2112.10752, 2021-12-20. arxiv.org/...2112.10752. Accessed 2026-05-24.
15. ^1 ^2Wikipedia, "Midjourney", Wikipedia, 2026 revision. en.wikipedia.org/...Midjourney. Accessed 2026-05-24.
16. ^1 ^2Hachman, M., "Midjourney's enthralling AI art generator goes live for everyone", PCWorld, 2022-07-26. pcworld.com/...rneys-ai-art-goes-live-for-everyone. Accessed 2026-05-24.
17. ^1 ^2 ^3 ^4Aituts, "All Midjourney Versions (V1-V6) Compared: The Evolution of Midjourney", Aituts, 2024. aituts.com/midjourney-versions. Accessed 2026-05-24.
18. ^1 ^2 ^3 ^4Smithsonian Magazine, "Art Made With Artificial Intelligence Wins at State Fair", Smithsonian Magazine, 2022-09-06. smithsonianmag.com/...olorado-state-fair-180980703. Accessed 2026-05-24.
19. ^1 ^2 ^3Roose, K., "An A.I.-Generated Picture Won an Art Prize. Artists Aren't Happy.", The New York Times, 2022-09-02. nytimes.com/...ai-artificial-intelligence-artists. Accessed 2026-05-24.
20. ^1 ^2 ^3 ^4Stokel-Walker, C., "We Spoke To The Guy Who Created The Viral AI Image Of The Pope That Fooled The World", BuzzFeed News, 2023-03-27. buzzfeednews.com/...ourney-image-creator-interview. Accessed 2026-05-24.
21. ^1 ^2 ^3 ^4Berkowitz, A., "Pope Puffer: Midjourney Image of Pope in Balenciaga Is a Warning About AI Future", Bloomberg, 2023-03-28. bloomberg.com/...-warning-about-ai-future-lfssbcan. Accessed 2026-05-24.
22. ^1 ^2Lamsoge, P. C., "The New Midjourney Website Launched in Beta!", Medium, 2024. medium.com/...ebsite-launched-in-beta-c749c2678969. Accessed 2026-05-24.
23. ^Geeky Gadgets, "Midjourney 6 Alpha web browser version now available", Geeky Gadgets, 2024-01. geeky-gadgets.com/midjourney-web-browser-version. Accessed 2026-05-24.
24. ^1 ^2 ^3 ^4 ^5 ^6 ^7Midjourney, "V7 Alpha", Midjourney Updates, 2025-04-04. updates.midjourney.com/v7-alpha. Accessed 2026-05-24.
25. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10Midjourney Docs, "Version", Midjourney Documentation, 2026 (states that V8.2 became the default on July 24, 2026, and that V8.1 was released April 14, 2026 and was default from June 10 to July 23, 2026). docs.midjourney.com/...32199405667853-Version. Accessed 2026-10-27.
26. ^1 ^2 ^3 ^4 ^5 ^6 ^7Midjourney, "Introducing Our V1 Video Model", Midjourney Updates, 2025-06-18. updates.midjourney.com/...ucing-our-v1-video-model. Accessed 2026-05-24.
27. ^1 ^2 ^3 ^4 ^5 ^6Wiggers, K., "Midjourney launches its first AI video generation model, V1", TechCrunch, 2025-06-18. techcrunch.com/...rst-ai-video-generation-model-v1. Accessed 2026-05-24.
28. ^1 ^2 ^3 ^4Midjourney, "Looping and End Frame for Video + Video in the Discord Bot", Midjourney Updates, 2025-07-25. updates.midjourney.com/...video-in-the-discord-bot. Accessed 2026-05-24.
29. ^1 ^2 ^3 ^4Midjourney, "Personalized feeds and midjourney.tv", Midjourney Updates, 2025-07-30. updates.midjourney.com/...-feeds-and-midjourney-tv. Accessed 2026-05-24.
30. ^1 ^2 ^3 ^4Blockchain Council, "Midjourney's Midjourney TV", Blockchain Council, 2025-08. blockchain-council.org/...midjourney-tv. Accessed 2026-05-24.
31. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9Wiggers, K., "Meta partners with Midjourney on AI image and video models", TechCrunch, 2025-08-22. techcrunch.com/...ney-on-ai-image-and-video-models. Accessed 2026-05-24.
32. ^1 ^2 ^3Lee, A., "Meta is licensing Midjourney's AI image and video tech", Engadget, 2025-08-22. engadget.com/...-ai-image-and-video-tech-120012178. Accessed 2026-05-24.
