“Move Fast and Break Things” Is Dead. What Counts in Enterprise AI Now?

The "move fast and break things" motto that propelled the AI boom is no longer effective, as costs have become more expensive and results have ceased to meet expectations. In a PwC survey of 4,454 CEOs, 56% reported no significant financial benefit from AI so far. Unexpected AI cost spikes have become a major issue, with 72% of organizations experiencing them.

Companies like Uber and Facebook have implemented measures to control AI costs. Uber limited each employee to $1,500 a month per AI coding tool, while Facebook changed its motto to "move fast with stable infra." CFOs now prioritize AI cost management as an architecture issue rather than a budget line item. Finance teams focus on understanding unit price to make architecture decisions.

Optimal use of the "cost-to-outcome" frontier is crucial, as customers measure the cost of tokens against the outcomes they produce. FinOps for AI is becoming increasingly important, with a focus on tagging every model call by team and feature to prevent runaway expenses. The new challenge for businesses is managing AI infrastructure spend, as finance teams now have veto power over AI investments.

Despite the shift in focus, big tech companies like Amazon, Meta, Microsoft, and Alphabet continue to spend heavily on AI, as they build capacity and fund it with large cash flows, betting on future demand.