Every venture capitalist in Silicon Valley begged David Holz to take their money. He said no. Not once, not in some dramatic press conference moment, but repeatedly, quietly, over years. While his competitors raised billions in venture capital, burned through it building massive organizations, and diluted their founders down to single-digit ownership stakes, Holz just kept building. Midjourney now generates roughly $500 million a year in revenue with about forty employees and zero outside funding. That is $12.5 million in revenue per person. And it breaks every rule in the startup playbook.
I have been following the AI industry closely for years, and the Midjourney story still catches me off guard every time I look at the numbers. Not because a bootstrapped company can be successful — that is common enough in software. But because a bootstrapped company can be THIS successful in the most capital-intensive sector in technology, competing directly against companies that have raised tens of billions of dollars, and winning on the metrics that actually matter.
The Numbers That Should Not Exist
Let me put Midjourney's efficiency in context, because the comparison is almost absurd.
OpenAI employs over 3,000 people and generates approximately $13 billion in annual revenue. That works out to about $4.3 million per employee. Anthropic has more than 1,200 employees generating roughly $5 billion, which is about $4.2 million per employee. Both of those are impressive numbers by any standard. Most technology companies would be thrilled with that kind of revenue productivity.
Midjourney generates $12.5 million per employee. Three times more than OpenAI. Three times more than Anthropic. With zero outside capital.
The estimated profit margin is between 70 and 80 percent, which means Midjourney is likely clearing $350 million to $400 million in annual profit. That profit funds all of their research and development internally. They do not need to raise money because they are already making more money than most funded AI companies ever will.
And the subscriber base is enormous: over 16 million paying users at an average of roughly $31 per month. Not free users. Not people on a trial. Paying subscribers who hand over their credit card information month after month because the product delivers value they cannot get anywhere else.
How a Forty-Person Company Outearns Thousands
The question everyone asks when they hear these numbers is: how? How does a company with forty people generate half a billion dollars in annual revenue? The answer is a combination of radical simplicity, a product that markets itself, and a founder who understood something about business that most of Silicon Valley has forgotten.
Start with the product. Midjourney runs on Discord. Not a custom-built enterprise platform. Not a sleek web application with dozens of features. Discord. Users type a text prompt, Midjourney generates images, and those images appear in a Discord channel. The interface could not be simpler. And that simplicity is not a limitation — it is the entire strategy.
The pricing is equally simple. Four tiers. Basic at $10 per month gives you 200 image generations. Standard at $30 per month gives you fifteen hours of fast generation. Pro at $60 per month gives you thirty hours plus stealth mode. Mega at $120 per month gives you sixty hours. No free tier. No freemium bait-and-switch. No ad-supported model. No enterprise sales team negotiating custom contracts over months-long procurement cycles.
Every single customer pays the same way. That means Midjourney needs zero salespeople. Zero account managers. Zero solutions engineers. Zero customer success representatives doing quarterly business reviews. The entire revenue engine is self-service, which is why forty people can handle sixteen million paying subscribers.
Why David Holz Said No to Venture Capital
This is the part of the story that fascinates me, because it goes against everything the startup ecosystem tells founders they should do. The conventional wisdom is clear: raise as much money as possible, as fast as possible, from the most prestigious investors possible. Growth at all costs. Market share first, profitability later. Winner take all.
Holz looked at that playbook and saw something different. He saw dilution. He saw loss of control. He saw pressure to grow headcount before growing revenue. He saw board meetings and investor updates and governance obligations that would pull his attention away from the product. And he decided that none of that was necessary.
The result? David Holz owns the entire company. At a conservative 10x revenue multiple — which is actually below the norm for high-growth AI companies — Midjourney is worth at least $5 billion. Holz did not give up a single percentage point. In an industry where founders routinely raise billions and end up owning 5 to 10 percent of their own company, that is a staggering outcome.
But it is not just about the money. Ownership means control. Holz can make decisions about the product, the team, the strategy, and the pace of growth without needing approval from a board of directors or managing the expectations of investors who need a specific return by a specific date. That freedom shows up in everything about how Midjourney operates.
They stayed small on purpose. Not because they could not afford to hire, but because Holz recognized that adding people does not automatically add value. Every person at Midjourney is essential. There is no middle management. No bureaucracy. No department heads managing department heads. When you have forty people generating half a billion dollars, everyone is doing work that directly creates value. That is an incredibly rare thing in a company of any size.
The Product IS the Marketing
Midjourney spends effectively zero on marketing. No paid ads. No content marketing team. No social media managers. No influencer partnerships. No growth hackers running conversion optimization experiments. The marketing budget is, as far as anyone can tell, approximately zero.
How do sixteen million people find and pay for a product with no marketing? Because every image Midjourney creates is an advertisement. When someone shares a stunning AI-generated image on social media, in a presentation, in a pitch deck, in a blog post — that image sells Midjourney to everyone who sees it. The distinctive aesthetic quality of Midjourney images has become recognizable enough that the product markets itself through its own output.
