San Diego has a way of making things feel more consequential than they are. But standing in the conference hall this week, watching the FinOps Foundation announce the Tokenomics Foundation and unveil FOCUS 1.4, I couldn’t shake the feeling that something genuinely important was happening, and that not everyone in the room had clocked it yet.
Let’s be honest about something. For most of its existence, FinOps has been a niche discipline. Cloud spend management matters, of course it does, but of the vast number of organisations running workloads on AWS, Azure, or GCP, only a fraction ever felt the discipline strongly enough to build a practice around it. For most, “FinOps” meant a dashboard showing who left a dev environment running over the weekend, and a monthly conversation with procurement.
That’s not a criticism. It’s a structural observation. The economics of cloud computing, for most enterprises, were manageable enough that the ROI on a serious FinOps function was marginal. Visibility into spend: sure. Tying cost to business outcomes in a rigorous way? Nice to have, rarely imperative.
AI changes this completely. The cost dynamics of token usage are categorically different from those of compute or storage. Token spend is tied directly to workload behaviour in ways that are both highly variable and highly consequential. A poorly prompted agentic workflow doesn’t just cost more than a well-designed one; it can cost orders of magnitude more, and the feedback loop is opaque unless you instrument it deliberately. Goldman Sachs projects global token usage to multiply roughly 24 times between 2026 and 2030. If that lands anywhere near true, organisations that don’t build the discipline to measure token spend against business outcomes aren’t running a FinOps gap. They’re running a financial risk.
This is the structural shift. AI is the vector by which the core proposition of FinOps, tying cost to value, finally gets traction at enterprise scale. Not because FinOps practitioners weren’t right before. But because the magnitude of the problem just changed.
The announcement that attracted the most attention at FinOps X was probably Tokenomicon. But the more consequential release was quieter: FOCUS 1.4, now generally available, and the roadmap to FOCUS 1.5 with native AI token tracking.
FOCUS is one of those ideas that looks obvious in retrospect and was genuinely hard to achieve. An open specification that normalises billing data across cloud, SaaS, data centre, and other technology vendors. Not a proprietary schema owned by any provider, not a vendor-specific export format, but a shared taxonomy that lets organisations reason about spend across their entire technology estate. Version 1.4 closes the last major gaps around cost recognition and invoice reconciliation. Version 1.5, now on the roadmap, will bring unit economics and token tracking into the same framework.
For FinOps practitioners, this is genius. And for a very practical reason that a conversation here this week crystallised for me.
I spoke with a FinOps lead from a large US organisation, someone who manages technology spend at a scale where third-party tooling is a six-figure line item in itself. Her take on FOCUS was unambiguous: “Why would I use a FinOps vendor when I can build my own tooling?” She’s not being contrarian. She’s being rational. With FOCUS, she can download the taxonomy, export her cloud spend as a structured dataset, and build exactly the dashboards her organisation needs, without the overhead of a platform vendor’s pricing model, integration complexity, or opinionated UX.
In an AI-augmented world, this capability becomes even more accessible. The workflow she described, download the FOCUS taxonomy, output cloud spend as CSV, use a model like Claude to generate tailored dashboards and analysis, is not a technical stretch. It’s an afternoon’s work for a capable analyst with basic prompting skills. The barriers to building your own FinOps tooling have dropped substantially. FOCUS provides the conceptual scaffold; AI provides the execution leverage.
What this means for the ecosystem of FinOps platform vendors is genuinely uncertain. The value proposition of “we normalise your billing data” weakens considerably when the normalisation standard is open and the implementation effort is low. Vendors who survive this shift will do so by providing either depth of insight that isn’t easily replicated with a general-purpose model, or workflow integration that reduces friction in ways pure analysis doesn’t. The “we make the data legible” layer of the market is under real pressure.
J.R. Storment has navigated this tension with considerable skill. Getting hyperscalers and major SaaS vendors to endorse an open billing standard that, taken to its logical conclusion, reduces their customers’ dependency on proprietary tooling, is not a trivial achievement. The foundation model of neutral convener, practitioner community, and vendor ecosystem has held together precisely because FOCUS is framed as expanding the market for cloud optimisation, not cannibalising it. Whether that framing survives the economics of what’s coming is another question.
The Tokenomics Foundation announcement, and with it Tokenomicon as a standalone conference, deserves a more sceptical reading than the enthusiasm in the room suggested.
The framing is that token economics represents a new discipline requiring its own organisational home. I’d push back on that. Token spend is a subset of technology spend. The questions practitioners are asking, how do I attribute token costs to business outcomes, how do I govern model usage, how do I compare cost efficiency across providers, are structurally identical to the questions FinOps has been building frameworks to answer for years. Tokens are a new unit of measurement. They are not a new discipline.
Having two foundations, two streams within a conference, and two practitioner communities for what is fundamentally one problem space creates coordination costs, fragments practitioner attention, and risks building incompatible frameworks where a unified one would serve better. The organisations that most need this capability, enterprises moving AI workloads from pilot to production, don’t want to navigate two ecosystems. They want one place to find the answers.
The counterargument, which is not without merit, is that the FinOps Foundation’s existing community skews toward cloud infrastructure practitioners, and that token economics draws in a different population: ML engineers, AI product managers, procurement teams who haven’t engaged with FinOps before. A new brand creates on-ramps for that audience. Tokenomicon in San Diego in June 2027 may well attract tens of thousands of people who have never attended FinOps X.
In addition, and as an initiative that brings these two threads together, FOCUS 1.5 is the technical bridge that makes token economics legible within the existing framework. FOCUS as the unifying lens is exactly right. One specification, one taxonomy, one system of record for cloud, SaaS, data centre, and AI token spend alike. That’s the north star worth building toward. Organisational structures should serve the goal, not fragment it.
FinOps X 2026 was the year the discipline’s theoretical promise met a practical imperative. The token reckoning is coming regardless of how the foundations are structured, regardless of what the platform vendors do, regardless of how practitioners organise. Enterprises will face token budgets that dwarf their cloud bills. Some will have the tooling and discipline to manage them intelligently. Most, initially, will not.
The practitioners who will matter most in the next three years are the ones who can bridge the financial and the technical, who can look at a generative AI workload, understand the token cost structure, attribute it to business outcomes, and build the feedback loops that make spending decisions rational. FOCUS 1.5 will give them the vocabulary. AI tooling will lower the barrier to building the instrumentation. The community infrastructure, whether one foundation or two, needs to keep pace.
The thing my conversation partner understood intuitively, the FinOps lead with her CSV exports and her Claude dashboards, is that the tools for doing this work have fundamentally changed. The discipline still matters. The craft still matters. But the leverage available to a single skilled practitioner has increased dramatically.
That’s worth building a conference around. It’s worth building a foundation around. The question is whether the institutions being built now are designed for that future or for the one that just passed.
