FOCUS specification eyes AI token economics as AI billing complexity hits a new frontier

AI token economics is creating a data normalization crisis across the technology stack as enterprises scramble to apply consistent financial controls to GPU clusters, AI factories and token-based consumption that existing cloud billing frameworks were never designed to handle.

As the FinOps Foundation Project, a Series of LF Projects LLC, community gathers in San Diego for FinOps X 2026, attention is turning to FOCUS — the FinOps Open Cost and Usage Specification — and whether the open standard can extend its multicloud billing lingua franca to cover a new generation of AI spending. Among organizations spending $100 million or more annually, roughly 68% are already using or experimenting with FOCUS-formatted data, signaling broad momentum for a common billing language, according to Shawn Alpay (pictured, right), director of data engineering at FinOps.

“The data is not normalized across providers, across technology categories, public cloud, AI, SaaS, etc,” Alpay said. “Being able to tell that story with the same names of the columns, with the same definitions of the allowed values … being able to have that story told the same way across all providers — it’s incredibly valuable. And then to be able to land that into your data lake, your data store, your data warehouse, just shorten the time it takes to get to market for your organization’s analytical needs, your accounting needs, your financial needs.”

Alpay and Matt Cowsert (left), principal product manager at FinOps, spoke with theCUBE’s John Furrier and Paul Nashawatyat at FinOps X 2026, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed the evolution of the FOCUS specification, AI token economics and how the open standard is expanding to serve new personas and technology categories. (* Disclosure below.)

FOCUS specification expands to AI token economics and new enterprise personas

With FOCUS 1.4 launching at the conference, the working group is pushing into new datasets and personas well beyond its public cloud roots. FOCUS now serves not just data engineers and FinOps practitioners but also finance, leadership and procurement teams, Cowsert noted.

“FOCUS is also a language that you’re then able to leverage across your organization,” Cowsert said. “We’ve done a really solid job of making sure that those additional personas in your organization actually know the value that they can get out of FOCUS as well.”

The harder frontier is AI token economics. Measuring the cost of inference requires visibility down to the per-user, per-session and per-request level — a level of cardinality that introduces significant data engineering challenges, Alpay explained. The working group is actively weighing whether to create a net-new observability dataset or integrate with existing frameworks such as the OpenTelemetry project, he added.

“When it comes to AI and tokens in that realm, you’re starting to talk about per user, per session, per request, per operation — and the farther down this ladder you go, the larger the cardinality of the data and the harder it is to gather that information,” Alpay said. “It could end up being that a non-trivial percentage of your cost is having to be attributed to just getting your data and putting it through a pipeline. It’s a non-zero cost just to tell the story. We have to be really thoughtful about how we deliver that information, potentially allowing you as a consumer to choose whether to consume data at that grain or not.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of FinOps X 2026:

(* Disclosure: TheCUBE is a paid media partner for the FinOps X event. Neither the FinOps Foundation, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE