This one almost didn’t get written. Not because the question is bad — honestly, it’s the most honest question a new creator can ask. It’s because nearly every answer I’ve seen online is junk.
You know the type. Thumbnails screaming “I MADE $47,000 WITH AI VIDEOS.” Screenshots that may or may not be real. Gurus selling courses who never, ever show their own analytics.
So let me try something different. No income screenshots. No promises. No “and YOU can too.” Just how YouTube pay actually works, the variables that decide what any given channel earns, realistic ranges, and the honest reason most AI channels make almost nothing. Want fantasy? Plenty of other tabs open for that. Want to understand the math before you sink six months into this? Keep reading.
How YouTube Pay Actually Works (RPM in One Minute)
Two numbers. That’s it. Everyone mixes them up, and that one mix-up is why half the “how much will I make” advice on the internet is wrong.
CPM (Cost Per Mille) is what advertisers pay YouTube for 1,000 ad impressions. That’s not your number. That’s YouTube’s number. It’s usually higher than what you see — often $10 to $30 or more in valuable niches — because YouTube takes its cut before you ever see a cent.
RPM (Revenue Per Mille) is what YOU actually earn per 1,000 views, after YouTube’s cut. For long-form content, you keep 55% of ad revenue. RPM is the number sitting in your YouTube Studio dashboard. It’s the only one that matters for planning.
Why does the distinction matter? Someone tells you “finance videos have a $25 CPM.” Sounds amazing. Right? Except the creator’s actual RPM might be $8–12 after YouTube’s cut, ad blockers, non-monetized views, and Premium revenue blending. That’s less than half of what the CPM suggested. So: always think RPM. CPM is a vanity number for you.
Shorts run on a different model entirely — pooled revenue, 45% creator share — and the RPMs are dramatically lower. Way lower. I break that down separately below, and I go deep on what actually sets your rate in YouTube RPM for AI videos.
Realistic Ranges by Content Type
I’m going to give you ranges. Not promises. Ranges. Based on how YouTube’s ad system works and what’s widely reported by creators — zero guarantee of what you’ll earn, because that comes down to your niche, audience geography, video length, season, and advertiser demand. Anyone who hands you a guaranteed number is selling something.
Long-form AI content, by niche type:
- High-value niches (finance, software, business, professional education): advertisers here pay premium rates, and RPMs land at the higher end of the spectrum. But — and this is the part people skip — these niches demand genuine expertise. An AI channel faking financial knowledge won’t survive a review, let alone earn from it.
- Mid-range niches (technology explainers, history, science, productivity): solid advertiser demand, moderate competition. Honestly? This is where most successful AI educational channels actually live. Not glamorous. Just works.
- Lower-range niches (entertainment, comedy, general viral content, gaming): enormous view potential but way less advertiser willingness to pay per view. You need serious volume for the math to work. Like, serious volume.
What I’d plan around: new AI channel, mid-range niche, mostly US audience? Model conservatively. Take the lower end of whatever range you’re hoping for, then be pleasantly surprised if you beat it. For Shorts, assume a fraction of long-form rates. A small fraction. I’ve watched too many creators build financial plans on best-case RPMs and then panic when reality came in lower.
The variables that move your RPM up or down:
- Audience geography: US, UK, Canadian, and Australian viewers generate significantly higher RPMs than viewers in regions with less advertiser spending. Two channels with identical view counts can earn completely different amounts. Wild, but true.
- Video length: longer videos (8+ minutes) can run mid-roll ads, which pushes revenue per view up. That’s a structural advantage long-form has over Shorts — built into the format itself.
- Seasonality: Q4 (October–December) usually has the highest advertiser demand and RPMs. Q1 (January–March) is typically the lowest. Plan your expectations around that rhythm.
- Niche advertiser competition: more advertisers bidding for your audience means higher rates. This is why finance beats comedy, every time.
- Engagement signals: better retention and engagement can nudge RPM indirectly by signaling quality to YouTube’s ad systems. Not a direct lever, but it matters.
Shorts vs Long-Form: The Math Side by Side
Let me put the two formats next to each other, because this is where I keep pushing long-form for AI creators who actually care about revenue.
Long-form: you earn 55% of ad revenue from ads on your videos. RPMs are measured in dollars per thousand views. A video with 100,000 views in a decent niche generates meaningful revenue — enough to notice, enough to reinvest into the channel.
Shorts: revenue gets pooled across all Shorts, then split by view share. You keep 45% of your allocation. RPMs are measured in cents per thousand views. A Short with 100,000 views generates pocket change. Literally pocket change.
To make it concrete: the revenue from 1 million long-form views might take 50 to 100 million Shorts views to match. I’m not handing you exact dollar figures because they’d be misleading — the ranges are too wide. But the order-of-magnitude gap is real. It’s structural. It’s not that Shorts are “worse.” They’re just a completely different economic model.
Doesn’t mean Shorts are worthless, by the way. They’re great for discovery, great for audience building, great for funneling people to your long-form stuff — read the full Shorts monetization breakdown before you pick a lane. But if someone tells you they’ll build a full-time income purely from AI Shorts ad revenue? Ask to see their analytics. The math is brutal at anything below massive scale.
