Cathie Wood: AI Inference Costs Fall 99.99% a Year, Fueling ‘Benign Deflation’

Wood's framing matters less as a forecast than as a lens: collapsing inference costs paired with rising AI revenue imply demand elasticity strong enough to absorb the price decline, which is the core condition for deflation to stay benign rather than margin-destructive. The tension worth watching is the rates call — her scenario has inflation undershooting while real growth runs hot, a combination that historically supports higher, not lower, yields, and that sits awkwardly with crypto's usual rate-cut reflex. Whether inference cost declines actually flow into broad productivity data, rather than being captured mostly by model providers, is the open question.