This paper investigates whether acquisitions involving public high technology firms are best understood in terms of acquirers taking over firms with ‘superior’ innovation performance to access their assets, or acquiring firms with ‘inferior’ innovation performance to turn them around. Innovation performance is proxied by R&D-intensity (R&D expenditure over assets), patent-intensity (patents per US$million of assets), i.e. the R&D productivity of a firm's assets, and the patent stock, i.e. the accumulated R&D output. We find substantial overlaps between target and non-acquired firm characteristics. Nevertheless targets have a relatively high R&D-intensity and a large patent stock, which is consistent with acquirers targeting firms with a superior innovation performance. However, these targets have significantly lower pre-acquisition patent-intensity and hence a lower R&D productivity. The targets are also experiencing weak financial performance. Our results are consistent with a selection process in which acquirers seek out firms that have a superior past innovation performance, but that are failing in terms of recent R&D productivity and financial performance. A comparison of the performance of the targets with their acquirers reinforces this conclusion.