How Solo Founders Make Money With AI
Most "make money with AI" advice is noise: a screenshot of a dashboard, a vague promise, and a course at the end. This is the opposite. Below are five real builders whose numbers have been reported by reputable outlets or shared openly by the founders themselves. Each one is broken down as a playbook, meaning the actual mechanism that turned AI into revenue, plus the lesson you can copy. The point is not to admire these people. It is to steal the repeatable part.
If you are a solo founder or thinking about your first AI side hustle, the question is never "can AI make money." It clearly can. The real question is which mechanism fits your skills, your audience, and your tolerance for risk. These five answer that in five different ways.
Key takeaways
- Distribution beats the model. Every builder here owned an audience or a search channel before the AI product paid off. The AI was the easy part.
- Charge money on day one. Each playbook sells a paid outcome (a headshot, an app, a tool, a template), not "access to AI."
- Small and profitable can be a destination, not a stepping stone. Several of these founders deliberately stayed solo.
- Speed is a moat for individuals. Shipping a working product in days, in public, generated the attention that paid the bills.
- Pick one money mechanism and go deep, rather than copying five at once.
The five playbooks at a glance
| Builder | What they built | Money mechanism | Lesson |
|---|---|---|---|
| Pieter Levels | A portfolio of small AI and web products (PhotoAI, RemoteOK, Nomad List) | Subscriptions and listings across many products, marketed by building in public | A portfolio of small bets beats one big swing |
| Danny Postma | HeadshotPro, an AI headshot generator | Paid one-off purchases plus affiliate revenue, fed by programmatic SEO | Own a search channel and the traffic compounds for free |
| Maor Shlomo | Base44, a prompt-to-app builder | Usage-based subscriptions, then an acquisition | A profitable solo product can become an exit fast |
| Marc Lou | A portfolio of developer tools and courses (ShipFast, CodeFast, DataFast) | One-time license sales and SaaS subscriptions across many products | Ship constantly, keep the winners, kill the rest |
| Nick Dobos | BoredHumans, a single site hosting 100+ free AI tools | Ad-supported free tools at high traffic volume | Volume and free access can be a business model of their own |
Playbook 1: Pieter Levels and the portfolio of tiny bets
Pieter Levels (known online as levelsio) is the closest thing the indie world has to a patron saint. He runs a stack of small products by himself, with no co-founders and no outside funding, and the combined revenue lands around $3 million a year. His biggest single product is PhotoAI, which has been documented passing roughly $130,000 in monthly recurring revenue, alongside older properties like RemoteOK and Nomad List.
The detail people miss is the failure rate. In his long conversation on the Lex Fridman podcast, Levels describes building a huge number of products before any of them stuck. PhotoAI was not luck. It was the survivor of a deliberate process: launch fast, charge immediately, and kill anything that does not find customers within a few weeks.
The mechanism
The money comes from monthly subscriptions and listing fees spread across many products, so no single one has to carry the whole business. AI lowered the cost of building each new product to almost nothing, which means he can run more experiments per year than a funded team. The marketing engine is "build in public": he posts revenue, screenshots, and lessons, and that audience becomes the launch channel for whatever he ships next.
The transferable lesson
You do not need a single brilliant idea. You need a system for launching cheap experiments quickly and the discipline to cut the losers. Treat your first ten AI projects as lottery tickets you can print for free, then pour your time into the one that sells.
Playbook 2: Danny Postma and programmatic SEO
Danny Postma had already done the hard thing once: he built an AI copywriting tool called Headlime and sold it, a story he shared openly on Indie Hackers. Then in 2023 he launched HeadshotPro, which turns a few selfies into a set of professional headshots. It grew fast. Within roughly a year, multiple write-ups documented it clearing six figures in monthly revenue, with a meaningful slice coming from affiliate payouts.
The mechanism
The product is a one-off purchase: you pay, you get your headshots, done. But the engine behind the sales is search. An Indie Hackers breakdown of Postma's SEO approach shows he generated hundreds of location pages (think "professional headshots San Francisco") and ranked for high-volume keywords like "professional headshots" inside a few months. That is programmatic SEO: build many pages from one template, each targeting a long-tail search, so Google sends you buyers around the clock for no marginal cost.
The transferable lesson
If your AI product solves a problem people actively search for, your cheapest growth channel is owning those search results. Pick a niche with real search volume, build pages that genuinely answer the query, and let the traffic compound while you sleep. This is one of the most durable make money with AI mechanisms because it does not depend on paid ads or a personal following.
Playbook 3: Maor Shlomo and the solo exit
Maor Shlomo built Base44, a tool that lets anyone build a working app by typing a prompt, no coding required. He started it as a side project. Six months later, Wix bought it. According to TechCrunch, the deal was $80 million in cash, and Shlomo was the sole shareholder, with a small team of eight who shared a retention bonus.
The numbers underneath were real before the acquisition. As CTech reported, Base44 reached over 100,000 users and was profitable, generating about $189,000 in profit in a single month, nearly double Shlomo's own forecast. He documented his costs, including the heavy AI token spend, in public.
The mechanism
Base44 charged usage-based subscriptions for building and hosting apps. Because the product itself was AI doing the coding, one person could serve a fast-growing user base without a large engineering team. The profitability is what made the exit possible: a buyer was paying for a proven, growing, cash-generating asset, not a hope.
The transferable lesson
Profit changes your options. A small AI product that already makes money is not just a paycheck, it is a sellable asset. If you build something with real usage and clean economics, an acquisition can arrive far sooner than the old startup playbook suggests. Build it as if someone might want to buy it, because someone might.
