Midjourney Revenue: How It Grew to $300M With No VC Funding

Midjourney did something unusual when it launched its AI image generator.

It did not start with a polished website.

It put the product inside Discord.

A user would enter a channel, type /imagine, and describe an image. While their image was being generated, hundreds of other prompts and images moved through the same channel.

It looked chaotic.

But that chaos became one of Midjourney's biggest growth advantages.

New users could see what other people were making. They could read their prompts, copy ideas, learn what worked, and immediately try again.

Every person using Midjourney was also showing other people how to use Midjourney.

That helped turn a small research lab into one of the biggest AI image companies in the world. Forbes, citing PitchBook, reports that Midjourney generated $300 million in revenue in 2024. The company says it remains self-funded and today has a team of about 60 people.

The interesting part is how it got there.

How Midjourney grew

Midjourney grew by turning product usage into distribution.

It launched inside Discord, where prompts and generated images were public by default. This made it easy for new users to learn from experienced users.

The best images then escaped Discord and spread across Reddit, X, Instagram, Facebook, and the rest of the internet. People saw an image, asked how it was made, discovered Midjourney, and started creating images of their own.

Midjourney also used invite-only access to seed its early community, expanded its bot into millions of other Discord servers, used community feedback to improve its models, and charged users early enough to cover its expensive GPU costs.

The result was a powerful loop:

public creation → people learn → better images → images get shared → new users join → more feedback → better product

Midjourney made every part of that loop stronger over time.

Midjourney growth in numbers

Discord members should not be confused with active users or paying customers. Midjourney has not publicly disclosed a comparable number of paid subscribers.

The revenue number also deserves careful wording. The strongest public figure is the reported $300 million in revenue for 2024. Higher estimates for later years have circulated online, but Midjourney has not confirmed them.

1. Midjourney chose Discord instead of a normal app

Before Midjourney became a Discord product, founder David Holz tested more conventional ways to use it.

There was a problem.

Someone would type something simple like "dog," get an image of a dog, and stop.

They did not know what else the model could do.

Image generation was new. Most people had never written prompts before. Giving someone an empty text box and asking them to be creative created a blank-page problem.

The experience changed when people used the model together.

One person might ask for a dog.

Another person would see it and try a dog in space.

Someone else would turn that into an Aztec space dog.

Each result gave the next person another idea.

Holz later explained the thinking simply: people wanted to make things together.

Midjourney's team also tested where its early users wanted to interact and found that AI enthusiasts already liked Discord.

So Discord became more than a place to access the bot.

It became part of the product.

2. Public prompts solved the onboarding problem

Most software products teach new users with onboarding screens, tutorials, templates, and help docs.

Midjourney had something better.

Other users.

Prompts and generated images appeared together in public channels. A new user could scroll for a few minutes and quickly understand what was possible.

They could see:

• what people were asking for
• which words people added to prompts
• what the model produced
• which images people chose to upscale
• how people changed prompts after a bad result
• which creations received reactions

Learning happened by watching.

Midjourney also generated images gradually instead of showing a blank loading screen until the result was finished.

While users waited for their own image, they watched other images appear.

The wait became entertainment.

Discord channels for themes and featured creations gave people more examples. Mentions brought users back when their image finished. Buttons under images made it easy to create variations or improve a result.

The whole room acted like a live tutorial.

This lowered the amount of work Midjourney had to do to explain the product.

3. Every generated image could become marketing

Midjourney had another advantage: its output was naturally shareable.

People did not need to write a long post explaining why the product was interesting.

They could simply post the image.

That created a growth loop outside Discord:

1. Someone created an interesting image with Midjourney.
2. They posted it on Reddit, X, Instagram, or another platform.
3. Other people saw it and asked how it was made.
4. They discovered Midjourney.
5. They joined and made images of their own.
6. Some of those images were shared again.

The product created its own marketing assets.

And as Midjourney's models improved, those assets became more impressive.

A strange, funny, beautiful, or realistic image could reach millions of people without Midjourney buying the distribution.

This became especially clear in 2022 and 2023.

Jason Allen used Midjourney to create an image that won a digital-art prize at the Colorado State Fair in 2022. The result started a large debate about AI and art.

In March 2023, an AI-generated image of Pope Francis wearing a large white puffer jacket spread across the internet. Many people initially thought it was real.

