diff --git a/tests/kernels/moe/test_cutedsl_moe.py b/tests/kernels/moe/test_cutedsl_moe.py
index 66a97b48b..bca3eba0f 100644
--- a/tests/kernels/moe/test_cutedsl_moe.py
+++ b/tests/kernels/moe/test_cutedsl_moe.py
@@ -17,7 +17,7 @@ from flashinfer import fp4_quantize
 from torch.nn import functional as F
 
 from vllm.model_executor.layers.activation import SiluAndMul
-from vllm.model_executor.layers.fused_moe.flashinfer_cutedsl_moe import (
+from vllm.model_executor.layers.fused_moe.experts.flashinfer_cutedsl_moe import (
     flashinfer_cutedsl_moe_masked,
 )
 from vllm.utils.flashinfer import (
diff --git a/vllm/model_executor/layers/fused_moe/flashinfer_cutedsl_moe.py b/vllm/model_executor/layers/fused_moe/experts/flashinfer_cutedsl_moe.py
similarity index 100%
rename from vllm/model_executor/layers/fused_moe/flashinfer_cutedsl_moe.py
rename to vllm/model_executor/layers/fused_moe/experts/flashinfer_cutedsl_moe.py
diff --git a/vllm/model_executor/layers/fused_moe/oracle/nvfp4.py b/vllm/model_executor/layers/fused_moe/oracle/nvfp4.py
index 031aca388..35451e87d 100644
--- a/vllm/model_executor/layers/fused_moe/oracle/nvfp4.py
+++ b/vllm/model_executor/layers/fused_moe/oracle/nvfp4.py
@@ -86,7 +86,7 @@ def backend_to_kernel_cls(
         return [FlashInferExperts]
 
     elif backend == NvFp4MoeBackend.FLASHINFER_CUTEDSL:
-        from vllm.model_executor.layers.fused_moe.flashinfer_cutedsl_moe import (
+        from vllm.model_executor.layers.fused_moe.experts.flashinfer_cutedsl_moe import (  # noqa: E501
             FlashInferCuteDSLExperts,
         )