Pieter Levels revenue: the never-sell math behind a $3M solo portfolio

"I don't like hiring, because I need to manage people, and I don't really like managing people, because it's not my thing. I like to create things."

Pieter Levels, speaking on Stripe's Cheeky Pint, July 2025

Every few months a screenshot goes around: Pieter Levels, one guy, no team, doing numbers most funded startups never touch. The screenshots are real. They are also the least interesting part of the story.

The number nobody frames correctly is not how much Pieter Levels revenue adds up to in a given month. It is what he does with a business after it works: nothing. He keeps it. He has kept almost all of them, for a decade, on one server, and that refusal to sell is the quiet engine doing most of the compounding.

Quick answer

As of July 2025, Pieter Levels (known online as @levelsio) said his companies together make about $3.1 million a year, all run solo with no employees and no outside funding. In that same first-party interview he put Remote OK at $3.4 million in lifetime revenue, Nomad List at about $700,000 in ARR, and PhotoAI at roughly $600,000 in ARR. Earlier, on his own blog, he reported PhotoAI peaking near $105,000 per month in March 2024. He has never raised venture capital and has never sold a company. The durable lesson is not the size of any single number. It is the never-sell, near-zero-cost portfolio structure underneath them.

The numbers, dated and attributed

Levels is unusually public, which makes the figures both checkable and inconsistent across time. Here is the honest reconciliation, each number tied to a date and a source rather than a vibe.

A few things jump out. PhotoAI's monthly run-rate looks lower in mid-2025 (about $50,000 a month implied by $600,000 ARR) than the roughly $105,000 a month he reported at its March 2024 peak. AI photo generation got crowded fast, and the revenue came off the top with it. That is not a failure. It is what happens to one hot product in a commoditizing category, and it is exactly the thing the rest of the portfolio is there to absorb.

One caution on the bigger aggregates floating around: some 2025 write-ups cite "$5.3M" across his projects. Treat that as a looser cumulative number. The first-party, current run-rate he gave in July 2025 was about $3.1 million a year.

Why he doesn't sell

Most indie playbooks end with an exit. Build a thing, grow it, sell it, start again. Arvid Kahl did exactly that with FeedbackPanda, and it was the right call for him. We wrote about that build-and-sell path earlier today.

Levels runs the opposite play.

In the Cheeky Pint interview, Stripe's founder-conversation podcast, he was blunt about the VC pitch he keeps turning down. "If you're a founder and you raise money, you kind of need to go big or bust," he said. An indie with a $10-million-a-year company, he argued, "is amazing," and there is no reason to trade that certainty for a lottery ticket.

So he keeps the companies. All of them. He talks about dozens of projects "running on my server," some with code he "didn't change in 10 years," still PHP, still paying. Go Effing Do It, an old motivation app, he says still makes about $50 a month. He does not turn it off. Fifty dollars is fifty dollars, and it costs almost nothing to keep alive.

That is the never-sell math. A portfolio of small, durable annuities you never liquidate compounds differently from a career of one-time exits. You are not repeatedly resetting to zero.

The price lever, not the user-count lever

Watch what he optimizes and what he ignores. He is not chasing a bigger user number for its own sake.

Nomad List has run a paid membership for years.

PhotoAI sells generation packs and subscriptions.

Remote OK monetizes job posts.

When a product finds demand, the move is to price it properly, not to inflate a vanity metric. For a solo founder this is the only sane lever. You cannot out-hire a competitor. You cannot out-spend one. But with a fixed, tiny cost base you can absolutely out-margin one, and price is where that margin lives.

The real moat is the cost base

Here is the line that should reframe every one of those revenue screenshots. On his blog, Levels has described serving roughly 4 billion requests a year across his sites on personal servers costing about $244 a month, at what he calls "99% profit margins" with heavy automation. For community moderation he leans on the GPT API rather than a support team.

Run the logic. When your costs are a rounding error and you have no payroll, revenue and take-home are nearly the same thing. A product that slipped from $105,000 to $50,000 a month is still, for a one-person company with a $244 server bill, an extraordinary amount of money to keep. The cost discipline is not a footnote to the revenue story. It is the story.

The part the trophy screenshots leave out

Two of Levels' most-shared moments are spikes, not baselines. His revenue "record" of about $420,000 in a single month came right after a September 2024 appearance on the Lex Fridman podcast. His browser flight simulator, fly.pieter.com, hit $1M in ARR in 17 days in February 2025 on the back of a viral moment.

Spikes are real, and they are wonderful. They are also not the plan. The plan is the boring line: many small products, kept for years, priced with nerve, run at near-zero cost. The viral months are a bonus the structure happens to be ready to catch. Tony Dinh's solo AI products show the same shape, big waves riding on a patient base, which we covered in the story of his $500k ChatGPT wrapper.

What you can actually copy

You are probably not going to reproduce the audience or the timing. You can copy the structure.

1. Build annuities, not exits. A handful of products doing $50,000 to $700,000 a year that you keep will out-compound a series of small sales you cash out and reset from.
2. Keep the quiet ones running if they cost nothing. The $50-a-month app stays on. Dead-simple products with no maintenance are free optionality.
3. Guard the cost base like it is the product. Near-100% margins are what turn a "declining" line into money you barely notice losing. Cheap hosting and automation over headcount is the whole game.
4. Pull the price lever before the growth lever. When something works, charge for it properly. That is the move a solo operator can actually execute alone.

Levels is not proof that everyone should stay solo forever. He is proof that "never sell, spend almost nothing, price with nerve" is a real business model, and a surprisingly sturdy one. The screenshots will keep going viral. The structure underneath them is the part worth stealing.

Keep reading

• The founder who reached $55K MRR with no audience, then sold
• The $500k solo ChatGPT wrapper, and what came after the spike