diff --git a/python/sglang/srt/layers/quantization/w8a8_int8.py b/python/sglang/srt/layers/quantization/w8a8_int8.py
index c9af7ae29..19cf49c9b 100644
--- a/python/sglang/srt/layers/quantization/w8a8_int8.py
+++ b/python/sglang/srt/layers/quantization/w8a8_int8.py
@@ -754,6 +754,8 @@ class NPU_W8A8LinearMethod(LinearMethodBase):
         x: torch.Tensor,
         bias: Optional[torch.Tensor] = None,
     ) -> torch.Tensor:
+        from sglang.srt.layers.linear import RowParallelLinear
+
         if isinstance(layer, RowParallelLinear):
             tp_rank = get_tensor_model_parallel_rank()
             return self.quant_method.apply(layer, x, bias, tp_rank)