The secure element is responsible for gating retries, and like all HSMs is designed specifically to prevent tampering, so everything - including retry counts and delays - is theoretically rendered untamperable.

I am aware of two bugs in that logic over the years - I can’t find the articles off the top of my head. One was essentially a TOCTOU bug that could be triggered via voltage spikes to reset the device after you tried to unlock but before it updated the retry count. The other required imaging and restoring the flash between each attempt. I don’t know how that was fixed, but it should hopefully be obvious that That is going to take more than a minute to brute force a 6 digit passcode.