AI YouTube Channels: Real Earnings Data From People Who Tried

YouTube is a real business. It is also one of the most overhyped career pivots in circulation. Both things are true, and the gap between them is where most people lose time and money.

Jonathan Laramy quit his customer service job in June 2025 to make AI-generated history videos full-time. He had no filmmaking background. His channel, Chloe VS History, now has 381,000 subscribers, and individual videos regularly pull 271,000 to 528,000 views. By any reasonable definition, it's working. Here's what the profile pieces don't lead with: a single long-form episode takes weeks to produce, costs up to $1,070 in AI generation fees alone, and requires 10 to 15 revision rounds per scene. "People don't realize how expensive they are to produce," Laramy said. He's right. And that's before you get to the part where most people who try this earn almost nothing.

If you're weighing whether to build a YouTube-based side hustle or career pivot, read this before you buy any equipment or sign up for any course.

The Real Shape of the Market

The platform is genuinely large and genuinely paying creators—but the money concentrates at the top of a very steep curve.

YouTube paid more than $32 billion to partners globally in 2024, and more than 3 million people earn through the platform. That's real money moving through a real system. It proves the opportunity exists. What it doesn't prove is that the opportunity is evenly distributed.

A peer-reviewed academic study of YouTube's historical data found that roughly 85% of all views go to about 3% of channels. That's not an accident of content quality—it's how recommendation systems work. Scarce attention flows to a small number of channels that have already captured it, making the discovery problem harder for everyone entering the market.

The income picture is equally sobering. A 2025 survey of more than 3,000 creators found that 57% of self-described full-time creators earned below a $44,000 U.S. living wage from content alone. "Full-time creator" is a job title you can give yourself before the income supports it. And among YouTube Partner Program channels that earned five figures or more in 2024, over 50% earned revenue from sources other than ads and YouTube Premium. The people doing best aren't relying on AdSense as their primary engine.

This doesn't require a filmmaking background to understand—it's the same concentration dynamic that applies to Etsy shops, freelance consulting, or any other platform where a small number of operators dominate recommendation surfaces. The relevant skill is figuring out how to not be in the invisible 97%.

Which raises the question the success stories almost never answer honestly: between two people who both use AI tools end-to-end, why does one build 381,000 subscribers while the other gets 74 views and quits?

What Actually Separates the Results

The tools are not the differentiator. Two people can run an identical AI production stack and get radically different results, because what happens before the generation is what determines whether viewers have a reason to come back.

Consider DrowningInMayo, an anonymous Reddit creator who posted a detailed account of a failed experiment. The workflow looked complete: ChatGPT for scripts, ElevenLabs for voice, Pexels for stock footage, CapCut for editing, Canva for thumbnails. The niche was AI productivity—a category with real search volume. Eleven videos went up in three weeks. The best result was 74 views. Most were in the 20s or lower. Zero subscribers. The creator's own diagnosis: "The AI voiceovers sounded off. Flat. The scripts didn't hook people. No emotion, no personality. Just generic tips."

Anyone can sign up for the software and generate images or videos. The difficult part is figuring out what people actually want to watch and then building a process that consistently delivers it.
— Jonathan Laramy, Creator, Chloe VS History

Laramy's workflow includes Claude for ideation, multiple AI image and video tools, and an AI voice model for character consistency—a comparable stack, arguably more complex. But Laramy reports spending more time refining his character's personality and dialogue than generating her appearance. Every scene gets reviewed, adjusted, and regenerated until it works. The generation is the raw material. His judgment is the product.

His own framing makes the variable explicit: "Anyone can sign up for the software and generate images or videos. The difficult part is figuring out what people actually want to watch and then building a process that consistently delivers it."

AI tools lower the barrier to making a video. They don't lower the barrier to making a video someone would choose to watch over the alternatives. That still requires a specific audience, a clear promise, and a repeatable reason to return.

If you're planning to use AI to build a YouTube channel, the honest question to answer first is: what specific thing do I understand—about a topic, a person, a problem, a format—that will make my version better than a prompt? If the answer is "nothing yet," that's where to start. Not in the tool stack.

But even when that editorial judgment exists, there's a harder economic reality that catches serious operators by surprise long after they've passed YouTube's monetization threshold.

Run the Math Before You Commit

Reaching YouTube's monetization threshold and running a profitable YouTube business are completely separate events. Most creator content deliberately obscures that gap.

Evan Porter ran a faceless channel converting articles from his existing website into voiceover videos with stock footage. He published consistently for nearly two years before enabling ads in March 2022. His first two monetized months produced $411 and $427, on roughly 65,000 views each. His ongoing production cost was about $300 per video, two videos per month—$600 monthly. His own conclusion: "Even though the channel is monetized and making half-decent money, it's still not profitable." The channel passed YouTube's threshold and still failed the business test.

The arithmetic explains why. At a $6 to $7 RPM—Porter's reported niche rate—covering $600 per month in production costs alone requires roughly 90,000 to 100,000 monthly monetized views. That's before your own labor, software subscriptions, or taxes. At Laramy's production costs of up to $1,070 per long-form video, the break-even math looks significantly worse per episode.

You can follow someone else's blueprint, but if it doesn't align with who you are and how you want to operate, the journey will always feel crippling — even if the cage is golden.
— Vanessa Lau, YouTube Creator

There's an escape route, and it's the one the successful operators actually found. Among YPP channels earning five figures or more in 2024, over 50% earned revenue beyond ads—the same statistic from earlier, now read differently. A channel that generates five qualified consulting leads per month at 3,000 views may be a better business than one chasing 90,000 entertainment views, because each lead is worth $1,000 to $5,000, not fractions of a cent.

This math is the same whether you're a marketer, an HR professional, a software developer, or a customer service worker considering this pivot. The niche and RPM will vary, but the calculation is identical: total cost divided by revenue per view equals required monthly views. Run that number before you spend anything.

The business model question has to come before the production question. If your plan is AdSense to income, you need an unusually large and consistent view volume. If your plan is YouTube to customer to higher-value transaction, the required audience can be much smaller and the path is measurably faster.

That calculation is where the real decision lives—and it's the one most YouTube course sellers have no incentive to walk you through honestly.

What to Do Before You Invest a Dollar

Return to Laramy's core insight: the channel works not because the AI tools are exceptional—anyone can access them—but because he spent the most time on the question that comes before generation. What does a specific person want to watch, and how do you build a process that reliably delivers it? That question is available to anyone. The tools are not the barrier.

YouTube is not a content business for most people who succeed at it. It's a customer-acquisition business, a lead-generation business, or a product-distribution business wearing a content business's clothes. The channels that generate real income treat the video as the top of a funnel, not as the product itself. That reframe changes what you optimize for, what audience you need, and how quickly you can find out whether the thing is working.

Before investing in equipment, software, or production: publish three test videos on an existing channel, or create one and publish three. Track two numbers after 30 days—click-through rate and average view duration. If neither metric is meaningful after three uploads with honest thumbnails and genuine content, the problem is the premise, not the production quality. Fix the premise before you scale anything.

The question isn't whether you can make YouTube videos. It's whether you have something specific that a specific person needs to watch—and that answer is available before you spend a dollar.

Recommended Tools & Resources