Other Midjourney images showed fake scenes involving politicians, celebrities, movies, architecture, fashion, and almost anything else people could imagine.

Some of this attention created serious moderation and misinformation problems.

But from a distribution point of view, every viral image introduced more people to AI image generation.

Midjourney was often the product they discovered.

4. Better images made the growth loop stronger

Distribution alone would not have made Midjourney this large.

The output had to be worth sharing.

Midjourney became known early for images that felt polished and visually interesting. As new versions launched, image quality improved quickly.

Better models made it easier for ordinary users to create something impressive.

That mattered because the product's output was also its marketing.

If generated images were boring, the sharing loop stopped.

When the images became better, the loop accelerated:

better model → better images → more sharing → more users → more usage and feedback → better model

Midjourney's product quality and distribution were tied together.

The company is still using this system today. Community ratings continue to help Midjourney improve areas such as image ranking, recommendations, model personalization, and style generation.

5. Invite-only access helped seed the community

Early Midjourney access was limited.

The company started with a small private beta and a waitlist.

At one point, paid users could invite five other people.

That did two useful things.

First, it created demand.

People were asking for invitations in online communities because they wanted to try the new tool.

Second, it controlled growth.

AI image generation was expensive. Midjourney needed GPUs for every generation. Letting unlimited numbers of free users arrive at once could have hurt the product for everyone.

Invitations gave the company more control while the model and infrastructure were still changing.

There was also a social advantage.

People often joined because someone they knew invited them. That person could explain how prompting worked and help them through the strange Discord interface.

Midjourney eventually made its Discord server publicly joinable in July 2022.

According to Discord's case study, the server reached its previous cap of one million members within three months of private testing.

The early invite system had already created the base for a much larger community.

6. Midjourney moved into other people's Discord servers

Midjourney's distribution expanded again when people could add the bot to their own Discord servers.

Now someone did not have to join Midjourney's enormous official community to discover the product.

The bot could appear inside a gaming community, art server, startup team, friend group, or private community they already used.

This changed Midjourney from a destination into infrastructure inside Discord.

Discord says the Midjourney app has now been installed in around seven million servers.

That created another distribution loop.

A single Midjourney user could add the bot to a community and expose many more people to it.

Those people could start generating images without leaving the Discord environment they already understood.

It also solved part of the chaos problem.

The official Midjourney server moved extremely fast. In a private server, a small group could generate images without losing their work in a constant stream of prompts.

7. The community helped Midjourney improve the product

Midjourney's users did more than market the product.

They helped improve it.

David Holz said that one of Midjourney's early model improvements came from learning which images users preferred and how they interacted with the system.

Midjourney later made rating images a formal product feature.

Users could choose between images they preferred. That data helped Midjourney improve its algorithms, fill its Explore page, and build personalization systems.

At different points, Midjourney even rewarded its most active image raters with free GPU time.

The company also used:

• feedback polls
• founder office hours
• alpha tests
• community support
• feature discussions
• prompt-sharing channels
• ratings
• likes and preferences

A small team suddenly had a huge group of people showing it what they liked.

That feedback loop became more valuable as the community grew.

More users meant more generations.

More generations created more examples.

More examples created more preference data.

And more preference data helped Midjourney improve what people saw and generated.

The community was helping with both distribution and product development.

8. Midjourney charged users early

Many software companies grow a large free user base and worry about monetization later.

Midjourney had less room to do that.

Generating AI images costs money.

Every prompt uses computing power.

Holz described the early business model in simple terms: if people enjoyed using Midjourney, they needed to pay enough to cover the cost of generating the images and leave some money for the company.

Early subscriptions started around $10 and $30.

Midjourney was already profitable by August 2022, according to Holz, only weeks after the public beta.

The company had roughly ten employees at the time.

Today, Midjourney offers four main subscription levels, from $10 to $120 per month.

The plans charge more for things heavy users care about: more fast GPU time, more generation capacity, and privacy.

This meant Midjourney did not need to choose between growth and revenue for years.

People were paying while the product was still spreading.

Subscriptions helped pay for the infrastructure required to serve the next wave of users.

9. Midjourney even turned privacy into a paid feature

Public creation helped Midjourney grow.

But not everyone wanted their work public.

A designer working on a client project may not want competitors seeing prompts and concepts. A company exploring a new campaign may want the work kept private.

