In 2024, forecasting AI spend was easy, as it was embedded in the product. AI was treated as another service. In 2025, that math didn’t just break, it evaporated. At Palo Alto Networks, we watched AI spend surge from APIs and tokens to agents. Agents don’t behave like standard cloud services; they decide how much to spend and how hard they ‘think’. The shift from simple chatbots to autonomous agents has introduced a new, volatile bill of materials.
This session outlines our journey through the AI FinOps Maturity Model, bridging the ‘Actionability GAP’ between high-level visibility and engineering-led cost targets. We will discuss the evolution of AI use cases and break down the AI stack into ‘Metal’, ‘Workbench’, and ‘Brain’ layers to show how to build a 2026 budget that accounts for the variability of agentic behavior.
