Is AI Profitable Yet? What 106 AI Businesses Actually Bank
Yes — for small teams selling a narrow product, not for anyone hoping the technology itself is the business. Across the 106 products built with AI coding…
Yes — for small teams selling a narrow product, not for anyone hoping the technology itself is the business. Across the 106 products built with AI coding tools that we index, 32 disclose a clean monthly figure with a median of $25K/mo and a range from $37/mo to $2M/mo.
The condition that flips that verdict is distribution. Every entry near the top of our range had an audience before it had a product; every entry near the bottom built first and looked for buyers afterwards. That is the honest version of "it depends", and it is checkable.
Table of contents
Is AI profitable yet, or still a promise?
Profitable, narrowly and unevenly. 71 of the 106 are one-person businesses, so the profitable cases are mostly individuals, not funded teams.
Those numbers are revenue, not profit margin, and they come from businesses that chose to publish. Our honest caveat, stated once: this index over-represents operations that survived and then talked about money, so the median is a comparison tool, never an expectation for your own launch. We record 7 documented failures here and 38 across the full 406-business index, far below true attrition.
The U.S. Census Bureau Business Trends and Outlook Survey AI use data is a better adoption baseline than any vendor's chart.
The case for, in numbers
The strongest outcomes in our index are real, verified, and run by small teams. They are also not typical, and reading them as a forecast is the error this article exists to prevent.
- ·[The Viral App Monetization Machine (100 Apps Dissected)](/projects/paywall-machine) — Cal AI and Lerna at $2M/mo each; LazyFit, CoinSnap and Impulse at $700K/mo each. Evidence class: ✅ Verified · Hard Data. Small teams throughout; the creator behind the dissection reports PuffCount at $1,300/day and Posted at $150K+/mo.
- ·[Cluely](/projects/cluely-stealth-ai-assistant) — $500K/mo (~$6M ARR). Evidence class: 🗣 Founder-Reported. Roy Lee, co-founder Neil, ~13-14 people plus 60 contracted creators. The contracted-creator count is the tell: this is a distribution operation with an AI product attached.
- ·[Niche Identifier App Portfolio (CoinSnap as the Model)](/projects/coinsnap-niche-id-app-empire) — $500K/mo as the top identifier app per Sensor Tower. Evidence class: ✅ Verified · Hard Data. Run by a portfolio operator, not a solo founder.
- ·[Gravl](/projects/gravl-ai-gym-workout-app) — $440K/mo with 70K+ subscribers. Evidence class: ✅ Verified · Hard Data. Started with Julian and two others; now 13-14 people. Headcount followed revenue.
- ·[Grower](/projects/grower-turkey-ai-marketing-platform) — $180K/mo MRR, growing ~10% month over month, 50+ employees, 200+ companies applying daily, and monthly revenue 3x the total investment ever raised. Evidence class: 🗣 Founder-Reported. Built by Barış Korkmaz out of Istanbul.
Four of those five are verified or on the record. That is the case for.
The case against, in the same numbers
The same dataset argues the other way. 74 of the 106 indexed businesses here disclose no clean monthly figure at all. Most readers will land nearer the middle than the ceiling, and the middle is the only honest planning number.
In our AI Clone of Floe-Style Agency SaaS entry, the original benchmark does $1.3M/mo while the person building the clone on camera does $0 — creator-reported, one person aided by a voice assistant. Two numbers, one entry, and the impressive one belongs to someone else.
The structure repeats. TinyURL Clone Link Shortener (Replit + AdSense) quotes the benchmark TinyURL at $200K/mo on 40M monthly visits — the builder is a solo non-coder named Mikey, and 40M visits is the actual requirement. The AI-Built Single-Page Read-Later App cites $60K/mo for the benchmark while noting the builder's own numbers are unverified. And our Receipt-Scanning Apps entry lists SimplyWise and a neighbour at an estimated $60,000/month each and a third at $80,000/month — all on-screen App Store estimates from an unnamed source tool.
Now the small end. AI Co-founder Club did $8,374 MRR from 270 paying customers 13 days after launch, and NoFap / Men's Self-Improvement App did $6K/mo from ~1,100 paying users in its first month live. Real businesses, an order of magnitude below the median.
How much is each of these numbers actually worth?