33. ^1 ^2 ^3 ^4Midjourney, "V8 Alpha", Midjourney Updates, 2026-03-17. updates.midjourney.com/v8-alpha. Accessed 2026-05-24.
34. ^1 ^2 ^3Midjourney, "V8.1 Alpha", Midjourney Updates, 2026-04-14. updates.midjourney.com/v8-1-alpha. Accessed 2026-05-24.
35. ^1 ^2 ^3 ^4 ^5Britannica, "Midjourney: AI Technology, Discord, Tools, Versions, & History", Encyclopaedia Britannica, 2025. britannica.com/...Midjourney. Accessed 2026-05-24.
36. ^1 ^2Ho, J. et al., "Denoising Diffusion Probabilistic Models", arXiv:2006.11239, 2020-06-19. arxiv.org/...2006.11239. Accessed 2026-05-24.
37. ^LAION, "LAION-5B: An open large-scale dataset for training next generation image-text models", LAION blog, 2022-03. laion.ai/...laion-5b. Accessed 2026-05-24.
38. ^Cantor, M., "Leaked: the names of more than 16,000 non-consenting artists allegedly used to train Midjourney's AI", The Art Newspaper, 2024-01-04. theartnewspaper.com/...used-to-train-midjourney-ai. Accessed 2026-05-24.
39. ^1 ^2 ^3 ^4Franzen, C., "Midjourney V6 is here with in-image text and completely overhauled prompting", VentureBeat, 2023-12-21. venturebeat.com/...completely-overhauled-prompting. Accessed 2026-05-24.
40. ^Mad Genius Club, "Navigating MidJourney: V4", Mad Genius Club, 2022-11-12. madgeniusclub.com/...navigating-midjourney-v4. Accessed 2026-05-24.
41. ^1 ^2 ^3The Decoder, "Midjourney details plans for AI video model and officially launches 'Orb' hardware project", The Decoder, 2024-08-30. the-decoder.com/...y-launches-orb-hardware-project. Accessed 2026-05-24.
42. ^1 ^2Midjourney Docs, "Editor", Midjourney Documentation, 2025. docs.midjourney.com/...32764383466893-Editor. Accessed 2026-05-24.
43. ^1 ^2 ^3TitanXT, "A Look Inside Midjourney V7: Personalization, Draft Mode, and Faster Creating", TitanXT, 2025. titanxt.io/...ation-draft-mode-and-faster-creating. Accessed 2026-05-24.
44. ^Midjourney Docs, "Style Reference", Midjourney Documentation, 2025. docs.midjourney.com/...80011136653-Style-Reference. Accessed 2026-05-24.
45. ^Midjourney Docs, "Character Reference", Midjourney Documentation, 2025. docs.midjourney.com/...7505293-Character-Reference. Accessed 2026-05-24.
46. ^1 ^2Midjourney Docs, "Omni Reference", Midjourney Documentation, 2025. docs.midjourney.com/...36285124473997-Omni-Reference. Accessed 2026-05-24.
47. ^1 ^2Midjourney, "Omni-Reference --oref", Midjourney Updates, 2025. updates.midjourney.com/omni-reference-oref. Accessed 2026-05-24.
48. ^Midjourney Docs, "Moodboards", Midjourney Documentation, 2025. docs.midjourney.com/...39193335040013-Moodboards. Accessed 2026-05-24.
49. ^Midjourney Docs, "Getting Started", Midjourney Documentation, 2025. docs.midjourney.com. Accessed 2026-05-24.
50. ^Aituts, "Best Prompts for Midjourney V5", Aituts, 2023. aituts.com/midjourney-v5. Accessed 2026-05-24.
51. ^PetaPixel, "Midjourney v5 Creates Photorealistic Images and Even Does Hands Correctly", PetaPixel, 2023-03-17. petapixel.com/...ges-and-even-does-hands-correctly. Accessed 2026-05-24.
52. ^Midjourney, "Version 6.1", Midjourney Updates, 2024-07-30. updates.midjourney.com/version-6-1. Accessed 2026-05-24.
53. ^Midjourney Docs, "Legacy Features", Midjourney Documentation, 2025. docs.midjourney.com/...29788681101-Legacy-Features. Accessed 2026-05-24.
54. ^Nijijourney, "Welcome to Niji V6!", Nijijourney Blog, 2024-06-07. nijijourney.com/...niji-updates-welcome-to-niji-v6. Accessed 2026-05-24.
55. ^Midjourney, "Niji V7!", Midjourney Updates, 2026-01-09. updates.midjourney.com/niji-v7. Accessed 2026-05-24.
