After setting up a new Compute Engine instance in zone us-central1-b, you encounter an error when attempting to attach a GPU requested by your data engineers. What is the most probable reason for this error?