TL;DR
Pure AI-voiceover + AI-image slideshow channels were demonetized en masse in 2024–25 under YouTube's "mass-produced or repetitive content" policy. What works now: documentary niches with human research, real archival footage, original commentary. Realistic ramp: 12–18 months to monetization, $0–$30k/mo at scale, heavily niche-dependent.
The 2024–25 reset, in one paragraph
YouTube updated its monetization policy on July 15, 2025 to explicitly target "mass-produced and repetitive" content — directly aimed at the AI-text-to-AI-voice-over-stock-footage channels that dominated 2023. Channels with hundreds of formulaic videos got demonetized in waves. Many lost ad revenue overnight. The crackdown is real, ongoing, and unevenly enforced.
What survived
- Documentary-style channels with human research, original scripts, and original or properly-licensed B-roll.
- Compilations with commentary — top-10s with a recognizable host voice (even if anonymous) and substantial editorial work.
- Tutorials and explainers in niches where the content has objective utility (coding, finance, woodworking).
- Music channels (lo-fi, instrumental) with original or licensed compositions.
- Animated explainers with original animation, even if the voice is AI.
Real CPMs by niche (2026)
| Niche | RPM range | Notes |
|---|---|---|
| Finance / investing | $15–$45 | Top tier. Audience high-value. |
| Business / B2B SaaS | $20–$60 | Highest CPMs, smaller audiences. |
| Real estate / mortgages | $25–$70 | Seasonal spikes during home-buying season. |
| Tech tutorials | $8–$25 | US-heavy audience helps. |
| True crime / documentary | $4–$12 | Big audiences, moderate RPM. |
| History / educational | $3–$10 | Lower RPM but extremely watch-time-friendly. |
| Gaming compilations | $2–$8 | High volume, low RPM. Need scale. |
| Reaction / commentary | $3–$10 | Copyright risk is the bigger issue. |
| Lo-fi music | $0.50–$2 | 24/7 livestream model only. |
Source: composite of public creator-disclosed RPM data, 2026.
The faceless production stack that actually works
- Research: you, doing it. Notion or Roam for the research database. AI is a research-assistant, not the researcher.
- Script: human-written, AI-assisted for polish. Original takes are the moat.
- Voiceover: ElevenLabs with a custom voice clone (yours or a paid voice actor's). Avoid the default "AI voices" — they're a tell.
- B-roll: Storyblocks, Artgrid, or properly licensed archival. Getty Embed for some news/sports clips. Avoid pure stock slideshow.
- Editing: DaVinci Resolve (free) or CapCut (free for most cases). Premiere Pro if you want subscriptions.
- Thumbnails: Photoshop or Photopea. This is where you spend disproportionate time — thumbnail click-through is half the game.
- Upload cadence: 1–2 long-form videos per week. Shorts daily if you're chasing growth.
The honest time + money math
| Phase | Time/week | Cost/mo | Realistic outcome |
|---|---|---|---|
| Month 1–3: foundation | 15–25 hrs | $50 | 0–500 subs. Not monetized. |
| Month 4–8: iterating | 15–25 hrs | $100–$200 | 500–5,000 subs. First monetization milestone (1k subs + 4k watch hours). |
| Month 9–18: scaling | 20–30 hrs | $300–$800 | $200–$2,000/mo from AdSense. First sponsorships. |
| Year 2+: established | 20+ hrs or hire editor | $500–$2,000 | $2,000–$30,000+/mo across AdSense + sponsors + affiliates. |
What kills faceless channels
When to skip faceless YouTube
Skip if: you want monthly cash within 6 months, you can't commit 15+ hours/week, you have no edge or insight in the niche you'd pick, you can't tolerate a long horizon of no measurable progress.
Run it if: you have genuine subject-matter expertise (engineering, history, finance, etc.), you can write, you can edit, and you treat the first 18 months as the cost of building an actual media asset.