56. ^1 ^2 ^3Aitoolanalysis, "Midjourney Review 2025: V7 Just Launched + Video Generation", Aitoolanalysis, 2025. aitoolanalysis.com/midjourney-review. Accessed 2026-05-24.
57. ^Demand Sage, "Midjourney Statistics 2026 (Active Users & Revenue)", Demand Sage, 2026. demandsage.com/midjourney-statistics. Accessed 2026-05-24.
58. ^1 ^2Cooper, H., "Midjourney V8 is live on alpha: Here's what's actually different", Substack, 2026-03. heatherbcooper.substack.com/...live-on-alpha-heres. Accessed 2026-05-24.
59. ^1 ^2 ^3Roose, K. and Mac, R., "We Spoke to the Cartoonist Behind a Pointed Comic About AI Art", The Verge / The New York Times coverage, 2022-12. Accessed 2026-05-24.
60. ^The Art Newspaper, "Leaked: the names of more than 16,000 non-consenting artists allegedly used to train Midjourney's AI", The Art Newspaper, 2024-01-04. theartnewspaper.com/...used-to-train-midjourney-ai. Accessed 2026-05-24.
61. ^VentureBeat, "Midjourney founder says 'the world needs more imagination'", VentureBeat, 2022. venturebeat.com/...he-world-needs-more-imagination. Accessed 2026-05-24.
62. ^1 ^2U.S. Copyright Office Review Board, "Theatre D'opera Spatial Decision Letter", U.S. Copyright Office, 2023-09-05. copyright.gov/...Theatre-Dopera-Spatial.pdf. Accessed 2026-05-24.
63. ^Copyright Alliance, "Takeaways from the Andersen v. Stability AI Copyright Case", Copyright Alliance, 2023. copyrightalliance.org/...ability-ai-copyright-case. Accessed 2026-05-24.
64. ^NYU JIPEL, "Andersen v. Stability AI: The Landmark Case Unpacking the Copyright Risks of AI Image Generators", NYU Journal of Intellectual Property & Entertainment Law, 2024. jipel.law.nyu.edu/...-risks-of-ai-image-generators. Accessed 2026-05-24.
65. ^Crowell & Moring, "California Federal Court Trims Lawsuit Against Stability AI, Midjourney, and DeviantArt in Generative AI Artwork Case", Crowell & Moring LLP, 2024-08. crowell.com/...ntart-in-generative-ai-artwork-case. Accessed 2026-05-24.
66. ^1 ^2 ^3Allyn, B., "Disney and Universal sue AI firm Midjourney for copyright infringement", NPR, 2025-06-12. npr.org/...djourney-copyright-infringement-lawsuit. Accessed 2026-05-24.
67. ^1 ^2Georgetown Law Tech Institute, "Disney, NBC Universal, and DreamWorks File Major IP Lawsuit Against AI Image Generator Midjourney", Georgetown Law, 2025-06. law.georgetown.edu/...i-image-generator-midjourney. Accessed 2026-05-24.
68. ^Maddaus, G., "Warner Bros. Joins Studios' AI Copyright Battle Against Midjourney", Variety, 2025-09-04. variety.com/...ney-lawsuit-ai-copyright-1236508618. Accessed 2026-05-24.
69. ^IPWatchdog, "Warner Bros. Complaint Alleges Midjourney's Copyright Infringement is 'Systematic' and 'Willful'", IPWatchdog, 2025-09-08. ipwatchdog.com/...-infringement-systematic-willful. Accessed 2026-05-24.
70. ^1 ^2Mealey's, "Judge Consolidates Disney, Warner Bros. AI Suits Against Midjourney", LexisNexis Mealey's, 2025-11. mealeys.com/...er-bros-ai-suits-against-midjourney. Accessed 2026-05-24.
71. ^Cabina.AI, "Midjourney Banned Words: What Users Need to Know", Cabina.AI, 2025. cabina.ai/...-banned-words-what-users-need-to-know. Accessed 2026-05-24.
72. ^Midjourney Docs, "Community Guidelines", Midjourney Documentation, 2025. docs.midjourney.com/...484109-Community-Guidelines. Accessed 2026-05-24.
73. ^Midjourney Docs, "Personalization", Midjourney Documentation, 2025. docs.midjourney.com/...33330574221-Personalization. Accessed 2026-05-24.
74. ^DataCamp, "Midjourney V7: A Guide With 8 Practical Examples", DataCamp, 2025. datacamp.com/...midjourney-v7. Accessed 2026-05-24.
