Not true for GPU instances, they're stuck 5 minutes in a stopping state because they run some GPU health checks.
Not true for GPU instances, they're stuck 5 minutes in a stopping state because they run some GPU health checks.