diff --git a/cloud/ec2 b/cloud/ec2
index 90dcd833b39..fbcb2d018d7 100644
--- a/cloud/ec2
+++ b/cloud/ec2
@@ -142,6 +142,13 @@ options:
     required: false
     default: null
     aliases: []
+  placement_group:
+    version_added: "1.3"
+    description:
+      - placement group for the instance when using EC2 Clustered Compute
+    required: false
+    default: null
+    aliases: []
   vpc_subnet_id:
     version_added: "1.1"
     description:
@@ -223,6 +230,7 @@ def main():
             ec2_url = dict(aliases=['EC2_URL']),
             ec2_secret_key = dict(aliases=['EC2_SECRET_KEY'], no_log=True),
             ec2_access_key = dict(aliases=['EC2_ACCESS_KEY']),
+            placement_group = dict(),
             user_data = dict(),
             instance_tags = dict(),
             vpc_subnet_id = dict(),
@@ -247,6 +255,7 @@ def main():
     ec2_url = module.params.get('ec2_url')
     ec2_secret_key = module.params.get('ec2_secret_key')
     ec2_access_key = module.params.get('ec2_access_key')
+    placement_group = module.params.get('placement_group')
     user_data = module.params.get('user_data')
     instance_tags = module.params.get('instance_tags')
     vpc_subnet_id = module.params.get('vpc_subnet_id')
@@ -310,7 +319,6 @@ def main():
         count_remaining = count_remaining - len(running_instances) 
     
     # Both min_count and max_count equal count parameter. This means the launch request is explicit (we want count, or fail) in how many instances we want.
-
     
     if count_remaining > 0:
         try:
@@ -320,7 +328,8 @@ def main():
                       'min_count': count_remaining, 
                       'max_count': count_remaining,
                       'monitoring_enabled': monitoring,
-                      'placement': zone, 
+                      'placement': zone,
+                      'placement_group': placement_group, 
                       'instance_type': instance_type,
                       'kernel_id': kernel,
                       'ramdisk_id': ramdisk,