Wes Roth

AI Apps Making $20,000+ per month with 1 person teams.

⏱ 21 min video · 3 min read12 Jul 2026Worth watching

TL;DR

Wes Roth launches a new series on building AI-powered solo businesses, using real case studies like David Brereton's Formula Bot ($226K/month MRR by 2025) and PDF.ai ($1.5M ARR peak) to show what is realistically achievable. The video frames expectations around a 1-5 year horizon and encourages beginners to start using Claude or ChatGPT immediately to build and ship small apps.

Key points

David Brereton built Formula Bot (Excel formula helper) with no coding skills using Bubble AI, grew it from $25K/month in 2023 to $226K/month (~$2.7M ARR) by 2025, and turned down acquisition offers from Microsoft and VC money from Sequoia.

ThumbnailTest.com was generating $20K/month in 2023 and sold for low-to-mid six figures in 2024, showing a viable exit path for small AI apps.

PDF.ai was purchased for $20K from a dead GitHub repo, grew to $1.5M ARR at peak, and stabilised around $500K ARR — still a strong return on a $20K investment.

People consistently overestimate 1-year business results and underestimate 5-year results; Wes says he would not bet on a short-term goal but would comfortably bet on a high six-to-seven figure outcome over five years if the founder keeps working.

The recommended starting path is: get a paid Claude or ChatGPT account, maximise your usage quota daily, build small apps, then ship something publicly — even if imperfect — to get real user feedback.

Actionable insights

→

Get a paid ChatGPT or Claude account and treat your monthly usage quota as a resource to burn through productively every period — building apps, doing research, automating parts of your workflow.

→

Use Claude or ChatGPT for deep research and bullet-point cheat sheets, but write or speak your own final output — never let the AI become your mouthpiece or outsource your understanding.

→

Ship something small publicly as fast as possible, even if unpolished — just get a few real humans to visit and interact with it. The goal is to get comfortable with the build-and-ship loop, not to launch a perfect product.

→

When evaluating a business idea, think in a 5-year window rather than 1 year. The case studies here (Formula Bot, PDF.ai) all took 1-4 years to reach significant recurring revenue.

→

No-code and AI-assisted tools like Bubble AI mean you do not need to be a developer to build a profitable SaaS product — David Brereton reached $2.7M ARR and still cannot code.

Notable quotes

“People that teach you how to make money, but the way that they've made their money is teaching people how to make money — that's just in general a huge problem and it ends in tears.”

“We tend to be optimists for the near-term and kind of pessimists for the long term. But with building businesses, it tends to be the other way around.”

“This is your enemy. This is the goal. This is the thing to burn through — your goal should be every day, every week, to put all of your allowed quota towards building or researching or doing something useful.”

Worth watching?

✅

Worth watching the full video?

Watch if you want the concrete case studies and revenue figures — the full video also includes a live walkthrough of how Wes uses Claude and ChatGPT in his own workflow, which is genuinely instructive for beginners.

Topics

Business OpenAI

Explore more summaries on these topics →

Saved you some time? The creator still deserves a like.

Watch on YouTube →

More like this