AI Spend Becomes a Capital Decision: FinOps Report Reveals 26% Waste and Ownership Gaps

A recent report by Harness, based on a survey of 700 engineering leaders and practitioners, reveals that AI spend has rapidly escalated to become a major financial commitment for enterprises. For 20% of organizations, AI now represents a monthly expenditure exceeding $1 million, with two-thirds spending over $250,000 per month. The critical finding is that approximately 26% of this AI spend is estimated to be wasted across infrastructure, software, models, and services. This development is highly significant for FinOps practitioners, platform teams, and engineering leadership. The report underscores that AI costs are no longer a minor operational expense but a strategic capital decision. The lack of clear ownership, with accountability split across engineering, FinOps, finance, and IT, is a primary driver of this waste and makes effective cost management challenging. The patterns observed—invoice shock, ownership confusion, and governance gaps—mirror the early days of cloud adoption, but are unfolding at an accelerated pace. This trend fits squarely within the broader evolution of FinOps, which has consistently aimed to bring financial accountability and cost optimization to dynamic technology environments. Just as FinOps matured to address the complexities of cloud spend, it must now rapidly adapt to the unique characteristics of AI workloads. The shift to GPU-heavy and LLM workloads has fundamentally altered the cost equation, demanding real-time decisioning and autonomous optimization capabilities that traditional rule-based cost management struggles to provide. The FinOps Foundation's recent focus on expanding its framework to include AI economics and token pricing further emphasizes this critical shift. In practice, this means organizations must prioritize establishing clear ownership for AI costs, implementing granular visibility into token-based pricing and inference volumes, and developing robust attribution models. Practitioners should focus on integrating AI cost management into existing FinOps frameworks, leveraging tools that can provide per-model token visibility and identify idle provisioned throughput. Furthermore, the report highlights that tracing cost spikes often takes days or even weeks, leading to significant financial losses. Therefore, investing in automated anomaly detection and remediation for AI spend is crucial. The goal is to move beyond reactive cost management to proactive optimization, ensuring that AI investments deliver measurable business value rather than contributing to an ever-growing pool of wasted resources.