The evidence is that quotas exist, as seen here, and are low enough that people are hitting them regularly. When was the last time you hit your quota of Google searches? When was the last time you hit your quota of StackOverflow questions? When was the last time you hit your quota of YouTube videos? Any service will rate limit abuse, but if abuse is indistinguishable from regular use from the provider's perspective, that's not a good sign.

This article convinced me otherwise https://www.wheresyoured.at/the-subprime-ai-crisis-is-here/

Inference cannot happen without training the model first, so the distinction is quite pointless.

The majority of accounts are free - these are profitable?

IMO they need as many users before their IPO - then the changes will really begin.

Inference for API or subscriptions? There is a massive price difference between the two.

You're assuming they can just stop training. For the entirety of these companies' existence, they have done training. It is part of their price. They must keep pushing out better and better models. That's like saying Nvidia can just stop making new GPUs, they're obviously making so much money with their current models now.

source?