75. ^Fello AI, "Midjourney V8.1 Review: HD by Default, 5x Faster (2026)", Fello AI, 2026-04. felloai.com/midjourney-v8-1-review. Accessed 2026-05-24.
76. ^AIBase, "Big Move! Midjourney Confirms Entry into Hardware Field", AIBase News, 2025. news.aibase.com/...11355. Accessed 2026-05-24.
77. ^1 ^2 ^3Midjourney, "V8.1 Updates", Midjourney Updates, 2026-04-30. updates.midjourney.com/v8-1-updates. Accessed 2026-06-02.
78. ^Midjourney, "Relax mode for V8 Alpha", Midjourney Updates, 2026-03-21. updates.midjourney.com/relax-mode-for-v8-alpha. Accessed 2026-06-02.
79. ^Midjourney, "High-res rating", Midjourney Updates, 2026-04-27. updates.midjourney.com/high-res-rating. Accessed 2026-06-02.
80. ^Midjourney, "Web Updates", Midjourney Updates, 2026-05-27. updates.midjourney.com/web-updates-5. Accessed 2026-06-02.
81. ^1 ^2 ^3 ^4GetLatka, "Midjourney Revenue 2024: $200M (Bootstrapped)", GetLatka company profile, updated 2026-10-13. getlatka.com/...midjourney. Accessed 2026-06-20.
82. ^1 ^2 ^3Robb, D. / Allyn, B., "Disney and Universal v. Midjourney complaint (2:25-cv-05275, C.D. Cal.)", coverage via NPR and CNBC, 2025-06-11. cnbc.com/...disney-universal-midjourney-ai-copyright. Accessed 2026-06-20.
83. ^McKool Smith, "AI Infringement Case Updates: Disney/Warner Bros. v. Midjourney consolidation and schedule", McKool Smith AI Litigation Tracker, 2025-11-24. mckoolsmith.com/newsroom-ailitigation-46. Accessed 2026-06-20.
84. ^Midjourney, "V8 Alpha", Midjourney Updates, 2026-03-17 (quoted: "image generation is ~5x faster than before"; "text rendering works better than ever (when specified in 'quotes')"; HD/q4/sref/Moodboard jobs "4x as slow ... each costs 4x"). updates.midjourney.com/v8-alpha. Accessed 2026-06-20.
85. ^1 ^2 ^3 ^4 ^5Moon, M., "Midjourney, the AI image generator, is developing a full-body ultrasonic scanner", Engadget, 2026-06-18 (updated 2026-06-23). engadget.com/...urney-full-body-ultrasonic-scanner. Accessed 2026-06-27.
86. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12Midjourney, "A New Era of Midjourney", Midjourney Medical blog, 2026-06. midjourney.com/...blogpost. Accessed 2026-06-27.
87. ^1 ^2 ^3 ^4Butterfly Network, "Butterfly Network Provides Commentary on Midjourney Medical's Full Body Ultrasound Scanner Announcement", BusinessWire / Butterfly Network investor relations, 2026-06-18. ir.butterflynetwork.com/...default. Accessed 2026-06-27.
88. ^1 ^2 ^3iatrox, "Midjourney Medical Explained: What Is the 60-Second Full-Body Ultrasound Scanner? (2026)", iatrox, 2026-06. iatrox.com/...ll-body-ultrasound-scanner-explained. Accessed 2026-06-27.
89. ^Butterfly Network, "Form 8-K" (Item 1.01, Co-Development and Licensing Agreement with Midjourney, Inc. dated November 17, 2025), U.S. Securities and Exchange Commission, signed 2025-11-18. sec.gov/...bfly-20251117. Accessed 2026-10-27.
90. ^Fello AI, "Midjourney Scanner: AI Firm's Full-Body CT Device", Fello AI, 2026-06. felloai.com/midjourney-scanner. Accessed 2026-06-27.
91. ^1 ^2 ^3OnHealthcare, "How Midjourney's 60-second full-body ultrasound scanner became a LinkedIn fever dream", onhealthcare.tech, 2026-06. onhealthcare.tech/...djourneys-60-second-full-body. Accessed 2026-06-27.
92. ^1 ^2 ^3 ^4Midjourney, "Version 8.2", Midjourney Updates, 2026-07-24. updates.midjourney.com/version-8-2. Accessed 2026-10-27.
93. ^1 ^2 ^3 ^4Holz, D., "Midjourney's First Acquisition", Midjourney Updates, 2026-07-23. updates.midjourney.com/midjourneys-first-acquisition. Accessed 2026-10-27.
