How to Make Money With an AI YouTube Channel (Real Numbers)

An AI-built YouTube channel called Boring History made $24,559.78 in its last 90 days, shown live in the revenue tab, at a $29.11 RPM, with no face, no filming, and no editing by hand. That number, not the six-figure headlines, is the honest anchor for what an AI channel can do. It comes from this full walkthrough by TubeGen’s founder, who helped build the channel and opens the analytics on screen. The channel is owned by two people in his mentorship program and was built end to end on TubeGen. Here’s the actual model behind it, numbers intact.

Why the numbers look strange (and why that’s the point)

Boring History had roughly 900,000 views but made $24,559 in 90 days, a huge earning rate for that view count. The reason is RPM. One video with 11,000 views made $323, because its RPM was $29.11 per thousand. That’s the entire thesis. In the right niche, views matter far less than what each thousand of them is worth.

Two things drive that high RPM here. The niche (sleep-history) attracts an audience advertisers value, and the videos are long — two to three hours each, which means more ad slots per view. Long, high-value, and repeatable is the combination that pays, and it’s the opposite of chasing raw view counts in a cheap niche.

The repeatability that makes it work

Scroll the channel and every video looks similar: same art style, same text placement, same color scheme, same title pattern, similar length. That sameness isn’t laziness. It’s the model. Repeatable format means endless video ideas, and it means the AI can study which titles and thumbnails perform, then generate more like the winners. (Worth noting: this same repeatability, pushed too far, is what triggers reused-content demonetization — so vary enough to stay authentic, which we cover in the demonetization breakdown.)

The two mistakes that kill AI channels

Per the founder, two errors sink most AI channels. The first is under-posting. When AI generates everything, the marginal cost of another video is tiny, so posting once every three days wastes the whole advantage. Easy niches attract competition, and the counter is volume: out-publish everyone. The second is publishing content that can’t monetize. The July 2025 YouTube update scared creators, but it targeted spam, not AI. The fix is simply quality: a voiceover that doesn’t sound like robotic text-to-speech, and production that isn’t low-effort. Keep it high-quality and monetization isn’t the problem people fear.

Finding the niche, then validating it

The model starts with finding a proven, recent channel to model, not inventing a niche cold. TubeGen’s Niche Finder, a searchable database of channels with an AI sleep-niche filter, surfaces candidates fast. The walkthrough pulls up three (Universe Before Sleep, Echo 7, The World Before Dawn) in seconds.

Then validate with real data before committing. Using the free SocialBlade (or ViewStats), you check each channel’s daily views and estimate revenue at a comparable RPM: 1,500 views a day at a $29 RPM is roughly $43 a day, about $1,300 a month, which isn’t worth it. A channel doing 12,300 views a day pencils out near $10,400 a month. You want the channel with the most views, the fewest videos, and the newest start date, since that signals a niche that’s working and not yet saturated. The World Before Dawn won that comparison on all three.

The build, end to end

Once you’ve picked a channel to model, the production is one connected flow. In TubeGen, Copy Style (a one-click tool that downloads the reference channel’s transcripts and trains the script writer to match its style) sets up the whole project, auto-selecting a fitting voice, language, and word count. The script writer then drafts in that style; a 12,900-word script generates in about five minutes. The voiceover, partnered with InWorld for natural narration, reads it, and a title comes from the generate-title-by-channel tool, which bases options on the reference channel’s best performers.

Visuals are where the timing happens. In the visuals advanced mode, you split the timeline at the script’s natural sentence breaks and treat each segment differently. Animate the first 40 seconds, then run a fresh image roughly every 35 seconds for the rest, matching the reference channel’s pattern. TubeGen averages the images to that interval while snapping to natural breaks. The editor imports everything already timed to the script, where you add zoom-ins and crossfades and render.

The cost is concrete: that example 18-minute video came to about $11.24 in credits, and animations were most of it, roughly $6 to $7 of the $11. Drop the animations and use lowest-quality static images (which, thanks to the model partners, still look fine) and the same video runs closer to $2. Finally, the thumbnail generator takes a reference thumbnail, applies the changes you describe, and outputs a matching design, though for a channel you plan to scale, varying thumbnails matters for staying monetizable.

What’s real here, and what isn’t

Keep the claims straight. The $24,559 in 90 days and the ~$10,000 a month are one channel’s reported figures, shown on screen, not a typical result and not something the tooling guarantees. The $29.11 RPM and the $323 video are that channel’s analytics, not a TubeGen platform metric. Your RPM depends on your niche and audience. What’s safe to say: Boring History was built entirely with AI on TubeGen, and the model (high-RPM niche, long repeatable videos, daily posting, quality production) is replicable. The earnings are not promised.

Before you start, the channel still has to clear YouTube’s bar; the monetization requirements guide covers the thresholds, and the best niches breakdown ranks lanes by the RPM economics that make this model work.

What it actually costs to run

The economics only work if you know the real per-video cost, and the walkthrough is specific about it. At TubeGen’s best rate, roughly 457 credits cost a dollar, and the example 18-minute video used about 5,000 credits, landing near $11.24. Animations drove most of that spend, so the single biggest cost lever is how much of the video you animate versus leave as timed static images. A creator optimizing for margin animates only the opening stretch, then runs still images the rest of the way, which is exactly what the modeled channel did.

Scale that to a two-to-three-hour daily upload and each video might run anywhere from about $50 to $150 depending on animation and image-quality settings. That sounds steep until you set it against a channel reportedly earning around $10,000 a month: the production cost is a rounding error against the revenue when the niche and RPM are right. Get the niche wrong, though, and that same spend produces videos nobody pays to advertise against, which is why niche validation comes before the first render.

The short version

Making money with an AI YouTube channel isn’t about volume of views. It’s about a high-RPM niche, long repeatable videos, daily uploads, and production clean enough to monetize. One channel ran that exact model to $24,559 in 90 days. The results aren’t guaranteed, but the method is repeatable.

Build your channel on the same system Boring History used. Start with TubeGen →