We prove that any non-adaptive algorithm that tests whether an unknown Boolean function f : 0 1 n 0 1 is a k -junta or -far from every k -junta must make ( k 3 2 ) many queries for a wide range of parameters k and . Our result dramatically improves previous lower bounds from [BGSMdW13, STW15], and is essentially optimal given Blais's non-adaptive junta tester from [Blais08], which makes O ( k 3 2 ) queries. Combined with the adaptive tester of [Blais09] which makes O ( k log k + k ) queries, our result shows that adaptivity enables polynomial savings in query complexity for junta testing.