We show that there are infinitely many primes p such that p−1 is divisible by a square d2≥pθ for θ=1/2+1/2000. This improves the work of Matomäki (2009) who obtained the result for θ=1/2−ε (with the added constraint that d is also a prime), which improved the result of Baier and Zhao (2006) with θ=4/9−ε. As in the work of Matomäki, we apply Harman’s sieve method to detect primes p≡1(d2). To break the θ=1/2 barrier we prove a new bilinear equidistribution estimate modulo smooth square moduli d2 by using a similar argument to the one Zhang (2014) used to obtain equidistribution beyond the Bombieri–Vinogradov range for primes with respect to smooth moduli. To optimize the argument we incorporate technical refinements from the Polymath project (2014). Since the moduli are squares, the method produces complete exponential sums modulo squares of primes which are estimated using the results of Cochrane and Zheng (2000).