Amazon CEO Cites Inference Costs And AI Talent As Adoption Barriers
The Common Question
Why are inference costs blocking broader enterprise AI adoption at scale?
The Lesson in This Case
Fortune 500 spending on Amazon Bedrock ran roughly seven times higher year over year in Q3, even as CEO Andy Jassy identified inference costs and an AI talent shortage as the primary forces slowing broader enterprise adoption. Inference already generates the largest share of AWS AI revenue, and Jassy expects it to become the overwhelming majority of future AI spending as models run constantly at scale, driving Amazon to develop custom silicon and pursue efficiency gains in model architecture, distillation, and prompt caching. Senior leaders budgeting for AI deployment should treat inference as a compounding operating cost and press vendors on chip efficiency and talent support before expanding rollouts.
Originally reported by CFO Dive. Read the full story here.