diff --git a/tensornetwork/backends/pytorch/pytorch_backend.py b/tensornetwork/backends/pytorch/pytorch_backend.py
index 8467fac39..fb3ae238a 100644
--- a/tensornetwork/backends/pytorch/pytorch_backend.py
+++ b/tensornetwork/backends/pytorch/pytorch_backend.py
@@ -475,3 +475,17 @@ def sign(self, tensor: Tensor) -> Tensor:
   
   def item(self, tensor):
     return tensor.item()
+
+  def power(self, a: Tensor, b: Tensor) -> Tensor:
+    """
+    Returns the power of tensor a to the value of b.
+    In the case b is a tensor, then the power is by element
+      with a as the base and b as the exponent.
+    In the case b is a scalar, then the power of each value in a
+      is raised to the exponent of b.
+
+    Args:
+      a: The tensor that contains the base.
+      b: The tensor that contains the exponent or a single scalar.
+    """
+    return a ** b
diff --git a/tensornetwork/backends/pytorch/pytorch_backend_test.py b/tensornetwork/backends/pytorch/pytorch_backend_test.py
index f44aed5c6..82a3e6356 100644
--- a/tensornetwork/backends/pytorch/pytorch_backend_test.py
+++ b/tensornetwork/backends/pytorch/pytorch_backend_test.py
@@ -566,6 +566,18 @@ def test_matmul():
   np.testing.assert_allclose(expected, actual)
 
 
+def test_power():
+  np.random.seed(10)
+  backend = pytorch_backend.PyTorchBackend()
+  t1 = np.random.rand(10, 2, 3)
+  t2 = np.random.rand(10, 3, 4)
+  a = backend.convert_to_tensor(t1)
+  b = backend.convert_to_tensor(t2)
+  actual = backend.power(a, b)
+  expected = np.power(t1, t2)
+  np.testing.assert_allclose(expected, actual)
+
+
 @pytest.mark.parametrize("dtype", torch_randn_dtypes)
 @pytest.mark.parametrize("offset", range(-2, 2))
 @pytest.mark.parametrize("axis1", [-2, 0])
diff --git a/tensornetwork/backends/symmetric/symmetric_backend.py b/tensornetwork/backends/symmetric/symmetric_backend.py
index 9e55bc694..1b1c517e3 100644
--- a/tensornetwork/backends/symmetric/symmetric_backend.py
+++ b/tensornetwork/backends/symmetric/symmetric_backend.py
@@ -689,3 +689,17 @@ def matmul(self, tensor1: Tensor, tensor2: Tensor):
     if (tensor1.ndim != 2) or (tensor2.ndim != 2):
       raise ValueError("inputs to `matmul` have to be matrices")
     return tensor1 @ tensor2
+
+  def power(self, a: Tensor, b: Tensor) -> Tensor:
+    """
+    Returns the power of tensor a to the value of b.
+    In the case b is a tensor, then the power is by element
+      with a as the base and b as the exponent.
+    In the case b is a scalar, then the power of each value in a
+      is raised to the exponent of b.
+
+    Args:
+      a: The tensor that contains the base.
+      b: The tensor that contains the exponent or a single scalar.
+    """
+    return a ** b
diff --git a/tensornetwork/backends/symmetric/symmetric_backend_test.py b/tensornetwork/backends/symmetric/symmetric_backend_test.py
index a8b78051e..f38d8bdc5 100644
--- a/tensornetwork/backends/symmetric/symmetric_backend_test.py
+++ b/tensornetwork/backends/symmetric/symmetric_backend_test.py
@@ -609,6 +609,19 @@ def test_addition_raises(R, dtype, num_charges):
     backend.addition(a, c)
 
 
+@pytest.mark.parametrize("dtype", np_dtypes)
+@pytest.mark.parametrize("num_charges", [1, 2])
+def test_power(dtype, num_charges):
+  np.random.seed(10)
+  R = 4
+  backend = symmetric_backend.SymmetricBackend()
+  a = get_tensor(R, num_charges, dtype)
+  b = BlockSparseTensor.random(a.sparse_shape)
+  expected = np.power(a.data, b.data)
+  actual = backend.power(a.data, b.data)
+  np.testing.assert_allclose(expected, actual)
+
+
 @pytest.mark.parametrize("dtype", np_dtypes)
 @pytest.mark.parametrize("R", [2, 3, 4, 5])
 @pytest.mark.parametrize("num_charges", [1, 2])