All articles

Your AI channel makes $0. Mine makes $12k a month. Why?

Eli Whitmore·Jun 25, 2026·11 min read

Quick math up top so the title isn't doing too much work.

My channel makes about $12,000 a month right now. $7,500 is YouTube ad revenue, $3,500 is two brand integrations I run on rotation, $1,000 is a $27 ebook I wrote on a slow weekend. May was the first month it crossed $10k.

The "I broke all of them first" part is also true. My first 9 videos averaged 400 views. Combined ad revenue across those 9: one dollar and forty-three cents.

If you came here hoping "this is easy," sorry. The workflow is easy. The hard part is knowing what to do.

One thing I would tell past-me. Don't try to learn YouTube algorithms from YouTube. I spent the first month doing exactly that. Every channel had a different answer to the same question. A structured course with daily 15-minute lessons was the unlock. YouTube is great once you already know what to watch. It is hostile when you are still figuring that out.

One more thing. The most important rule is #13. I am saving it for the end on purpose.

About me

34, ex-graphic designer at an Atlanta agency, laid off in November when they "restructured." Started an AI animation Shorts channel in February. Animated character stories in a fantasy-farming world, same lane as the Roblox-game animation channels everywhere lately. I picked it because the visual style was familiar, a channel in the lane was already pulling real money, and the audience overlaps with kids and teens, where most YouTube Shorts watch time actually lives.

Now the 13.

Channel setup

Rule 1. Account ≥ 7 days old, with watch history. My first attempt was a brand-new account that morning. Got 12 views in 48 hours. YouTube treats empty fresh accounts as bots. Use the account for a week first. Watch videos, subscribe, comment. My third attempt I used my actual high-school-era Google account. Views immediately normalized.

Rule 2. "Not made for kids." Always. Even if your content IS kid-friendly. The COPPA toggle turns off personalized ads, comments, end screens, and brand integrations. Earnings tank by about 80%.

Rule 3. Handle + profile picture. Done. I spent two days on a banner, polished description, and a logo I redesigned 14 times. None of it moved views. The first 3 seconds of your videos decide everything.

The hook

Rule 4. Motion in the first frame. Non-negotiable. Static openings are swipe traps. My early videos opened on a slow logo-fade. Got 200 views. I switched to "character running straight at the camera" as the first frame. Swipe-through rate jumped from 45% to 72% on the very next upload.

Story structure

Rule 5. Conflict arc, not linear arc. Linear is hook → rising action → payoff. Conflict arc is hook → rising action → conflict → comeback → rising action → payoff. The conflict-comeback in the middle makes the story unpredictable. My average view duration jumped from 22 sec to 36 sec on 45-second videos after I switched.

Rule 6. Original ideas don't win. Better execution does. I spent six weeks trying to invent something nobody had done. Then I picked a niche where a real channel had already hit 100M+ views. I didn't clone them. YouTube doesn't push two of the same. But knowing the niche worked removed the biggest unknown.

Upload cadence (the one I broke worst)

Rule 7. Wait at least 48 hours between uploads for your first 5 uploads. After my 3rd video hit 500K views I rushed and uploaded the 4th the next morning. It flopped under 1,000. The 3rd stopped getting pushed too. Both videos lost. 48 hours is roughly how long YouTube needs to finish its test push and decide whether to push wider.

Rule 8. Watch real-time view candles. Only upload when they drop below 100/hr for 12+ hours. This is the actual signal. Above 100/hr means the previous video is still being pushed. Below 100/hr for 12+ straight hours means YouTube is done measuring.

I learned this from a mix of YouTube tutorials, a Skool community I paid into for 3 months, and a short course on Mindwand that broke channel-building into 15-minute daily lessons I actually finished. Coursera and YouTube tutorials cover similar ground. Pick one. Finish it.

Niche stability

Rule 9. Don't pivot in your first 5 uploads. One video flopped on my second channel and I panicked into a different niche. 8 videos later across two niches, the algorithm had no idea who my audience was. Zero broke 5K views. Third channel I made myself wait 5 uploads. By video 5 the algorithm had figured out my viewers.

