diff --git a/dask_expr/_collection.py b/dask_expr/_collection.py
index 97ebfb27c..e61457308 100644
--- a/dask_expr/_collection.py
+++ b/dask_expr/_collection.py
@@ -299,7 +299,7 @@ def dask(self):
         # cause very unfortunate materializations. Even a mere hasattr(obj,
         # "dask") check already triggers this since it's a property, not even a
         # method.
-        return self.__dask_graph_factory__().materialize()
+        return self.__dask_graph_factory__().lower_completely().materialize()
 
     def finalize_compute(self):
         return new_collection(Repartition(self.expr, 1))
diff --git a/dask_expr/tests/test_collection.py b/dask_expr/tests/test_collection.py
index d40653cb5..83cbe6ea9 100644
--- a/dask_expr/tests/test_collection.py
+++ b/dask_expr/tests/test_collection.py
@@ -864,9 +864,9 @@ def test_persist(pdf, df):
     b = a.persist()
 
     assert_eq(a, b)
-    assert len(a.__dask_graph__()) > len(b.__dask_graph__())
+    assert len(a.dask) > len(b.dask)
 
-    assert len(b.__dask_graph__()) == b.npartitions
+    assert len(b.dask) == b.npartitions
 
     assert_eq(b.y.sum(), (pdf + 2).y.sum())
 
@@ -1113,7 +1113,7 @@ def test_serialization(pdf, df):
 
     part = df.partitions[0].compute()
     assert (
-        len(pickle.dumps(df.__dask_graph__()))
+        len(pickle.dumps(df.__dask_graph_factory__()))
         < 1000 + len(pickle.dumps(part)) * df.npartitions
     )
 
@@ -1185,7 +1185,7 @@ def test_tree_repr(fuse):
 
 def test_simple_graphs(df):
     expr = (df + 1).expr
-    graph = expr.__dask_graph__()
+    graph = expr.materialize()
 
     assert graph[(expr._name, 0)] == (operator.add, (df.expr._name, 0), 1)
 
@@ -1239,7 +1239,7 @@ def test_repartition_divisions(df, opt):
     assert_eq((df + 1)["x"], df2)
 
     # Check partitions
-    for p, part in enumerate(dask.compute(list(df2.index.partitions))[0]):
+    for p, part in enumerate(dask.compute(list(df2.index.partitions))):
         if len(part):
             assert part.min() >= df2.divisions[p]
             assert part.max() < df2.divisions[p + 1]
diff --git a/dask_expr/tests/test_fusion.py b/dask_expr/tests/test_fusion.py
index 00bc4056e..9bc64ae9b 100644
--- a/dask_expr/tests/test_fusion.py
+++ b/dask_expr/tests/test_fusion.py
@@ -114,7 +114,7 @@ def test_fuse_broadcast_deps():
     df3 = from_pandas(pdf3, npartitions=2)
 
     query = df.merge(df2).merge(df3)
-    assert len(query.optimize().__dask_graph__()) == 2
+    assert len(query.optimize().materialize()) == 2
     assert_eq(query, pdf.merge(pdf2).merge(pdf3))
 
 
diff --git a/dask_expr/tests/test_merge.py b/dask_expr/tests/test_merge.py
index a7f25ace9..ae4a778b4 100644
--- a/dask_expr/tests/test_merge.py
+++ b/dask_expr/tests/test_merge.py
@@ -230,9 +230,7 @@ def test_merge_combine_similar(npartitions_left, npartitions_right):
     query = df.merge(df2)
     query["new"] = query.b + query.c
     query = query.groupby(["a", "e", "x"]).new.sum()
-    assert (
-        len(query.optimize().__dask_graph__()) <= 25
-    )  # 45 is the non-combined version
+    assert len(query.optimize().materialize()) <= 25  # 45 is the non-combined version
 
     expected = pdf.merge(pdf2)
     expected["new"] = expected.b + expected.c