diff --git a/asv_bench/benchmarks/algorithms.py b/asv_bench/benchmarks/algorithms.py
index 436093ef195ef..7d97f2c740acb 100644
--- a/asv_bench/benchmarks/algorithms.py
+++ b/asv_bench/benchmarks/algorithms.py
@@ -2,6 +2,8 @@
 
 import numpy as np
 
+from pandas._libs import lib
+
 import pandas as pd
 from pandas.util import testing as tm
 
@@ -13,6 +15,19 @@
         pass
 
 
+class MaybeConvertObjects:
+    def setup(self):
+        N = 10 ** 5
+
+        data = list(range(N))
+        data[0] = pd.NaT
+        data = np.array(data)
+        self.data = data
+
+    def time_maybe_convert_objects(self):
+        lib.maybe_convert_objects(self.data)
+
+
 class Factorize:
 
     params = [[True, False], ["int", "uint", "float", "string"]]
diff --git a/doc/source/whatsnew/v0.25.0.rst b/doc/source/whatsnew/v0.25.0.rst
index 68ecb4c487a1e..8c472cb3121d2 100644
--- a/doc/source/whatsnew/v0.25.0.rst
+++ b/doc/source/whatsnew/v0.25.0.rst
@@ -940,7 +940,6 @@ Performance improvements
 - For :meth:`to_datetime` changed default value of cache parameter to ``True`` (:issue:`26043`)
 - Improved performance of :class:`DatetimeIndex` and :class:`PeriodIndex` slicing given non-unique, monotonic data (:issue:`27136`).
 
-
 .. _whatsnew_0250.bug_fixes:
 
 Bug fixes
diff --git a/pandas/_libs/lib.pyx b/pandas/_libs/lib.pyx
index 1df220029def6..1936404b75602 100644
--- a/pandas/_libs/lib.pyx
+++ b/pandas/_libs/lib.pyx
@@ -1955,6 +1955,7 @@ def maybe_convert_objects(ndarray[object] objects, bint try_float=0,
                 seen.timedelta_ = 1
             if not (convert_datetime or convert_timedelta):
                 seen.object_ = 1
+                break
         elif util.is_bool_object(val):
             seen.bool_ = 1
             bools[i] = val