A VPN is only a single end point just like your ISP meaning you are only shifting the problem to your VPN provider who admittedly is more trustworthy than your ISP but you are still putting an immense amount of trust into a single point of failure.

If you truly want to hide from your ISP or really anyone, your only options are to use TOR or I2P where your traffic is encrypted and tumbled through multiple servers.

So your threat model assumes an actor with a quantum computer capable of breaking RSA, but not a regular computer capable of filtering by IP address?