diff --git a/doc/source/whatsnew/v2.0.0.rst b/doc/source/whatsnew/v2.0.0.rst
index 4d5c676e941ea..bac404496fa53 100644
--- a/doc/source/whatsnew/v2.0.0.rst
+++ b/doc/source/whatsnew/v2.0.0.rst
@@ -348,6 +348,7 @@ Other API changes
 - Changed behavior of :class:`Index` constructor with sequence containing at least one ``NaT`` and everything else either ``None`` or ``NaN`` to infer ``datetime64[ns]`` dtype instead of ``object``, matching :class:`Series` behavior (:issue:`49340`)
 - :func:`read_stata` with parameter ``index_col`` set to ``None`` (the default) will now set the index on the returned :class:`DataFrame` to a :class:`RangeIndex` instead of a :class:`Int64Index` (:issue:`49745`)
 - Changed behavior of :class:`Index` constructor with an object-dtype ``numpy.ndarray`` containing all-``bool`` values or all-complex values, this will now retain object dtype, consistent with the :class:`Series` behavior (:issue:`49594`)
+- Changed behavior of :meth:`DataFrame.shift` with ``axis=1``, an integer ``fill_value``, and homogeneous datetime-like dtype, this now fills new columns with integer dtypes instead of casting to datetimelike (:issue:`49842`)
 - :meth:`DataFrame.values`, :meth:`DataFrame.to_numpy`, :meth:`DataFrame.xs`, :meth:`DataFrame.reindex`, :meth:`DataFrame.fillna`, and :meth:`DataFrame.replace` no longer silently consolidate the underlying arrays; do ``df = df.copy()`` to ensure consolidation (:issue:`49356`)
 -
 
diff --git a/pandas/core/frame.py b/pandas/core/frame.py
index ea7a2096a7d21..f3b55533919dd 100644
--- a/pandas/core/frame.py
+++ b/pandas/core/frame.py
@@ -5720,14 +5720,7 @@ def shift(
                 #  keep the same dtype (i.e. the _can_hold_element check)
                 #  then we can go through the reindex_indexer path
                 #  (and avoid casting logic in the Block method).
-                #  The exception to this (until 2.0) is datetimelike
-                #  dtypes with integers, which cast.
                 not can_hold_element(arrays[0], fill_value)
-                # TODO(2.0): remove special case for integer-with-datetimelike
-                #  once deprecation is enforced
-                and not (
-                    lib.is_integer(fill_value) and needs_i8_conversion(arrays[0].dtype)
-                )
             ):
                 # GH#35488 we need to watch out for multi-block cases
                 # We only get here with fill_value not-lib.no_default
diff --git a/pandas/tests/frame/methods/test_shift.py b/pandas/tests/frame/methods/test_shift.py
index c44c0bd566585..4a207eac9fa80 100644
--- a/pandas/tests/frame/methods/test_shift.py
+++ b/pandas/tests/frame/methods/test_shift.py
@@ -500,7 +500,6 @@ def test_datetime_frame_shift_with_freq_error(
         with pytest.raises(ValueError, match=msg):
             no_freq.shift(freq="infer")
 
-    @td.skip_array_manager_not_yet_implemented  # TODO(ArrayManager) axis=1 support
     def test_shift_dt64values_int_fill_deprecated(self):
         # GH#31971
         ser = Series([pd.Timestamp("2020-01-01"), pd.Timestamp("2020-01-02")])
@@ -516,17 +515,15 @@ def test_shift_dt64values_int_fill_deprecated(self):
         df2 = DataFrame({"A": ser, "B": ser})
         df2._consolidate_inplace()
 
-        with pytest.raises(TypeError, match="value should be a"):
-            df2.shift(1, axis=1, fill_value=0)
+        result = df2.shift(1, axis=1, fill_value=0)
+        expected = DataFrame({"A": [0, 0], "B": df2["A"]})
+        tm.assert_frame_equal(result, expected)
 
-        # same thing but not consolidated
-        # This isn't great that we get different behavior, but
-        #  that will go away when the deprecation is enforced
+        # same thing but not consolidated; pre-2.0 we got different behavior
         df3 = DataFrame({"A": ser})
         df3["B"] = ser
         assert len(df3._mgr.arrays) == 2
         result = df3.shift(1, axis=1, fill_value=0)
-        expected = DataFrame({"A": [0, 0], "B": df2["A"]})
         tm.assert_frame_equal(result, expected)
 
     @pytest.mark.parametrize(