diff --git a/onnxruntime/core/providers/migraphx/gpu_data_transfer.cc b/onnxruntime/core/providers/migraphx/gpu_data_transfer.cc
index 77c5e18a5878e..a643f6b208f94 100644
--- a/onnxruntime/core/providers/migraphx/gpu_data_transfer.cc
+++ b/onnxruntime/core/providers/migraphx/gpu_data_transfer.cc
@@ -67,6 +67,9 @@ common::Status GPUDataTransfer::CopyTensorAsync(const Tensor& src, Tensor& dst,
     } else if (src_device.Type() == OrtDevice::GPU) {
       // copying between GPU, this is non-blocking
       HIP_CALL_THROW(hipMemcpyAsync(dst_data, src_data, bytes, hipMemcpyDeviceToDevice, static_cast<hipStream_t>(stream.GetHandle())));
+    } else {
+      // copy from other CPU memory to GPU, this is blocking
+      HIP_CALL_THROW(hipMemcpyWithStream(dst_data, src_data, bytes, hipMemcpyHostToDevice, static_cast<hipStream_t>(stream.GetHandle())));
     }
   } else if (src_device.Type() == OrtDevice::GPU) {
     // If dest are not pinned, the memory copy will be performed synchronously.