Midjourney's higher Pro and Mega plans include Stealth Mode.

This creates an interesting pricing model.

Public creation helps the network because more work can be discovered, studied, and shared.

Users who need more privacy can pay more for it.

The same behavior that helped the product grow also created a reason for professional users to upgrade.

There is a tradeoff.

A public-by-default system can make users uncomfortable. Midjourney has faced complaints about prompt copying, public images, client privacy, and confusion about which creations are visible.

But the growth logic is clear.

Visible work created more opportunities for discovery.

10. Midjourney stopped free trials when it no longer needed them

Free generations were useful early on.

They let curious people experience AI image generation without paying.

But by March 2023, Midjourney stopped its free trials.

The timing led some people to assume the change was caused by viral fake images of Donald Trump and Pope Francis.

Holz gave a more practical explanation.

Demand was becoming too high.

People were creating disposable accounts to get more free generations. A viral tutorial in China brought another wave of users. GPU availability was becoming a problem.

Free usage was starting to hurt the experience for people who paid.

By then, Midjourney had already proved that people wanted the product.

The company no longer needed unlimited free use to create demand.

It needed to protect the product for customers willing to pay.

That is an important shift in Midjourney's growth story.

Early on, free access helped create the network.

Later, the network itself created enough demand.

Midjourney's growth flywheel

The easiest way to understand Midjourney's growth is as a loop.

1. People create in public

Prompts and outputs are visible to other users.

↓

2. New users learn by watching

They do not have to start from an empty text box. They can copy patterns and ideas that already work.

↓

3. Users create better images

Better prompts and better models produce more interesting results.

↓

4. The best images spread outside Midjourney

People share them on social platforms, communities, and websites.

↓

5. New people discover Midjourney

They want to know how the image was made and try the product themselves.

↓

6. More usage creates more feedback

Prompts, ratings, likes, and user behavior give Midjourney more information about what people want.

↓

7. Midjourney improves the product

Better models produce better results.

↓

8. Better results give people more to share

And the loop starts again.

The important idea is simple:

Midjourney made product usage visible.

A user's work could teach the next user, attract the next user, and help improve the product for the next user.

Midjourney's growth timeline

Why Midjourney eventually moved beyond Discord

The Discord strategy worked extremely well.

Eventually, its success created a problem.

The official server became enormous.

A new user could enter a channel and watch their generation disappear up the screen beneath hundreds of new messages.

People lost images.

Professional users moved to private servers or direct messages.

Other potential customers simply did not want to use Discord.

Midjourney began building its own web experience.

At first, access was limited to users who had generated large numbers of images. The threshold gradually fell.

On August 23, 2024, Midjourney opened web creation to everyone.

But Midjourney did not throw away everything that made Discord work.

Its web product kept many of the same ideas:

• public images
• Explore
• ratings
• profiles
• personalization
• moodboards
• community feedback

The interface changed.

The idea of learning from visible creation stayed.

That matters because Discord itself was never the entire strategy.

The deeper strategy was making creation social and observable.

Discord happened to be an excellent place to start.

The downsides of Midjourney's growth model

The same decisions that helped Midjourney grow created problems later.

Discord became overwhelming

A public feed works when a community is small.

With millions of people, it becomes hard to follow.

The web app helped Midjourney give users a calmer workspace.

Public creation created privacy concerns

Some people do not want prompts, client projects, reference images, or experiments appearing publicly.

Midjourney added paid privacy options, but public-by-default creation remains a tradeoff.

Moderation became expensive

A huge public creative community needs rules.

Midjourney has had to deal with spam, sexual content, harassment, political images, deepfakes, and requests involving public figures.

As AI images became more realistic, those problems became harder.

Viral output created legal exposure

The same cultural references that helped Midjourney images travel also attracted attention from artists and rights holders.

Disney and Universal sued Midjourney for copyright infringement in 2025. Midjourney has disputed claims against its practices and argued for fair use.

The cases and wider legal questions around generative AI remain important risks.

Growth made Midjourney more visible to customers.

It also made the company more visible to critics, regulators, and rights holders.

What startups can learn from Midjourney

The lesson from Midjourney is not "start a Discord."

Discord worked because it matched the way people used the product.

The more useful lesson is to ask:

Can using your product help another person understand why the product is valuable?