94. ^1 ^2 ^3Novak, M., "'We're at a Crazy Moment in History': Midjourney Buys Astrology App Co-Star", Gizmodo, 2026-07-24. gizmodo.com/...ys-astrology-app-co-star-2000790434. Accessed 2026-10-27.
95. ^1 ^2Midjourney, "Edit Model for V8", Midjourney Updates, 2026-08-27. updates.midjourney.com/edit-model-for-v8. Accessed 2026-10-27.
96. ^Midjourney, "Alpha Changelog - 9/23/26", Midjourney Updates, 2026-09-24. updates.midjourney.com/alpha-changelog-9-23-26. Accessed 2026-10-27.
97. ^1 ^2Midjourney, "Alpha Changelog - 10/7/26", Midjourney Updates, 2026-10-08. updates.midjourney.com/alpha-changelog-10-7-26. Accessed 2026-10-27.
98. ^Midjourney, "Testing Thinking Mode", Midjourney Updates, 2026-10-08. updates.midjourney.com/testing-thinking-mode. Accessed 2026-10-27.
99. ^Midjourney, "Changelog - 8/20/26", Midjourney Updates, 2026-08-21. updates.midjourney.com/changelog-8-20-26. Accessed 2026-10-27.
100. ^1 ^2Kanter, J., "Midjourney: Report Claims AI Model Is Worst Offender For Flouting Hollywood Copyright", Deadline, 2026-10-14. deadline.com/...ing-hollywood-copyright-1237154802. Accessed 2026-10-27.
101. ^Maddaus, G., "Midjourney Seeks to Reveal Studios' Use of AI in High-Stakes Copyright Battle", Variety, 2026-07-02. variety.com/...s-ai-copyright-discovery-1236800902. Accessed 2026-10-27.
102. ^McKool Smith, "AI Infringement Case Updates: January 19, 2026", McKool Smith, 2026-01-19. mckoolsmith.com/newsroom-ailitigation-51. Accessed 2026-10-27.
103. ^1 ^2Holz, D. (@DavidSHolz), post on X, 2025-08-22 ("We remain an independent, community-backed research lab, with no investors..."). x.com/...1958984943566299632. Accessed 2026-10-27.
104. ^1 ^2 ^3 ^4 ^5 ^6Claburn, T., "Holz, founder of AI art service Midjourney, on future images", The Register, 2022-08-01. theregister.com/...david_holz_midjourney. Accessed 2026-10-27.
105. ^1 ^2 ^3 ^4Vincent, J., "'An engine for the imagination': an interview with David Holz, CEO of AI image generator Midjourney", The Verge, 2022-08-02. theverge.com/...ey-multiverse-interview-david-holz. Accessed 2026-10-27.
106. ^Riley, D., "Leap Motion sold for fire-sale price of $30M to Ultrahaptics", SiliconANGLE, 2019-05-30. siliconangle.com/...re-sale-price-30m-ultrahaptics. Accessed 2026-10-27.
107. ^1 ^2 ^3TechCrunch, "Midjourney says it's 'getting into hardware'", TechCrunch, 2024-08-28. techcrunch.com/...y-says-its-getting-into-hardware. Accessed 2026-10-27.
108. ^1 ^2 ^3 ^4Welch, C., "AI Startup Midjourney Pivots to Health With Ultrasound Machine", Bloomberg Law, 2026-06-18. news.bloomberglaw.com/...h-with-ultrasound-machine. Accessed 2026-10-27.
109. ^1 ^2 ^3Constantin, A. M., "Midjourney says its new full-body scanner beats an MRI. It has never built hardware before.", The Next Web, 2026-06-18. thenextweb.com/...er-midjourney-medical-ultrasound. Accessed 2026-10-27.
110. ^1 ^2 ^3Vigliarolo, B., "Midjourney pivots from AI image generation to body scanning medical spa where patients bathe in 'golden light'", The Register, 2026-06-18. theregister.com/...5258429. Accessed 2026-10-27.
111. ^RTTNews, "Butterfly Network Stock Jumps 33% On Midjourney Medical Partnership Update", RTTNews, 2026-06-18. rttnews.com/...djourney-medical-partnership-update. Accessed 2026-10-27.
112. ^Butterfly Network, "Butterfly Network Reports Second Quarter 2026 Financial Results", Exhibit 99.1 to Form 8-K, U.S. Securities and Exchange Commission, 2026-07-30. sec.gov/...bfly-20260730xexx991. Accessed 2026-10-27.
113. ^1 ^2Midjourney, "Hardware announcement livestream", Midjourney Updates, 2026-06-16. updates.midjourney.com/...-announcement-livestream. Accessed 2026-10-27.

Show all 113 references