diff --git a/pandas/core/dtypes/cast.py b/pandas/core/dtypes/cast.py
index 12974d56dacdc..4627100161b40 100644
--- a/pandas/core/dtypes/cast.py
+++ b/pandas/core/dtypes/cast.py
@@ -241,10 +241,6 @@ def maybe_downcast_numeric(result, dtype: DtypeObj, do_round: bool = False):
         # e.g. SparseDtype has no itemsize attr
         return result
 
-    if isinstance(result, list):
-        # reached via groupby.agg._ohlc; really this should be handled earlier
-        result = np.array(result)
-
     def trans(x):
         if do_round:
             return x.round()
diff --git a/pandas/core/groupby/groupby.py b/pandas/core/groupby/groupby.py
index 23f0e178130be..947f18901775b 100644
--- a/pandas/core/groupby/groupby.py
+++ b/pandas/core/groupby/groupby.py
@@ -51,7 +51,7 @@ class providing the base-class of operations.
 from pandas.errors import AbstractMethodError
 from pandas.util._decorators import Appender, Substitution, cache_readonly, doc
 
-from pandas.core.dtypes.cast import maybe_downcast_to_dtype
+from pandas.core.dtypes.cast import maybe_downcast_numeric
 from pandas.core.dtypes.common import (
     ensure_float,
     is_bool_dtype,
@@ -1178,7 +1178,7 @@ def _python_agg_general(self, func, *args, **kwargs):
             key = base.OutputKey(label=name, position=idx)
 
             if is_numeric_dtype(obj.dtype):
-                result = maybe_downcast_to_dtype(result, obj.dtype)
+                result = maybe_downcast_numeric(result, obj.dtype)
 
             if self.grouper._filter_empty_groups:
                 mask = counts.ravel() > 0
@@ -1188,7 +1188,7 @@ def _python_agg_general(self, func, *args, **kwargs):
                 if is_numeric_dtype(values.dtype):
                     values = ensure_float(values)
 
-                result = maybe_downcast_to_dtype(values[mask], result.dtype)
+                    result = maybe_downcast_numeric(values[mask], result.dtype)
 
             output[key] = result
 
diff --git a/pandas/core/indexes/interval.py b/pandas/core/indexes/interval.py
index 2f86d9c20bfe8..0c5fbef6cb087 100644
--- a/pandas/core/indexes/interval.py
+++ b/pandas/core/indexes/interval.py
@@ -20,7 +20,7 @@
     find_common_type,
     infer_dtype_from_scalar,
     maybe_box_datetimelike,
-    maybe_downcast_to_dtype,
+    maybe_downcast_numeric,
 )
 from pandas.core.dtypes.common import (
     ensure_platform_int,
@@ -1274,7 +1274,7 @@ def interval_range(
         breaks = np.linspace(start, end, periods)
         if all(is_integer(x) for x in com.not_none(start, end, freq)):
             # np.linspace always produces float output
-            breaks = maybe_downcast_to_dtype(breaks, "int64")
+            breaks = maybe_downcast_numeric(breaks, np.dtype("int64"))
     else:
         # delegate to the appropriate range function
         if isinstance(endpoint, Timestamp):
diff --git a/pandas/core/tools/numeric.py b/pandas/core/tools/numeric.py
index 4af32b219d380..dd7373927ed9b 100644
--- a/pandas/core/tools/numeric.py
+++ b/pandas/core/tools/numeric.py
@@ -2,7 +2,7 @@
 
 from pandas._libs import lib
 
-from pandas.core.dtypes.cast import maybe_downcast_to_dtype
+from pandas.core.dtypes.cast import maybe_downcast_numeric
 from pandas.core.dtypes.common import (
     ensure_object,
     is_datetime_or_timedelta_dtype,
@@ -180,8 +180,9 @@ def to_numeric(arg, errors="raise", downcast=None):
         if typecodes is not None:
             # from smallest to largest
             for dtype in typecodes:
-                if np.dtype(dtype).itemsize <= values.dtype.itemsize:
-                    values = maybe_downcast_to_dtype(values, dtype)
+                dtype = np.dtype(dtype)
+                if dtype.itemsize <= values.dtype.itemsize:
+                    values = maybe_downcast_numeric(values, dtype)
 
                     # successful conversion
                     if values.dtype == dtype: