He Built a $75K/Month App in 3 Hours (Pieter Levels AI Coding Breakdown)

3 hours. That's how long it took Pieter Levels to build a video game that now makes $75,000 a month.

[Beat]

No team. No investors. No game development experience. Just Cursor AI and a weekend afternoon.

And this isn't even his biggest product. This man makes $3 million a year with zero employees.

Let me break down exactly how he does it.

If you don't know Pieter Levels, here's the quick version:

Dutch programmer. Self-taught - didn't start coding until his late twenties. Digital nomad who's built over 40 projects.

His most famous projects: Nomad List - a community for digital nomads - and Remote OK - one of the largest remote job boards in the world.

But what makes him legendary in indie hacker circles isn't just his success. It's his philosophy.

His competitive advantage? Shipping before things are ready. Launching fast. Iterating based on real feedback, not assumptions.

And now, AI has supercharged that approach.

In early 2026, Pieter decided to build a flight simulator. Not because he had game development experience - he didn't. But because AI coding tools had reached a point where he could try.

He used Cursor - the AI-powered code editor that's taken over the developer world. Built on VS Code, powered by Claude and GPT-4.

3 hours of prompting and iteration. That's it.

The result? A low-polygon flight simulator that immediately went viral. Developers everywhere started building similar games, inspired by what he'd proven possible.

Revenue timeline:

Launch: Viral on Twitter/X

Days later: $57K MRR

March 2026: $75K MRR

Trajectory: Heading toward $100K MRR

A game built in an afternoon, making more per month than most senior developer salaries.

But here's what most people miss about Pieter's strategy. The flight simulator isn't his main business. It's part of a PORTFOLIO.

Photo AI - $138K/month: His biggest earner. AI-powered photo generation. Takes up 70% of his total revenue.

Interior AI - $45K/month: AI that redesigns room interiors. Here's the wild part: it runs with 99% profit margins. Why? Because it's 100% operated by AI - no human employees.

He said publicly: 'Interior AI now has >99% profit margins. GPU bill is tiny. It's 100% ran by AI robots, no people.'

Remote OK, Nomad List, and others: The rest of his portfolio fills in the gaps. Some built years ago, still generating passive income.

Combined: $250,000+ per month. $3 million per year. Zero employees. One laptop.

Now here's something that will surprise developers:

Pieter's tech stack is deliberately simple. Vanilla PHP. jQuery. SQLite.

[Beat]

No React. No microservices. No Kubernetes. No 47-step CI/CD pipeline.

Why? Because complexity is the enemy of shipping fast.

His approach with AI:

1. Use Cursor for rapid development

2. Ship ugly, ship fast

3. Let users tell you what to fix

4. Automate everything that can be automated

5. Keep the tech stack simple enough that AI can understand it

The last point is crucial. AI works better with straightforward code. The more complex your architecture, the more AI struggles to help.

Here's what Pieter represents that most people miss:

It's not about the tools. It's about the APPROACH.

Old Approach: Plan for months, Build the perfect product, Launch when ready, Hope people want it

Pieter's Approach: Find an interesting idea, Build an MVP in hours or days, Ship immediately, Iterate based on real usage

AI doesn't change WHAT to build. It changes HOW FAST you can test ideas.

Most founders fail because they spend 6 months building something nobody wants. Pieter can test an idea in a weekend. If it doesn't work? Move on. If it does? Double down.

AI gave him a 10x speed multiplier on a strategy that was already working.

Now, Pieter has said something that made a lot of indie hackers uncomfortable:

'Indie Hacking is dead.'

What he means isn't that solo founders can't succeed. Obviously they can - he's proof.

What he means is: the OLD way of indie hacking - grinding for months on a product, doing everything manually - that's dead.

The new way is AI-augmented, fast-shipping, portfolio-based. Build more. Ship faster. Let winners emerge.

And if you're still building the old way? You're competing with people like Pieter who can test 10 ideas in the time it takes you to build one.

Pieter Levels is one of dozens of documented success stories we've collected at End of Coding.

Solo founders. Non-technical entrepreneurs. Career changers. All using AI tools to build real businesses with real revenue.

We've got the tool comparisons, the tutorials, and the documented case studies - including how to get started with Cursor, the tool Pieter used.

Link in the description.

The question isn't whether you CAN build a $75K/month product with AI.

The question is: what idea are you going to test this weekend?