diff --git a/python/sglang/srt/layers/moe/cutlass_w4a8_moe.py b/python/sglang/srt/layers/moe/cutlass_w4a8_moe.py
index 0bf89b2d0..bed37dd78 100644
--- a/python/sglang/srt/layers/moe/cutlass_w4a8_moe.py
+++ b/python/sglang/srt/layers/moe/cutlass_w4a8_moe.py
@@ -3,7 +3,18 @@
 from typing import Optional
 
 import torch
-from sgl_kernel import cutlass_w4a8_moe_mm, get_cutlass_w4a8_moe_mm_data, silu_and_mul
+
+from sglang.srt.utils import is_cuda_alike
+
+_is_cuda_alike = is_cuda_alike()
+
+if _is_cuda_alike:
+    from sgl_kernel import (
+        cutlass_w4a8_moe_mm,
+        get_cutlass_w4a8_moe_mm_data,
+    )
+
+from sgl_kernel import silu_and_mul
 
 from sglang.jit_kernel.per_tensor_quant_fp8 import per_tensor_quant_fp8
 from sglang.srt.distributed import get_moe_expert_parallel_world_size