Algrow

≈ $14,000 / month revenue

• Built by Sam, a university student with no prior coding background.
• First MVP written by pasting ChatGPT code into Notepad, then rebuilt and scaled in Cursor.
• Helps creators, dropshippers & influencers research and replicate viral video formats.
• Got its first ~400 users entirely through Discord communities.

LaunchFast

$0 → ≈ $21,800 MRR in 90 days

• Built by Hassam, a non-technical founder, in roughly 48 hours using Cursor.
• AI tool for Amazon private-label sellers to find, validate and source products.
• Distribution solved by partnering with LegacyX, a coaching company with an existing seller audience.

Goji Berry AI

$0 → ≈ $30,000 MRR in ~5 months

• Built by Roman, a former mechanical engineer who previously built and sold Coco AI.
• LinkedIn outreach tool built on "intent signals" — targeting prospects showing buying readiness.
• Mostly $99/mo plans. First customers acquired via honest storytelling posts on Reddit (the marketing channel, not a product feature).

Photo AI

≈ $105,000 / mo revenue · ≈ $80,000 / mo profit

• Run solo by Pieter Levels (@levelsio); large PHP codebase maintained with AI-assisted coding.
• Users upload selfies to train an AI model, then generate studio-grade photos and videos via prompts.
• Over 29 million photos generated to date.

fly.pieter.com

≈ $1M ARR in 17 days

• Built as a public experiment by Pieter Levels — his first "vibecoded" flight simulator.
• A multiplayer browser flight sim requiring zero downloads.
• Monetized through in-game advertising.

DataFast

19,161 users · plans from $9–19 / month

• Built by Marc Lou, who reports 24 startups and $2M+ earned online; "Built with ShipFast."
• Shows which marketing channels drive paying customers, not just pageviews.
• Ships an AI-agent CLI so Cursor, Claude Code and Codex can query revenue from a prompt.

Note: the site states its user count and pricing but does not publish DataFast's own MRR, so no revenue figure is claimed here.

Medvi

$401M revenue (2025) · $65M profit · $1.8B projected (2026)

• Built for ~$20,000 by Matthew Gallagher with a headcount of two, using ChatGPT, Claude & Grok for code.
• 250,000 customers; 16.2% net margin — roughly 3× Hims & Hers' margin.
• Infrastructure (physicians, pharmacy, compliance) rented from CareValidate & OpenLoop.

⚠ Per the same Forbes report: Medvi received an FDA warning letter (Feb 2026) over compounded GLP-1 marketing, and a researcher publicly reported a HIPAA/IDOR data breach. Its $1.8B projection depends on a regulatory window that may close.

Lovable

≈ $500M annualized revenue run rate

• Europe's fast-growing vibe-coding platform, founded late 2023 — not yet three years old.
• 50M+ projects built; usage now at ~1 million new projects per week.
• Users are primarily non-technical founders, designers and salespeople building monetizable software.

The Harsh Realities of Making Money with AI

The numbers you just saw are real — and they're the survivors. Here's what AI actually does, what it doesn't, and the failure modes hiding behind the headline figures.

Build ≠ Business

The build was never the hard part

In every verified case, distribution decided the outcome

• Algrow: first ~400 users earned by hand in Discord communities.
• LaunchFast: revenue only came after partnering with an audience that already existed (LegacyX).
• Goji Berry: $30K MRR built on months of honest Reddit storytelling.
• AI can give you a product in an afternoon. It cannot give you an audience, a niche, or trust — those are still slow, human, and unautomated.

You're looking at the winners

~1,000,000 new projects per week on Lovable alone

• The case studies are a handful of survivors selected from tens of millions of attempts.
• TechCrunch notes the platforms have not yet transparently reported abandonment rates — "the not-as-flattering stuff."
• The honest read: for every public success here, an unknown and almost certainly vast number of projects earned nothing and were quietly dropped.
• Survivorship bias is the single biggest distortion in any "I made $X with AI" story.

Speed creates real liabilities

Shipping fast is not the same as shipping safe

• Maintenance: TechCrunch's core warning — software sits on a shifting stack of dependencies that constantly break. Building is easy; keeping it running is the hard, ongoing part, and it's why many companies still buy instead of build.
• Hallucination: Medvi's AI chatbot fabricated drug prices (which the founder had to honor) and invented product lines that didn't exist.
• Security: Medvi shipped a textbook IDOR flaw exposing 250,000 patients' health records — no login required.
• Regulation: An FDA warning letter doesn't care that a solo founder built it with AI. The liability is fully human.

An outlier is not a template

Even Medvi's backers frame it carefully. VC Kobie Fuller called it "an extreme example" — an early data point, not a repeatable playbook. Sam Altman called the one-person billion-dollar company "unimaginable without AI," but it took a founder who was already technically fluent, marketing-savvy, and willing to absorb serious regulatory risk. AI lowered the floor for building. It did not lower the floor for judgment, distribution, or accountability — and those are still where most of the money is won or lost.

Building is cheap. Distribution & durability are the moat.

Across every verified case, the AI build was the easy part — the winners solved distribution (Discord, Reddit, partnerships) and the open question, per TechCrunch, is maintenance: whether vibe-coded software stays running as dependencies shift.