Playbook 4: Marc Lou and shipping in volume
Marc Lou is a one-person software studio. He runs a portfolio of products, mostly tools and courses aimed at other developers, and he reports his numbers publicly. In his year-end recap he wrote plainly: "I made $1,032,000 in 2025." His two anchor products, the ShipFast code starter kit and the CodeFast course, each bring in around $20,000 a month, with newer products like DataFast adding to the pile.
His public revenue dashboards and the breakdowns on Indie Hackers show the same pattern Levels uses: many products, most of them small, a few of them carrying the bulk of the revenue. In the same recap he openly lists products that failed that year. The wins are not luck. They are what survives a high shipping rate.
The mechanism
Lou sells one-time licenses (a code template you buy once) and subscriptions (analytics, courses). AI tools let him build and launch new products quickly, and his audience on X turns each launch into immediate sales. Crucially, he sells to developers, a group that pays for tools that save time. The buyer profile is as important as the product.
The transferable lesson
Pick an audience that already spends money to solve its problems, then ship for that audience relentlessly. Most of what you launch will flop. That is fine and expected. Your job is to keep the launch cost low, keep shipping, and double down the moment something sells.
Playbook 5: Nick Dobos and free tools at volume
Nick Dobos took the opposite approach to everyone above. Instead of charging for a single polished product, he built BoredHumans, a website hosting more than 100 free AI tools: image generators, chatbots, writing helpers, games, and more, all on one domain. The published reporting on his exact revenue comes mostly from secondary aggregators rather than top-tier outlets, so the specific monthly figure should be treated with caution. What is clearly verifiable is the model and the scale: a single solo-built site offering 100-plus free AI tools, monetized by advertising against a large volume of traffic.
The mechanism
The tools are free, which removes the friction of a purchase decision entirely. Revenue comes from ads served against the traffic, so the business scales with visitors rather than conversions. Hosting many tools on one domain means each tool can rank for its own search terms and feed the same ad inventory. It is the media business model wearing an AI costume: attention in, ad revenue out.
The transferable lesson
Paid subscriptions are not the only way to make money with AI. If you can attract large volumes of free users to genuinely useful tools, advertising can carry the business. This path rewards breadth and search visibility over a single premium feature, and it suits builders who would rather optimize for traffic than for a sales funnel. Just be honest with yourself: ad models need real scale before they pay, so this is a volume game, not a quick win.
What these five have in common
Strip away the specifics and the same skeleton appears every time. None of these builders sold "AI" as the product. They sold a concrete outcome: a headshot, an app, a starter kit, a free tool that does a job. The AI was an input that made the product cheaper or faster to build, not the pitch.
Second, every one of them owned a distribution channel before the money showed up. For Levels and Lou it was an audience built by sharing their work in public. For Postma it was search rankings. For Dobos it was traffic volume across many tools. For Shlomo it was a public building journey that turned into word of mouth. The model was commodity. The channel was the asset.
Third, they moved fast and were comfortable killing things. The failure count in these stories is high. What separates them from people who never earn a dollar is not a higher hit rate, it is a higher shot count combined with the willingness to abandon what does not work.
How to pick your first AI money play
Start from what you already have, not from the most exciting story above. Run yourself through four questions.
- Do you have an audience? If you already post and people listen, the Levels and Lou playbooks fit: ship small paid products to the people who follow you.
- Are you good at SEO or willing to learn it? If so, the Postma playbook is the most repeatable. Find a niche with search demand and build pages that win those searches.
- Can you build something with real, recurring usage? Then aim for the Shlomo model: a usage-based product with clean economics that could one day be acquired.
- Would you rather chase traffic than conversions? The Dobos model rewards free, high-volume tools monetized by ads, as long as you can reach real scale.
Then commit to one. The most common failure among beginners is trying to run all four playbooks at once and finishing none. Choose the mechanism that matches your unfair advantage, give it a real ninety days, charge money as early as you can, and let the results tell you whether to push harder or move on. The builders above did not win because AI made them special. They won because they picked one mechanism and ran it longer than everyone who quit.
Books that shaped these playbooks
These three books keep coming up among solo builders, because they argue for exactly the small, fast, profitable approach the playbooks above use in practice.
Frequently asked questions
What is the most realistic way for a beginner to make money with AI?
The most repeatable beginner path is the search-driven product, the Danny Postma playbook. Pick a narrow problem people actively search for, build an AI tool that solves it, and create pages that rank for those searches. It does not require an existing audience or paid ads, just a real niche and patience while the rankings build.
Do I need to know how to code to start an AI side hustle?
Not anymore. Prompt-to-app tools like the one Maor Shlomo built, and a wave of similar vibe coding platforms, let non-programmers build working products from plain-English descriptions. Coding still helps, but the bigger levers for most beginners are choosing a paying audience and owning a distribution channel.
How much money can a solo founder realistically make with AI?
It ranges enormously. Many solo AI products earn a few hundred to a few thousand dollars a month, which is already life-changing for some people. The headline cases here, roughly $1 million a year for Marc Lou or around $3 million a year for Pieter Levels, are real but rare, and they came after years of shipping and many failures. Treat them as proof the ceiling is high, not as a typical result.
Why do these builders give away revenue numbers publicly?
Building in public is itself a growth strategy. Sharing real revenue earns attention and trust, and that attention becomes a distribution channel for the next product. Levels and Lou both treat transparency as marketing, which is part of why their launches convert so well.
Should I build one big AI product or many small ones?
The portfolio approach used by Pieter Levels and Marc Lou is lower risk for a solo founder, because no single failure sinks you and your winners reveal themselves over time. The single-product path, like Base44, can pay off faster and bigger but concentrates your risk. If you are early, a portfolio of cheap experiments is usually the safer way to find your winner.