diff --git a/gemma/ops.h b/gemma/ops.h
index 2b5dc39..1988bd8 100644
--- a/gemma/ops.h
+++ b/gemma/ops.h
@@ -149,6 +149,8 @@ HWY_INLINE void MatVecAddLoop(const ArrayT& mat, const size_t mat_ofs,
   }
 }
 
+
+#if !defined(HWY_NATIVE_DOT_BF16) || !HWY_NATIVE_DOT_BF16
 template <bool kAdd, size_t kOuter, size_t kInner, typename VecT, typename AddT,
           size_t kCapacity>
 HWY_INLINE void MatVecAddLoop(
@@ -174,6 +176,7 @@ HWY_INLINE void MatVecAddLoop(
     }
   }
 }
+#endif
 
 template <size_t kOuter, size_t kInner, typename ArrayT, typename VecT>
 HWY_INLINE void MatVecLoop(const ArrayT& mat, const size_t mat_ofs,
@@ -344,6 +347,7 @@ HWY_INLINE void MatVecAdd(const ArrayT& mat, const size_t mat_ofs,
   constexpr size_t kRowsPerStrip = RowsPerStrip<kOuter>();
   constexpr size_t kNumStrips = kOuter / kRowsPerStrip;
 
+  #if !defined(HWY_NATIVE_DOT_BF16) || !HWY_NATIVE_DOT_BF16
   if constexpr (
     CompressTraits<typename ArrayT::value_type>::kSupportsEvenOdd
     && hwy::IsSameEither<VecT, float, hwy::bfloat16_t>()
@@ -352,10 +356,12 @@ HWY_INLINE void MatVecAdd(const ArrayT& mat, const size_t mat_ofs,
     ToEvenOddF32(vec_aligned, kInner, vec_dequant.get());
     detail::MatVecAddInner<true, kAdd, kOuter, kInner>(
       mat, mat_ofs, vec_dequant.get(), add, out, pool);
-  } else {
-    detail::MatVecAddInner<false, kAdd, kOuter, kInner>(
-      mat, mat_ofs, vec_aligned, add, out, pool);
+    return;
   }
+  #endif
+
+  detail::MatVecAddInner<false, kAdd, kOuter, kInner>(
+    mat, mat_ofs, vec_aligned, add, out, pool);
 }
 
 template <size_t kOuter, size_t kInner, typename ArrayT, typename VecT>