This group splits 20 third-party verified, 49 founder-reported, 20 creator-relayed, 17 unproven — 65% verified or founder-reported. Better than most categories, and still one in six figures here has no supporting artefact whatsoever.
| Evidence class | Count in this group | What it means |
|---|---|---|
| ✅ Third-party verified | 20 | Someone outside the business saw hard data — a dashboard, an audited record, an estimation platform |
| 🗣 Founder-reported | 49 | The operator stated the figure and put their name to it |
| 📎 Creator-relayed | 20 | Repeated by a host or writer who did not personally verify it |
| 🔮 Unproven | 17 | A claim with no supporting artefact at all |
This is the argument the entire site rests on. A founder-reported figure and a verified figure are not the same claim, and publishing them side by side in identical formatting is how the internet launders hype into fact. Gravl's verified $440K/mo and Upward's creator-reported $400K/mo from revenue-tracking tools, team not disclosed in the video, look alike on a list and are not alike as evidence. Rubric: how we grade the evidence; every entry in all indexed ideas carries its class.
What would have to be true for this to work for you
Three testable conditions, none about grit or vision. If any one fails, the median $25K/mo is not a number you should be planning against.
- 1.You already have a channel, or you can prove demand before you build. Quick Shorts made $500+ on launch day because one Reel about auto-posting Shorts with N8N hit 2.3M views, collected nearly 10K emails and a 1,000-person Discord — the product was built in Lovable only after demand was proven.
- 2.Your buyer is already saying the words out loud. The Intent-Signal Lead Finder entry suggests $500-$1,000 per lead or $5,000 per project on exactly this basis: skip the 1,000 cold emails and go to the 3% publicly asking, with a finished sample attached.
- 3.You are selling an outcome, not a technology. AI Co-founder Club's $8,374 MRR came from packaging a scary tool safely — fear converted into a monthly convenience fee, not "we use AI".
How to use this without fooling yourself
- 1.Plan against $25K/mo, not $2M/mo. Write the median at the top of your model. If the business only works at Cal AI's verified $2M/mo, you do not have a plan.
- 2.Test condition one this week. Post the demo before you build the product, the way Quick Shorts did. If nothing lands, you have saved yourself a build.
- 3.Label every benchmark in your deck. If your projection leans on Upward's $400K/mo, write "creator-reported" beside it; if on CoinSnap's $500K/mo, write "verified". Start from the Built With AI hub and check the class before the number.
- 4.Sanity-check the market, not the hype. Use the U.S. Census Annual Business Survey and the SBA Office of Advocacy small business profiles for the baseline. For build quality read Vibe Coding: Can It Build a Profitable Product?; for niche selection, Profitable Niches For Dropshipping applies the same method.
Frequently asked questions
Which 5 jobs will survive AI?
We do not forecast labour markets and will not invent a list. Our data shows something narrower and more useful: 71 of the 106 indexed products built with AI coding tools are run by one person, and 32 disclose revenue with a median of $25K/mo. The pattern is small operators shipping narrow products, not job categories disappearing.
What is a $900000 AI job?
That figure is not in our data, so we will not repeat or explain it. The highest verified figures we hold here are business revenues, not salaries: Cal AI and Lerna at $2M/mo each (verified hard data). Compensation claims circulating online rarely carry an evidence class, which is precisely why we do not publish them.
What will be AI after 10 years?
We index what businesses have banked, not what technology will do. Our only defensible statement about the future is the shape of the present: a $37/mo floor, a $2M/mo ceiling, a $25K/mo median, and 17 of 106 entries here whose claims have no supporting artefact. Anyone offering a ten-year figure is not working from evidence.
How reliable are these is ai profitable yet figures?
Reasonably, with the classes attached. This group splits 20 third-party verified, 49 founder-reported, 20 creator-relayed, 17 unproven — 65% verified or founder-reported, against 57 / 184 / 121 / 44 across the whole 406-business index. Only the verified fifth rests on data someone outside the business actually saw.
How many of the 106 indexed businesses actually disclose a monthly number?
32 of 106 in this group. Across the whole index, 106 of 406 businesses disclose a clean monthly figure, median $27K/mo, with 38 documented failures across 22 countries. The 74 here that disclose nothing are not missing data to be estimated — their silence is part of why headline AI revenue claims deserve scepticism.