diff --git a/pyathena/sqlalchemy/base.py b/pyathena/sqlalchemy/base.py
index 8eada2056..5295d08e8 100644
--- a/pyathena/sqlalchemy/base.py
+++ b/pyathena/sqlalchemy/base.py
@@ -8,6 +8,7 @@
 from typing import (
     TYPE_CHECKING,
     Any,
+    ClassVar,
     cast,
 )
 
@@ -143,7 +144,7 @@ class AthenaDialect(DefaultDialect):
     preparer: type[IdentifierPreparer] = AthenaDMLIdentifierPreparer
     statement_compiler: type[SQLCompiler] = AthenaStatementCompiler
     ddl_compiler: type[DDLCompiler] = AthenaDDLCompiler
-    type_compiler: type[GenericTypeCompiler] = AthenaTypeCompiler
+    type_compiler_cls: ClassVar[type[GenericTypeCompiler]] = AthenaTypeCompiler
     default_paramstyle: str = pyathena.paramstyle
     max_identifier_length: int = 255
     cte_follows_insert: bool = True
diff --git a/pyathena/sqlalchemy/compiler.py b/pyathena/sqlalchemy/compiler.py
index 4211643b5..8687224e0 100644
--- a/pyathena/sqlalchemy/compiler.py
+++ b/pyathena/sqlalchemy/compiler.py
@@ -1182,7 +1182,7 @@ def get_column_specification(self, column: Column[Any], **kwargs) -> str:
             type_ = "INT"
         else:
             # type_expression marks column DDL so STRUCT and MAP use Hive syntax.
-            type_ = self.dialect.type_compiler.process(column.type, type_expression=column)
+            type_ = self.dialect.type_compiler_instance.process(column.type, type_expression=column)
         text = [f"{self.preparer.format_column(column)} {type_}"]
         if column.comment:
             text.append(f"{self._get_comment_specification(column.comment)}")
diff --git a/tests/pyathena/sqlalchemy/test_base.py b/tests/pyathena/sqlalchemy/test_base.py
index bf010643f..d956232c9 100644
--- a/tests/pyathena/sqlalchemy/test_base.py
+++ b/tests/pyathena/sqlalchemy/test_base.py
@@ -19,11 +19,13 @@
 from sqlalchemy.sql.schema import Column, MetaData, Table
 from sqlalchemy.sql.selectable import TextualSelect
 
+from pyathena.aio.sqlalchemy.base import AthenaAioDialect
 from pyathena.converter import DefaultTypeConverter
 from pyathena.cursor import Cursor
 from pyathena.error import DatabaseError, OperationalError
 from pyathena.formatter import DefaultParameterFormatter
 from pyathena.sqlalchemy.base import AthenaDialect
+from pyathena.sqlalchemy.compiler import AthenaTypeCompiler
 from pyathena.sqlalchemy.types import (
     TINYINT,
     AthenaArray,
@@ -108,6 +110,14 @@ def close(self):
 
 
 class TestAthenaDialect:
+    @pytest.mark.parametrize("dialect_class", [AthenaDialect, AthenaAioDialect])
+    def test_type_compiler(self, dialect_class):
+        # SQLAlchemy 2.0 builds the type compiler from type_compiler_cls. A legacy
+        # type_compiler class attribute would take precedence over it.
+        assert not hasattr(dialect_class, "type_compiler")
+        assert dialect_class.type_compiler_cls is AthenaTypeCompiler
+        assert isinstance(dialect_class().type_compiler_instance, AthenaTypeCompiler)
+
     def test_columns_from_information_schema(self):
         # Rows arrive unordered, and Athena reports a missing comment as NULL.
         # The API cursor this path pins hands that over as None or as an empty