What AI Production Really Costs
Here’s the part most “AI passive income” content conveniently leaves out. AI video production isn’t free. And the costs stack up in ways people don’t see coming.
Direct costs:
- AI video generation: tools like Veo, Runway, and similar charge per generation or through subscriptions. Costs vary enormously — from a few cents per clip on some platforms to dollars per generation on others. And here’s the kicker: if you’re iterating (and you should be), multiply your per-video estimate by the number of attempts. For a full tool-by-tool comparison, see my 2026 pricing breakdown for AI video generation.
- AI voiceover: quality AI voices cost money, either per character or by subscription. Free options exist. The quality gap is audible, though.
- Scripting tools: using AI writing assistants? Factor those subscriptions in.
- Editing software: Premiere Pro, DaVinci Resolve Studio, or alternatives — subscription or one-time cost, either way it counts.
- Thumbnails: AI image generation for thumbnails, plus whatever design tools you use.
- Music licensing: royalty-free music subscriptions, unless you enjoy copyright claims.
Hidden costs:
- Your time. The big one. Even with AI doing the heavy lifting, a quality video takes hours — scripting, directing generations, editing, optimizing. If you value your time at anything at all, the “passive” narrative falls apart fast.
- Learning curve. Your first 20 videos will take three times longer than your 50th. Budget for that. Nobody’s fast on day one.
- Failed experiments. Not every video works. The flops still cost you production time and AI credits. That’s just the game.
I’m not saying the costs are prohibitive — AI is genuinely cheaper than traditional video production, no question. I’m saying: go in with real numbers, not the “AI is free” fantasy. Calculate your per-video cost honestly. Then figure out how many views you need just to break even. For most new AI channels, the answer is “more than you’ll get in the first six months.” And that’s fine — as long as you know it going in instead of finding out the hard way.
Why Most AI Channels Earn Little
Time to be blunt about why the vast majority of AI YouTube channels earn little or nothing. It’s not bad luck. It’s structural.
They’re in low-RPM niches. Most AI channels pile into entertainment, fun facts, and general viral content — the lowest-paying categories there are. The niches that actually pay demand expertise that’s hard to fake with AI. Niche selection is step one of the 2026 roadmap for monetizing AI videos on YouTube, and this is exactly why.
They don’t reach monetization thresholds. Getting to 1,000 subscribers and 4,000 watch hours (or 10 million Shorts views) is genuinely hard — I cover the exact requirements plus the AI-specific gotchas in my YouTube monetization requirements checklist. Most channels quit before they get there. And of the ones that make it to the line, plenty get rejected at review for inauthentic content.
Their content doesn’t hold viewers. AI-generated videos often have this novelty factor that gets the click but can’t keep attention. Low retention means fewer ads served, lower RPM, and algorithmic deprioritization. Triple penalty. One problem, three punishments.
They’re competing with everyone doing the exact same thing. When 10,000 channels are all making AI “top 10 facts” videos, the market saturates. Views split a thousand ways. Nobody wins.
They treat it as passive income. The channels that earn? Run like businesses. Consistent publishing, audience engagement, format evolution, analytics reviews. The “set and forget” AI channels earn exactly what that approach deserves: nothing.
None of this means AI channels can’t earn. It means the ones that do are run by creators who treat it seriously, pick defensible niches, and put in actual work. AI is a production-speed advantage. It’s not a substitute for strategy.
A 12-Month Curve That Isn’t Fantasy
Let me sketch what a realistic first year looks like for a serious AI creator. Not the best case. Not the worst case. The middle path:
Months 1–3: investment phase. You’re publishing, learning, and earning nothing. Your costs (tools, time) are higher than your revenue (zero). This is normal. If someone tells you they were profitable in month one, they’re either lying or they walked in with an existing audience.
Months 4–6: traction phase. A few videos start getting views. You might hit a few thousand subscribers if the content is good and the niche has demand. Revenue is still negligible — maybe enough to cover your tool subscriptions, maybe not.
Months 7–9: threshold phase. If you’ve stayed consistent and the content resonates, you’re approaching or hitting YPP thresholds. The review process takes time. Some creators get in here. Many need longer.
Months 10–12: early monetization. You’re in YPP. Revenue starts trickling in. For most channels at this point, it’s supplementary income — not rent money. The creators who eventually build something significant are the ones who keep going past this stage, reinvesting and improving.
Beyond 12 months, the range opens up wide. Some channels plateau. Some find their rhythm and grow steadily. A few break out. What separates them usually isn’t the AI tools — it’s the creator’s judgment, consistency, and willingness to adapt when things stop working.
Look, I know this isn’t the answer most people want. The honest answer never is. But I’d rather you start with clear eyes than with inflated expectations that make you quit in month three. YouTube rewards patience and punishes shortcuts. AI gives you a production advantage — so use it to make more good videos, not to skip the work of making good videos in the first place.