diff --git a/src/nvfp4_megamoe_kernel/cutlass_nvfp4_gemm/cutlass_nvfp4_gemm.cu b/src/nvfp4_megamoe_kernel/cutlass_nvfp4_gemm/cutlass_nvfp4_gemm.cu
index f005411f..d559cda8 100644
--- a/src/nvfp4_megamoe_kernel/cutlass_nvfp4_gemm/cutlass_nvfp4_gemm.cu
+++ b/src/nvfp4_megamoe_kernel/cutlass_nvfp4_gemm/cutlass_nvfp4_gemm.cu
@@ -20,7 +20,7 @@
 #include <cutlass/gemm/kernel/tile_scheduler_params.h>
 #include <cutlass/float_subbyte.h>
 #include <cutlass/util/packed_stride.hpp>
-#include <cutlass/device_memory.h>
+#include <cutlass/util/device_memory.h>
 
 #if defined(CUTLASS_ARCH_MMA_SM100_SUPPORTED)