Three years ago, FinOps meant one thing: controlling cloud infrastructure costs. Today, it means something fundamentally different.
The FinOps Foundation's 2025 framework redefined the discipline from "advancing people who manage the value of cloud" to "advancing people who manage the value of technology." That single word change — from cloud to technology — reflects a hard truth: for an enterprise with $100M in annual technology spend, public cloud typically represents only 30–40% of the bill. The rest is fragmented across SaaS, data centre infrastructure, legacy licensing agreements, and increasingly, generative AI consumption.
90% of respondents now manage SaaS. 64% manage licensing. 57% manage private cloud. 48% manage data centres. 98% now manage AI spend. An emerging 28% are beginning to include labour costs. FinOps is no longer cloud cost management — it is enterprise technology value management.
Not every organisation needs the same level of FinOps sophistication. A 50-person startup running a single AWS account has different needs than a 10,000-person enterprise managing $200M in annual technology spend across three cloud providers, 400 SaaS applications, and a growing portfolio of AI workloads.
Maturity assessment serves three purposes:
The FinOps Foundation defines three maturity phases that apply across all capabilities and scopes. These are not sequential gates — an organisation can be at different maturity levels for different capabilities simultaneously.
At this phase, the organisation has basic cost reporting but limited ability to act on it. Cloud bills arrive monthly and are reviewed after the fact. Tagging is inconsistent or absent. Cost allocation is manual and approximate. AI spend is either unmeasured or reported in aggregate with no granularity.
At this phase, the organisation has a functioning FinOps practice with defined roles, regular reporting cadences, and established governance processes. Cost allocation is automated for major services. Showback reports are published regularly, and chargeback may be in place for cloud and SaaS. AI spend is tracked at the model and team level.
At this phase, FinOps is not a separate function — it is embedded into how the organisation designs, builds, and operates technology. Cost considerations are part of architecture decisions, sprint planning, and product roadmaps. AI token economics are governed as a financial asset.
The 2025/2026 framework organises FinOps capabilities into six core domains. Each domain applies across all scopes — cloud, SaaS, licensing, data centre, and AI.
| Domain | What It Covers |
|---|---|
| Plan & Forecast | Budget development, usage forecasting, financial planning. Matures from annual spreadsheets to continuous AI-assisted forecasting. |
| Inform | Cost visibility, reporting, benchmarking, chargeback. The foundational domain — without accurate data, no other domain can function. |
| Optimise | Rate optimisation (committed use, reserved capacity, spot pricing), rightsizing, waste elimination. |
| Quantify Business Value | Unit economics, cost per outcome, ROI measurement. Transforms FinOps from cost management into value management. |
| Manage Commitment & Usage | Procurement, contract negotiation, commitment tracking. Enterprises with $10M+ in annual cloud spend leave significant money on the table without this. |
| Govern | Policy enforcement, compliance, approval workflows, access controls. Ensures FinOps decisions are consistent, auditable, and aligned with enterprise risk management. |
The most significant development in the 2025–2026 FinOps landscape is the formalisation of AI as a dedicated scope. The FinOps Foundation released a dedicated FinOps for AI certification in 2025, signalling that AI cost management is now a core competency.
AI presents unique challenges that test the flexibility of the FinOps framework:
- Less transparent pricing: AI workloads often have variable, less predictable pricing than traditional cloud services.
- Rapid model evolution: New models and pricing tiers appear monthly, requiring constant re-evaluation.
- Output quality tradeoffs: Cost optimisation must be balanced against output quality in ways that don't apply to infrastructure.
- Emerging observability: Token-level tracking and attribution tooling is still maturing.
Organisations that adapt their FinOps fundamentals to AI quickly will gain an early advantage in managing this fast-growing spend area. Those that treat AI as a separate, ungoverned domain will face the same bill shock that characterised the early years of cloud adoption.
A practical maturity assessment should evaluate five dimensions. For each, score your organisation on a 1–5 scale. The composite score provides a diagnostic baseline and identifies the specific capability gaps that will deliver the highest return when addressed.