diff --git a/doc/source/ray-core/accelerator-types.rst b/doc/source/ray-core/accelerator-types.rst
new file mode 100644
index 0000000000000..a4559a038ca45
--- /dev/null
+++ b/doc/source/ray-core/accelerator-types.rst
@@ -0,0 +1,11 @@
+:orphan:
+
+.. _accelerator_types:
+
+Accelerator Types
+=================
+
+Ray supports the following accelerator types:
+
+.. literalinclude:: ../../../python/ray/util/accelerators/accelerators.py
+    :language: python
\ No newline at end of file
diff --git a/python/ray/_private/worker.py b/python/ray/_private/worker.py
index a8f8baa393778..0dcd0d3ff00a9 100644
--- a/python/ray/_private/worker.py
+++ b/python/ray/_private/worker.py
@@ -3281,7 +3281,7 @@ def method(self):
             By default it is empty.
         accelerator_type: If specified, requires that the task or actor run
             on a node with the specified type of accelerator.
-            See `ray.util.accelerators` for accelerator types.
+            See :ref:`accelerator types <accelerator_types>`.
         memory: The heap memory request in bytes for this task/actor,
             rounded down to the nearest integer.
         max_calls: Only for *remote functions*. This specifies the
diff --git a/python/ray/actor.py b/python/ray/actor.py
index 7eb98cdd894bf..a6adfd5e862a7 100644
--- a/python/ray/actor.py
+++ b/python/ray/actor.py
@@ -353,6 +353,7 @@ class _ActorClassMetadata:
         resources: The default resources required by the actor creation task.
         accelerator_type: The specified type of accelerator required for the
             node on which this actor runs.
+            See :ref:`accelerator types <accelerator_types>`.
         runtime_env: The runtime environment for this actor.
         scheduling_strategy: Strategy about how to schedule this actor.
         last_export_session_and_job: A pair of the last exported session
@@ -591,7 +592,7 @@ def options(self, **actor_options):
                 This is a dictionary mapping strings (resource names) to floats.
             accelerator_type: If specified, requires that the task or actor run
                 on a node with the specified type of accelerator.
-                See `ray.util.accelerators` for accelerator types.
+                See :ref:`accelerator types <accelerator_types>`.
             memory: The heap memory request in bytes for this task/actor,
                 rounded down to the nearest integer.
             object_store_memory: The object store memory request for actors only.
diff --git a/python/ray/remote_function.py b/python/ray/remote_function.py
index c11495b8b171e..77212413ad150 100644
--- a/python/ray/remote_function.py
+++ b/python/ray/remote_function.py
@@ -174,7 +174,7 @@ def options(self, **task_options):
                 This is a dictionary mapping strings (resource names) to floats.
             accelerator_type: If specified, requires that the task or actor run
                 on a node with the specified type of accelerator.
-                See `ray.util.accelerators` for accelerator types.
+                See :ref:`accelerator types <accelerator_types>`.
             memory: The heap memory request in bytes for this task/actor,
                 rounded down to the nearest integer.
             object_store_memory: The object store memory request for actors only.
diff --git a/python/ray/serve/schema.py b/python/ray/serve/schema.py
index 5489353bc1c09..9bceff6c54cfc 100644
--- a/python/ray/serve/schema.py
+++ b/python/ray/serve/schema.py
@@ -185,6 +185,7 @@ class RayActorOptionsSchema(BaseModel):
         default=None,
         description=(
             "Forces replicas to run on nodes with the specified accelerator type."
+            "See :ref:`accelerator types <accelerator_types>`."
         ),
     )