diff --git a/h/celery.py b/h/celery.py
index 44a62904a76..78e9342849d 100644
--- a/h/celery.py
+++ b/h/celery.py
@@ -74,7 +74,6 @@
     # task's @app.task() arguments.
     task_time_limit=240,
     imports=(
-        "h.tasks.annotations",
         "h.tasks.cleanup",
         "h.tasks.indexer",
         "h.tasks.mailer",
diff --git a/h/tasks/annotations.py b/h/tasks/annotations.py
deleted file mode 100644
index 853d893c3ee..00000000000
--- a/h/tasks/annotations.py
+++ /dev/null
@@ -1,39 +0,0 @@
-from sqlalchemy import func, update
-
-from h.celery import celery, get_task_logger
-from h.models import Annotation, User
-
-log = get_task_logger(__name__)
-
-
-@celery.task
-def fill_pk_and_user_id(batch_size=1000):
-    """
-    Task to fill the new annotation.pk and annotation.user_id in batches.
-
-    Once most of the existing rows are done we'll make the code changes
-    to keep these up to date, make the column not nullable and remove this task.
-    """
-    # pylint: disable=no-member
-    db = celery.request.db
-
-    annotations = (
-        db.query(Annotation.id.label("annotation_id"), User.id.label("user_id"))
-        .join(
-            User,
-            User.username
-            == func.split_part(func.split_part(Annotation.userid, "@", 1), ":", 2),
-        )
-        .where(Annotation.pk.is_(None))
-        .order_by(Annotation.created.asc())
-        .with_for_update(skip_locked=True)
-        .limit(batch_size)
-    ).cte("annotations")
-
-    db.execute(
-        update(Annotation)
-        .values(pk=Annotation.pk_sequence.next_value(), user_id=annotations.c.user_id)
-        .where(Annotation.id == annotations.c.annotation_id)
-        # just update the rows in the DB, we don't need to refresh the objects in the session
-        .execution_options(synchronize_session=False)
-    )
diff --git a/tests/h/tasks/annotations_test.py b/tests/h/tasks/annotations_test.py
deleted file mode 100644
index c0970092721..00000000000
--- a/tests/h/tasks/annotations_test.py
+++ /dev/null
@@ -1,50 +0,0 @@
-import pytest
-
-from h.models import Annotation
-from h.tasks.annotations import fill_pk_and_user_id
-
-
-class TestFillPKAndUserId:
-    AUTHORITY_1 = "AUTHORITY_1"
-    AUTHORITY_2 = "AUTHORITY_2"
-
-    USERNAME_1 = "USERNAME_1"
-    USERNAME_2 = "USERNAME_2"
-
-    def test_it(self, factories, db_session):
-        author_1 = factories.User(authority=self.AUTHORITY_1, username=self.USERNAME_1)
-        author_2 = factories.User(authority=self.AUTHORITY_2, username=self.USERNAME_2)
-
-        annos_1 = factories.Annotation.create_batch(
-            5,
-            userid=author_1.userid,
-        )
-        annos_2 = factories.Annotation.create_batch(
-            5,
-            userid=author_2.userid,
-        )
-        factories.Annotation.create_batch(
-            5,
-            userid=author_2.userid,
-        )
-
-        fill_pk_and_user_id(batch_size=10)
-
-        # Only one batch of 10 was processed, those have PKs now
-        assert (
-            db_session.query(Annotation).filter(Annotation.pk.is_not(None)).count()
-            == 10
-        )
-        assert db_session.query(Annotation).filter(Annotation.pk.is_(None)).count() == 5
-
-        # Refresh data for the annotations
-        _ = [db_session.refresh(anno) for anno in annos_1 + annos_2]
-        # user_id was updated with the right value
-        assert {anno.user_id for anno in annos_1} == {author_1.id}
-        assert {anno.user_id for anno in annos_2} == {author_2.id}
-
-    @pytest.fixture(autouse=True)
-    def celery(self, patch, db_session):
-        cel = patch("h.tasks.annotations.celery", autospec=False)
-        cel.request.db = db_session
-        return cel