diff --git a/sdv/metadata/multi_table.py b/sdv/metadata/multi_table.py
index 924f433ae..a5f7264e0 100644
--- a/sdv/metadata/multi_table.py
+++ b/sdv/metadata/multi_table.py
@@ -600,11 +600,10 @@ def detect_from_csvs(self, folder_name, read_csv_parameters=None):
             raise ValueError(f"No CSV files detected in the folder '{folder_name}'.")
 
         data = {}
-        read_csv_parameters = read_csv_parameters or {}
         for csv_file in csv_files:
             table_name = csv_file.stem
-            self.detect_table_from_csv(table_name, str(csv_file), read_csv_parameters)
-            data[csv_file.stem] = pd.read_csv(str(csv_file), **read_csv_parameters)
+            data[table_name] = _load_data_from_csv(csv_file, read_csv_parameters)
+            self.detect_table_from_dataframe(table_name, data[table_name])
 
         self._detect_relationships(data)
 
diff --git a/tests/unit/metadata/test_multi_table.py b/tests/unit/metadata/test_multi_table.py
index df4d9bd3b..2749e277f 100644
--- a/tests/unit/metadata/test_multi_table.py
+++ b/tests/unit/metadata/test_multi_table.py
@@ -2380,18 +2380,20 @@ def test_detect_table_from_csv_table_already_exists(self):
         with pytest.raises(InvalidMetadataError, match=error_message):
             metadata.detect_table_from_csv('table', 'path.csv')
 
-    def test_detect_from_csvs(self, tmp_path):
+    @patch('sdv.metadata.multi_table._load_data_from_csv')
+    def test_detect_from_csvs(self, load_data_mock, tmp_path):
         """Test the ``detect_from_csvs`` method.
 
         The method should call ``detect_table_from_csv`` for each csv in the folder.
         """
         # Setup
         instance = MultiTableMetadata()
-        instance.detect_table_from_csv = Mock()
+        instance.detect_table_from_dataframe = Mock()
         instance._detect_relationships = Mock()
 
         data1 = pd.DataFrame({'col1': [1, 2], 'col2': [3, 4]})
         data2 = pd.DataFrame({'col1': [5, 6], 'col2': [7, 8]})
+        load_data_mock.side_effect = [data2, data1]
 
         filepath1 = tmp_path / 'table1.csv'
         filepath2 = tmp_path / 'table2.csv'
@@ -2406,13 +2408,18 @@ def test_detect_from_csvs(self, tmp_path):
         instance.detect_from_csvs(tmp_path)
 
         # Assert
-        expected_calls = [
-            call('table1', str(filepath1), {}),
-            call('table2', str(filepath2), {}),
+        expected_calls_load_data = [
+            call(filepath1, None),
+            call(filepath2, None),
         ]
+        load_data_mock.assert_has_calls(expected_calls_load_data, any_order=True)
 
-        instance.detect_table_from_csv.assert_has_calls(expected_calls, any_order=True)
-        assert instance.detect_table_from_csv.call_count == 2
+        expected_detect_calls = [
+            call('table1', data1),
+            call('table2', data2),
+        ]
+        instance.detect_table_from_dataframe.assert_has_calls(expected_detect_calls, any_order=True)
+        assert instance.detect_table_from_dataframe.call_count == 2
 
         instance._detect_relationships.assert_called_once()
         table1 = instance._detect_relationships.call_args[0][0]['table1']