diff --git a/pandas/core/frame.py b/pandas/core/frame.py
index 53ab03696551e..018059ce2784f 100644
--- a/pandas/core/frame.py
+++ b/pandas/core/frame.py
@@ -7279,7 +7279,7 @@ def value_counts(
             subset = self.columns.tolist()
 
         name = "proportion" if normalize else "count"
-        counts = self.groupby(subset, dropna=dropna).grouper.size()
+        counts = self.groupby(subset, dropna=dropna, observed=False).grouper.size()
         counts.name = name
 
         if sort:
diff --git a/pandas/tests/frame/methods/test_value_counts.py b/pandas/tests/frame/methods/test_value_counts.py
index 355f05cd5156c..c05a929360478 100644
--- a/pandas/tests/frame/methods/test_value_counts.py
+++ b/pandas/tests/frame/methods/test_value_counts.py
@@ -175,3 +175,17 @@ def test_data_frame_value_counts_subset(nulls_fixture, columns):
     )
 
     tm.assert_series_equal(result, expected)
+
+
+def test_value_counts_categorical_future_warning():
+    # GH#54775
+    df = pd.DataFrame({"a": [1, 2, 3]}, dtype="category")
+    result = df.value_counts()
+    expected = pd.Series(
+        1,
+        index=pd.MultiIndex.from_arrays(
+            [pd.Index([1, 2, 3], name="a", dtype="category")]
+        ),
+        name="count",
+    )
+    tm.assert_series_equal(result, expected)