What's interesting about this argument is that nothing is stopping you from doing this same thing using JWTs. Just generate your token and store it as a claim in the JWT. You can check the revocation of the stored token without even validating that the token is genuine, if that is fortuitous. What this buys you is the ability to attach clear text information with the token, and, if you are doing asymmetric validation, the ability to validate both the security credential and the included plaintext information in an untrusted (client side) setting.

I actually think the revocation argument is the over-engineering case here.

I'd argue that people who need to avoid hitting their database on every request outnumber people who need sub-minute revocation

I don't see how this solution means you cannot use JWTs.