AI spending is climbing the same growth curve cloud took a decade to scale. It's just moving twice as fast. Most engineering orgs don't have the visibility, attribution, or feedback loops to keep up. Cloud FinOps took years to mature into real tagging, allocation, and forecasting. AI isn't giving anyone that runway: leadership wants to know if it's paying off before rollout is even finished, often with no data to answer with and no link between a code change and what it costs to run.

In this session, we'll break down what actually separates AI cost management from cloud cost management: different unit economics, different owners, different attribution models, and a waste rate of 25% of cloud and AI spend, which took cloud a decade to reach and AI hit almost overnight. You'll see why treating AI spend like "cloud, but with GPUs" leaves most of it unaccounted for, what it takes to trace cost down to the commit, PR, or ticket it produced, and why closing that gap takes more than a better dashboard. It takes an agent that can act on what it finds, at whatever level of autonomy you're comfortable with.

Key Takeaways:

• Learn what really separates AI cost from cloud cost: unit economics, ownership, attribution

• See real cost per outcome numbers for AI coding tools, not just adoption metrics

• Understand why FinOps teams and AI COE teams end up siloed, and what closes that gap

• Get the real waste numbers behind the hype (25% of cloud and AI spend)

• race a cost spike back to the exact commit, PR, or ticket that caused it

• See what it looks like when cost governance acts before the bill arrives, not after