diff --git a/test/quantization/test_moe_quant.py b/test/quantization/test_moe_quant.py
index 842468a769..15039af798 100644
--- a/test/quantization/test_moe_quant.py
+++ b/test/quantization/test_moe_quant.py
@@ -1,5 +1,6 @@
 import unittest
 
+import pytest
 import torch
 from parameterized import parameterized
 
@@ -32,6 +33,12 @@
     is_sm_at_least_90,
 )
 
+if torch.version.hip is not None:
+    pytest.skip(
+        "ROCm support for MoE quantization is under development",
+        allow_module_level=True,
+    )
+
 
 class TestMoEQuantCompile(unittest.TestCase):
     DEFAULT_PARAMS = (512, 256, 8, 2)  # hidden_dim, expert_dim, num_experts, top_k