diff --git a/superset/commands/database/uploaders/base.py b/superset/commands/database/uploaders/base.py
index 18b4f8024f4..f8e2b389d84 100644
--- a/superset/commands/database/uploaders/base.py
+++ b/superset/commands/database/uploaders/base.py
@@ -159,6 +159,12 @@ class UploadCommand(BaseCommand):
         if not self._model:
             return
 
+        self._table_name, self._schema = (
+            self._model.db_engine_spec.normalize_table_name_for_upload(
+                self._table_name, self._schema
+            )
+        )
+
         self._reader.read(self._file, self._model, self._table_name, self._schema)
 
         sqla_table = (
diff --git a/superset/db_engine_specs/base.py b/superset/db_engine_specs/base.py
index 4113a3e8fe5..396c3805acd 100644
--- a/superset/db_engine_specs/base.py
+++ b/superset/db_engine_specs/base.py
@@ -1295,6 +1295,26 @@ class BaseEngineSpec:  # pylint: disable=too-many-public-methods
 
         return None
 
+    @classmethod
+    def normalize_table_name_for_upload(
+        cls,
+        table_name: str,
+        schema_name: str | None = None,
+    ) -> tuple[str, str | None]:
+        """
+        Normalize table and schema names for file upload.
+
+        Some databases (e.g., Redshift) fold unquoted identifiers to lowercase,
+        which can cause issues when the upload creates a table with one case
+        but metadata operations use a different case. Override this method
+        to normalize names according to database-specific rules.
+
+        :param table_name: The table name to normalize
+        :param schema_name: The schema name to normalize (optional)
+        :return: Tuple of (normalized_table_name, normalized_schema_name)
+        """
+        return table_name, schema_name
+
     @classmethod
     def df_to_sql(
         cls,
diff --git a/superset/db_engine_specs/redshift.py b/superset/db_engine_specs/redshift.py
index df8ee834fc4..ea49c479dea 100644
--- a/superset/db_engine_specs/redshift.py
+++ b/superset/db_engine_specs/redshift.py
@@ -183,6 +183,24 @@ class RedshiftEngineSpec(BasicParametersMixin, PostgresBaseEngineSpec):
         ),
     }
 
+    @classmethod
+    def normalize_table_name_for_upload(
+        cls,
+        table_name: str,
+        schema_name: str | None = None,
+    ) -> tuple[str, str | None]:
+        """
+        Redshift folds unquoted identifiers to lowercase.
+
+        :param table_name: The table name to normalize
+        :param schema_name: The schema name to normalize (optional)
+        :return: Tuple of (normalized_table_name, normalized_schema_name)
+        """
+        return (
+            table_name.lower(),
+            schema_name.lower() if schema_name else None,
+        )
+
     @classmethod
     def df_to_sql(
         cls,
diff --git a/tests/unit_tests/db_engine_specs/test_redshift.py b/tests/unit_tests/db_engine_specs/test_redshift.py
index 77022809075..d733767503f 100644
--- a/tests/unit_tests/db_engine_specs/test_redshift.py
+++ b/tests/unit_tests/db_engine_specs/test_redshift.py
@@ -20,6 +20,7 @@ from typing import Optional
 
 import pytest
 
+from superset.db_engine_specs.redshift import RedshiftEngineSpec
 from tests.unit_tests.db_engine_specs.utils import assert_convert_dttm
 from tests.unit_tests.fixtures.common import dttm  # noqa: F401
 
@@ -49,3 +50,32 @@ def test_convert_dttm(
     )
 
     assert_convert_dttm(spec, target_type, expected_result, dttm)
+
+
+@pytest.mark.parametrize(
+    "table_name,schema_name,expected_table,expected_schema",
+    [
+        ("BPO_mytest_2", "MySchema", "bpo_mytest_2", "myschema"),
+        ("MY_TABLE", None, "my_table", None),
+        ("already_lower", "lower_schema", "already_lower", "lower_schema"),
+    ],
+)
+def test_normalize_table_name_for_upload(
+    table_name: str,
+    schema_name: Optional[str],
+    expected_table: str,
+    expected_schema: Optional[str],
+) -> None:
+    """
+    Test that table and schema names are normalized to lowercase for Redshift.
+
+    Redshift folds unquoted identifiers to lowercase, so we need to normalize
+    table names to ensure consistent behavior when checking table existence
+    and performing replace operations.
+    """
+    normalized_table, normalized_schema = (
+        RedshiftEngineSpec.normalize_table_name_for_upload(table_name, schema_name)
+    )
+
+    assert normalized_table == expected_table
+    assert normalized_schema == expected_schema