The first directly imaged exoplanets indicated that wide-orbit giant planets could be more common around A-type stars. However, the relatively small number of nearby A-stars has limited the precision of exoplanet demographics studies to >~10%. We aim to constrain the frequency of wide-orbit giant planets around A-stars using the VLT/SPHERE extreme adaptive optics system, which enables targeting <~100 A-stars between 100 and 200pc. We present the results of a survey of 84 A-stars within the nearby ~5-17Myr old ScoOB2 association. The survey detected three companions-one of which is a new discovery (HIP75056Ab), whereas the other two (HD95086b and HIP65426b) are now-known planets that were included without a priori knowledge of their existence. We assessed the image sensitivity and observational biases with injection and recovery tests combined with Monte Carlo simulations to place constraints on the underlying demographics. We measure a decreasing frequency of giant planets with increasing separation, with measured values falling between 10% and 2% for separations of 30-100au, and 95% confidence-level upper limits of <~45%-8% for planets on 30-100au orbits, and <~5% between 200 and 500au. These values are in excellent agreement with recent surveys of A-stars in the solar neighborhood-supporting findings that giant planets out to separations of <~100 au are more frequent around A-stars than around solar-type hosts. Finally, the relatively low occurrence rate of super-Jupiters on wide orbits, the positive correlation with stellar mass, and the inverse correlation with orbital separation are consistent with core accretion being their dominant formation mechanism.