This is something that most AI companies fundamentally misunderstand about go-to-market strategy. They think they need a sales team, a marketing team, a content team, a growth team. Midjourney proves that if the product is good enough, and if the output of the product is inherently shareable, the traditional marketing machine is unnecessary. The product IS the marketing.
I have watched startups raise $50 million and spend $30 million of it on sales and marketing, and I have watched Midjourney spend nothing and grow faster. The lesson is not that marketing never matters. The lesson is that product quality is the most efficient customer acquisition channel ever invented, and most companies invest in marketing because their product is not good enough to sell itself.
What This Tells Us About AI Business Models
The Midjourney story challenges the dominant narrative about AI business models in ways that the industry still has not fully absorbed.
The dominant narrative says that AI requires massive capital investment. Foundation models cost hundreds of millions to train. Compute costs are astronomical. You need billions in funding just to stay competitive. That narrative is true for companies competing at the frontier of model development. It is completely false for companies that apply AI to specific problems and charge customers for the value they deliver.
Midjourney is not trying to build artificial general intelligence. They are not competing with OpenAI on reasoning capabilities or with Google on search. They are generating beautiful images for people who want beautiful images, and they charge a reasonable price for it. The scope of their ambition is precisely calibrated to what they can deliver profitably with a small team.
The dominant narrative also says that consumers will not pay for AI. Enterprise contracts are where the money is. You need a sales team calling on Fortune 500 companies and negotiating six-figure annual deals. Midjourney proved that this is completely wrong. Sixteen million individuals happily pay $10 to $120 per month for an AI tool that makes them more creative. The consumer AI market is enormous, and it is being systematically underestimated by VCs who are fixated on enterprise SaaS metrics.
And the dominant narrative says that you need to move fast and break things, hire aggressively, and scale at all costs. Midjourney moved deliberately, hired carefully, and scaled revenue instead of headcount. They proved that discipline is a competitive advantage, not a limitation.
The Profitability Gap
Here is something that really stands out when you compare Midjourney to its funded competitors: profitability.
OpenAI reportedly lost $5 billion in 2024 despite generating $6 billion in revenue. They are spending faster than they are earning, funding the gap with capital from investors who believe the losses are investments in future dominance. Anthropic burns capital at staggering rates, investing heavily in model development and safety research. Stability AI nearly ran out of money entirely before scrambling for emergency funding.
Midjourney has been profitable since approximately day one. Their estimated $350 million to $400 million in annual profit is not a goal or a projection. It is cash that has already been generated and that funds the company's continued development without any external dependency. If the venture capital market collapses tomorrow, if interest rates spike, if investor enthusiasm for AI evaporates — none of that affects Midjourney. They make more money than they spend. That is a form of resilience that no amount of venture capital can buy.
The Lessons That Apply Beyond Midjourney
I want to be careful here, because the wrong takeaway from this story is that every AI startup should bootstrap. That is not the lesson. Some AI companies genuinely need hundreds of millions of dollars to build their technology. If you are training frontier models or building autonomous vehicles, you cannot self-fund that research from subscription revenue.
The real lessons are more nuanced. First, venture capital is a tool, not a requirement. The fact that the most profitable company in the entire AI industry raised zero dollars should permanently end the assumption that raising money is necessary for building a valuable AI business. It is one path. It is not the only path.
Second, simplicity scales. Four pricing tiers. One platform. No enterprise sales. No free tier. That simplicity is not a weakness. It is what allows forty people to serve sixteen million customers. Every layer of complexity you add to your business requires people to manage it, and people are the most expensive thing in any organization.
Third, quality is the most efficient growth engine. If your product is remarkable enough that every use of it becomes an advertisement, you do not need a marketing budget. Invest in making the product extraordinary, and the customers will come.
Fourth, staying small is a legitimate strategy. The startup ecosystem treats headcount growth as a proxy for success. Midjourney demonstrates that revenue per employee is a far better metric than total employees. A company of forty people generating half a billion dollars is more impressive and more durable than a company of four thousand people generating $2 billion while losing money.
David Holz built something that most people in Silicon Valley would say is impossible. A bootstrapped AI company that generates more profit than most funded AI companies generate in revenue. The numbers are right there, and they are hard to argue with. $500 million. Forty people. Zero funding. That is what an efficiently run AI company actually looks like.
---
Keep Reading
- OpenAI's $13B ARR: The Fastest-Growing Company in History — the opposite end of the spectrum: the most-funded AI company and its revenue machine.
- Cursor Hit $9B Valuation — another AI startup with remarkable capital efficiency and explosive growth.
- How to Make Money with Claude AI — if Midjourney inspired you, here are practical ways to build income with AI tools.
- AI Kills Middle-Class Jobs — or Does It? — the workforce implications of companies that do more with fewer people.
