BUG: series sum changes dtype uint64 to int64

doc/source/whatsnew/v2.1.0.rst

@@ -351,6 +351,7 @@ Numeric
 - Bug in :meth:`Series.any`, :meth:`Series.all`, :meth:`DataFrame.any`, and :meth:`DataFrame.all` had the default value of ``bool_only`` set to ``None`` instead of ``False``; this change should have no impact on users (:issue:`53258`)
 - Bug in :meth:`Series.corr` and :meth:`Series.cov` raising ``AttributeError`` for masked dtypes (:issue:`51422`)
 - Bug in :meth:`Series.median` and :meth:`DataFrame.median` with object-dtype values containing strings that can be converted to numbers (e.g. "2") returning incorrect numeric results; these now raise ``TypeError`` (:issue:`34671`)
+- Bug in :meth:`Series.sum` converting dtype ``uint64`` to ``int64`` (:issue:`53401`)
 
 
 Conversion

pandas/core/nanops.py

@@ -324,8 +324,10 @@ def _get_values(
 def _get_dtype_max(dtype: np.dtype) -> np.dtype:
     # return a platform independent precision dtype
     dtype_max = dtype
-    if dtype.kind in "biu":
+    if dtype.kind in "bi":
         dtype_max = np.dtype(np.int64)
+    elif dtype.kind == "u":
+        dtype_max = np.dtype(np.uint64)
     elif dtype.kind == "f":
         dtype_max = np.dtype(np.float64)
     return dtype_max

pandas/tests/reductions/test_reductions.py

@@ -1110,6 +1110,13 @@ def test_idxminmax_with_inf(self):
             assert s.idxmax() == 0
             np.isnan(s.idxmax(skipna=False))
 
+    def test_sum_uint64(self):
+        # GH 53401
+        s = Series([10000000000000000000], dtype="uint64")
+        result = s.sum()
+        expected = np.uint64(10000000000000000000)
+        tm.assert_almost_equal(result, expected)
+
 
 class TestDatetime64SeriesReductions:
     # Note: the name TestDatetime64SeriesReductions indicates these tests