For Midjourney, the answer was yes.

One person's prompt showed another person what to try.

One person's image became another person's inspiration.

One person's social post became another person's first exposure to Midjourney.

The best products can create some version of this loop.

Figma does it when people collaborate on the same design.

GitHub does it when public code helps developers discover projects.

Canva does it when people share designs made with Canva.

Midjourney took the idea unusually far because almost every use of the product produced something visual that people could immediately understand and share.

There is another lesson too.

Product-led distribution only works when the product output is good.

Midjourney could make creation public, build a giant Discord server, and add invitation loops.

None of that would matter for long if the images were bad.

The quality of the product created the fuel.

The community multiplied it.

Most products can't make usage public the way Midjourney did. They can still show up where their users share work and ask for help. Okara's Reddit Agent finds those threads and drafts replies you post from your own account, and the UGC Videos Agent turns your product into short creator-style clips.

The real reason Midjourney grew

It would be easy to describe Midjourney's growth as a Discord success story.

Discord was important.

But the bigger idea was making product use observable.

Midjourney put new users in a room full of people already using the product.

They could watch experts.

They could copy good prompts.

They could see results before spending time learning the tool.

They could make something worth sharing within minutes.

And when they shared it, the output advertised Midjourney without looking like an advertisement.

As more people joined, they generated more images, prompts, ratings, ideas, and feedback.

That helped Midjourney improve the product.

A better product then created better things to share.

Midjourney turned creation into distribution.

That is the growth loop that took it from an unusual Discord bot to a company that reportedly generated $300 million in annual revenue.

More startup growth stories

• Higgsfield Revenue (2026): How It Grew to a $1B Run Rate
• HeyGen Revenue: How It Grew From a $5 Fiverr Order to $200M
• ElevenLabs ARR: How ElevenLabs Grew to $600M (2026)
• All startup growth case studies

Frequently asked questions

How did Midjourney grow so fast?

Midjourney grew by launching inside Discord and making prompts and generated images visible to other users. People learned from one another, shared their best images outside Discord, invited friends, and brought the Midjourney bot into other communities. The company also improved its models quickly and charged users early enough to fund its expensive GPU usage.

Why did Midjourney use Discord?

Midjourney found that creating images was more engaging when people could see what others were making. Discord let users generate images together, share prompts, react to results, help each other, and learn from examples in real time.

How did Discord help Midjourney grow?

Discord turned Midjourney into a social product. Public prompts made onboarding easier, public images became product demos, invitations helped seed the community, and the Midjourney bot could later spread into other Discord servers. Discord says Midjourney has been installed in around seven million servers.

How much revenue does Midjourney make?

Forbes, citing PitchBook, reports that Midjourney generated $300 million in revenue in 2024. Higher estimates for later years have circulated online, but Midjourney has not publicly confirmed them.

Is Midjourney profitable?

Yes. David Holz said Midjourney was already profitable in August 2022. Forbes also describes the company as profitable in its current company profile.

How many users does Midjourney have?

Midjourney's Discord community has accumulated tens of millions of members, but Discord members should not be treated as the same thing as active users or paying customers. Midjourney has not publicly disclosed a comparable total for active or paid users.

Did Midjourney raise venture capital?

Midjourney says it is self-funded and community-funded. It has no disclosed equity investors or traditional venture funding round.

What is Midjourney's business model?

Midjourney makes most of its money through subscriptions. Its current plans range from $10 to $120 per month. Higher plans include more GPU capacity and features such as Stealth Mode.

Why did Midjourney stop its free trial?

Midjourney stopped free trials in March 2023 after free usage began putting pressure on its GPU capacity. Holz said disposable accounts and a sudden wave of demand were affecting the experience for paying customers.

Sources

• Discord — How Midjourney Built a Business on Discord
• Discord Engineering — Maxjourney: Pushing Discord's Limits with a Million+ Online Users in a Single Server
• The Verge — David Holz interview on Midjourney and Discord
• The Register — David Holz on Midjourney's business model and early profitability
• Midjourney — The web is open to all!
• Midjourney — Comparing Midjourney Plans
• Midjourney — company About page
• Midjourney — updates on image rating and personalization
• Forbes — Midjourney company profile and reported 2024 revenue
• Reuters — Disney and Universal copyright lawsuit against Midjourney