Production

Rule 10. Use reference images for AI generation. Text-only prompts produce inconsistent characters and broken palettes. Generate one strong frame, save it, upload it as a reference for everything after. Characters stay the same character. Palettes hold.

Rule 11. Cut every clip to 1-3 seconds. Pacing carries the retention curve. CapCut handles it. Average view duration improved about 30% between my "long cuts" videos and my current "fast cuts" videos.

Rule 12. Realistic themes only. No gimmick AI niches. "Hyperrealistic AI history" channels were going viral in March. By June they were dead. Novelty kills brand equity. Animation niches work because viewers follow the character, not the AI.

The mindset rule

Rule 13. Your first upload is the least representative datapoint. The one nobody told me, almost made me quit. New channels go through a discovery phase where YouTube has not found your audience yet. Same video uploaded as #1 vs #3 can perform 10-50x differently. I had a channel where upload #1 hit 70K views, I quit, then looked back later and realized upload #2 would have probably hit a million. Plan for 30 videos before you judge anything.

The honest math

Channel started Feb 4. Today is June 25. 5 months. Video 26.

  • Videos 1-9: averaged 400 views. Total ad revenue: $1.43.
  • Videos 10-15: averaged 8K views. Ad revenue: $87.
  • Video 16: hit 1.2M views accidentally. $1,840 from that one video.
  • Videos 17-22: averaged 200K once the algorithm found my audience. $4,200.
  • Videos 23-26: averaged 500K. Ads + brand deals + ebook compound to June landing at ~$12k.

Averaged across 5 months, my hourly rate is around $20/hour. Better than agency design. Much worse than what the YouTube tutorials promise. Realistic.

Pick a niche someone has already proven works. Set up your account properly before you upload. Commit to 30 videos. And pick one structured course or community to finish before you start. I did Mindwand for the daily-lesson format. Could have done Skool. Could have done Coursera. Doesn't matter which. Matters that you commit and finish.

That's the list. I hope some of it lands.

Frequently asked questions

Tools I pay for: ChatGPT Pro plus ($20/mo), one AI image+video tool ($30-60/mo depending on volume), Skool community subscription for the first 3 months ($89/mo, dropped after), CapCut Pro ($10/mo though the free version works fine). Total around $60-80/mo on the lean stack once I dropped Skool. The first month it was higher because I was experimenting with several image-gen tools. Worth verifying current pricing on each platform yourself before committing prices change.

Realistic answer: you need 1,000 subscribers and either 4,000 long-form watch hours OR 10M Shorts views in the prior 90 days to qualify for YPP (the YouTube Partner Program). I hit the subscribers threshold around video 18 and the Shorts views around video 22. So somewhere between video 18 and 25 is realistic if your videos are landing. If your videos aren't landing, you're not getting close until you fix what's broken.

Then the niche is wrong OR the execution is wrong, and you can finally tell the difference. If all 30 underperformed, it's the niche pivot. If 5-10 of them hit and 20 didn't, study what made the 5-10 work and concentrate on that. The key is that you need 30 datapoints before this question is even answerable. Asking it after 5 videos is pointless.

Honestly, in some sub-niches yes (anything involving 'hyperrealistic AI history' or 'AI religious stories' is brutally saturated). In adjacent sub-niches no there's still real room in animated character stories, niche-game animations, animated wisdom-storytelling. Pick a sub-niche where you can find one shop / one channel with proven recent traction but where the field isn't 100 shops deep. That's the credibility-zone everyone fights over but most beginners overshoot trying to be original.

Yes, but be realistic about the hours. Each video takes me about 2 hours start-to-finish at my current pace (was 6+ hours at the beginning). Multiply that by 3 videos a week and you're looking at 6-8 hours of after-work production. If you can carve out 8 hours a week on top of your day job for 4-6 months without quitting, this is a real path. If you can only spare 2 hours a week, this isn't going to work and that's fine.

Keywords

AI ShortsYouTubeAI AnimationAI IncomeSide Hustle