You can use numpy.ufunc.reduce if you don't need the intermediary accumulated results of any ufunc.
>>> a = np.arange(9).reshape(3,3)
>>> a
>>>
array([[0, 1, 2],
[3, 4, 5],
[6, 7, 8]])
>>>
>>> np.add.reduce(a, axis=1)
>>> array([ 3, 12, 21])
However, in the case of sum, Willem's answer is clearly superior and to be preferred. Just keep in mind that in the general case, there's ufunc.reduce.
