I agree that they all add up, and I've dropped a few this past year. I'm just pointing out that I find more value in the YouTube content I do watch than I do from other sources in terms of the time I spend watching.

It's not just the ad revenue itself that has to be replaced.

In addition to that, you have substantial costs baked into supporting a subscriber program. Customer service costs, payment processing fees, development effort to manage your subscriber system itself, incremental development and creative costs to ensure your wireframes and functionality account for the ad-free and ad-filled versions (which may include multiple form factors and platforms for content delivery), marketing related costs (be it marketing labor, development labor, conversion rate optimization labor, etc) for user acquisition and retention work, and additional complexity in ensuring harmony between what you promise in your subscriber contract and what promises you're making in your advertising/data brokerage contracts.

In addition to all that overhead, the people that pay for subscriptions are those with disposable income to pay for subscriptions. Which substantially overlaps with who advertisers want to reach. So the more valuable part of your audience pool is self-selecting out of your advertising pool, which exposes you to substantial risk of devaluing your advertising pool and the CPMs you can command overall far in excess of the value of any single individuals ad- and data- related revenue (which only gets worse as your paid subscriber base gets larger and more noticeably absent from your advertising pool).

So you take on risk at the beginning in the form of overhead, and as your subscriber base scales the overhead becomes less of an issue as it's amortized across a larger base. But your risk itself still exists, it just shifts over to the advertising side of the house where your addressable audience is becoming less and less valuable (or maybe even just less in volume, if not in value).

And after all that, if they just broke even it wouldn't be worth the risk. So in addition to just paying for your equivalent value to an advertiser that month, you're paying for that + a portion of the overhead and incremental costs it takes to support the ad-free version + a portion of the long term devaluation of the advertising pool that results from having an ad-free version + a bit of profit to ensure it was all worth the hassle to begin with.