How Much I Make on YouTube (And How I Run EVERYTHING)

TL;DR

A nearly 1M-subscriber AI YouTube channel earns $6,000–$7,000/month from AdSense, with sponsorships exceeding that, plus a full behind-the-scenes of tools and workflows.

Key Points

• 1.AdSense revenue is $6,000–$7,000/month. On a nearly 1M-subscriber channel posting 1–2 videos per week, that's the AdSense figure; sponsorship/brand deals generate even more but were not disclosed in detail.
• 2.Intro animations are made using AI video generators via Leonardo. He records two frames — himself off-camera and seated — then uses VO3.1, Cling 3.0, or Runway/Seed Dance with a text prompt to animate the transition, keeping the AI-generated audio.
• 3.Editing uses a live-recording-plus-Recut-plus-DaVinci workflow. A Stream Deck handles real-time scene switching during recording; Recut strips silences (cutting 65 min to ~27 min automatically); final polish and zoom effects happen in DaVinci Resolve using Greg's Presets.
• 4.The studio is a self-described 'nerd dream office' with high-end hardware. Key gear includes a Mac Studio M3 Ultra, Nvidia RTX 5090 PC, DGX Spark (128GB VRAM for local models), Stream Deck XL, Rodecaster Pro, and multiple cameras with teleprompters.
• 5.Business automation runs on N8N, Make.com, and Cursor-written scripts. His N8N workflow auto-scrapes new AI tools, runs them through GPT-4o Mini (with Gemini fallback), and populates the FutureTools.io database; Make.com handles news ingestion via Raindrop.
• 6.He uses Cursor and Claude Code as his primary app-building tools. He vibe-coded an AMA overlay app and a Python script that scraped six weeks of YouTube comments to find questions for this very video.
• 7.AI is not killing creators, but AI-generated slop will proliferate. He argues audiences still want authentic human voices and opinions; Gen Alpha brain-rot content requires more effort than people assume, and most faceless AI video channels won't be profitable.
• 8.ChatGPT still dominates consumer AI despite narrative of 'bleeding' users. He credits Anthropic's edge to prioritizing coding models — which unlock reasoning, tool use, and memory — and says OpenAI has now adopted the same strategy with coding-focused releases.
• 9.He stays current via Feedly (200–400 items/day), curated newsletters, and an X list called 'AI is Awesome.' Newsletters include The Neuron, The Rundown, AI Breakfast, and Matt Wolfe's own FutureTools.io/news page auto-populated by his Make.com automation.