diff --git a/website/source/docs/providers/aws/r/autoscaling_group.html.markdown b/website/source/docs/providers/aws/r/autoscaling_group.html.markdown
index 552ff8d570..73715ee912 100644
--- a/website/source/docs/providers/aws/r/autoscaling_group.html.markdown
+++ b/website/source/docs/providers/aws/r/autoscaling_group.html.markdown
@@ -45,7 +45,8 @@ The following arguments are supported:
 * `max_size` - (Required) The maximum size of the auto scale group.
 * `min_size` - (Required) The minimum size of the auto scale group.
     (See also [Waiting for Capacity](#waiting-for-capacity) below.)
-* `availability_zones` - (Required) A list of AZs to launch resources in.
+* `availability_zones` - (Optional) A list of AZs to launch resources in.
+   Required only if you do not specify any `vpc_zone_identifier`
 * `launch_configuration` - (Required) The ID of the launch configuration to use.
 * `health_check_grace_period` - (Optional) Time after instance comes into service before checking health.
 * `health_check_type` - (Optional) "EC2" or "ELB". Controls how health checking is done.