gistfile1.py

import random
import timeit

def inter_reduce(sets):
    return reduce(set.intersection, sets)

def inter_for(sets):
    sets = iter(sets)
    result = next(sets)
    for s in sets:
        result.intersection_update(s)
    return result

random.seed(42)
sets = [set(random.randrange(1000) for i in range(4000))
        for j in range(100)]
print timeit.timeit("inter_reduce(sets)",
                    "from __main__ import inter_reduce, sets", number=1000)
print timeit.timeit("inter_for(sets)",
                    "from __main__ import inter_for, sets", number=1000)

^^ as an addendum to this, my fork has versions that are referentially transparent, but this definitely adds significant overhead (still cheaper than reduce though). I'm pretty sure there's a better way of doing it as well.