diff --git a/pytensor/tensor/rewriting/shape.py b/pytensor/tensor/rewriting/shape.py
index e86411dd9c..9462504e78 100644
--- a/pytensor/tensor/rewriting/shape.py
+++ b/pytensor/tensor/rewriting/shape.py
@@ -966,16 +966,15 @@ def local_reshape_to_dimshuffle(fgraph, node):
     inp, output_shape = node.inputs
     [output] = node.outputs
 
-    # Remove any broadcastable dimensions from the input
-    squeeze_axes = [i for i, bcast in enumerate(inp.type.broadcastable) if bcast]
-
     # Trivial case, all dimensions of input/output are known to be broadcastable:
     # there's nothing to reshape
     if all(inp.type.broadcastable) or all(output.type.broadcastable):
+        squeeze_axes = tuple(range(inp.type.ndim))
         new_output_shape = []
         expand_axes = tuple(range(output.type.ndim))
 
     else:
+        squeeze_axes = [i for i, bcast in enumerate(inp.type.broadcastable) if bcast]
         unpacked_shape = _unpack_shape_vector(output_shape)
         new_output_shape = []
         expand_axes = []
diff --git a/tests/tensor/rewriting/test_shape.py b/tests/tensor/rewriting/test_shape.py
index 27678bd630..43df9ffd23 100644
--- a/tests/tensor/rewriting/test_shape.py
+++ b/tests/tensor/rewriting/test_shape.py
@@ -445,6 +445,15 @@ def test_squeeze_of_alloc(self):
         new_out = rewrite_graph(out, include=("canonicalize", "ShapeOpt"))
         assert equal_computations([new_out], [pt.alloc(x, 12, 9)], strict_dtype=False)
 
+    def test_reshape_implies_size_1_input(self):
+        x = pt.matrix("x", shape=(None, None))
+        out = pt.reshape(x, (1, 1, 1))
+
+        new_out = rewrite_graph(out, include=("canonicalize",))
+        assert equal_computations(
+            [new_out], [x.dimshuffle("x", "x", "x")], strict_dtype=False
+        )
+
 
 def test_expand_dims_squeeze_reshape_fusion():
     x = pt.tensor("x", shape=(1, 9))