diff --git a/src/main/java/com/nvidia/spark/rapids/jni/Hash.java b/src/main/java/com/nvidia/spark/rapids/jni/Hash.java
index 96b66555a..2d23ae525 100644
--- a/src/main/java/com/nvidia/spark/rapids/jni/Hash.java
+++ b/src/main/java/com/nvidia/spark/rapids/jni/Hash.java
@@ -22,15 +22,15 @@
 import ai.rapids.cudf.NativeDepsLoader;
 
 public class Hash {
+  static {
+    NativeDepsLoader.loadNativeDeps();
+  }
+
   // there doesn't appear to be a useful constant in spark to reference. this could break.
   static final long DEFAULT_XXHASH64_SEED = 42;
 
   public static final int MAX_STACK_DEPTH = getMaxStackDepth();
 
-  static {
-    NativeDepsLoader.loadNativeDeps();
-  }
-
   /**
    * Create a new vector containing spark's 32-bit murmur3 hash of each row in the table.
    * Spark's murmur3 hash uses a different tail processing algorithm.