What is FOCUS, the open specification for cloud and AI cost data?

FOCUS (FinOps Open Cost and Usage Specification) is an open specification that defines how cloud, SaaS and AI vendors must structure the billing data they produce so it reaches the customer in a common format: same columns, same names, same calculation rules. Without a standard, every vendor formats its bill differently, and anyone who wants to know total spend first has to build and maintain a converter for each one. With FOCUS the cost of an hour of virtual machine time on AWS and the cost of credits consumed on Snowflake sit in the same table and, when billed in the same currency, add up directly. The specification is a Linux Foundation project, started within the FinOps Foundation and also supported by the Tokenomics Foundation and the ITAM Forum. FOCUS does not say how to cut spend: it supplies the clean data that FinOps works on.

See also: FinOps · The practice bringing financial accountability to the cloud: every team sees, understands and optimizes the cost of what it runs.

What it covers today and what is still missing

Version 1.0 became generally available in June 2024, fourteen months after work began in a FinOps Foundation working group. Version 1.2, from May 2025, added SaaS and PaaS plus virtual currencies, the credits and tokens some vendors price consumption in before converting it to dollars. Version 1.4, ratified on June 4, 2026, introduces two datasets, Invoice Detail and Billing Period, to reconcile consumption line by line with the invoices themselves: FOCUS starts talking to accounts payable, not just to the cloud team. Token counts already fit existing columns. Standard fields for the AI model and for input versus output tokens arrive with version 1.5, due for ratification on December 3, 2026, as recommended rather than mandatory properties, so today the per-model detail that AI tokenomics needs remains custom work. Meanwhile, according to the project's own list, AWS, Microsoft and Google Cloud export data in version 1.2.

Your own data warehouse or a FinOps platform?

When every bill came in a different format, normalizing them was a major reason to buy a FinOps platform. With FOCUS exports the vendor does most of the normalizing (different versions and vendor-specific columns remain), and loading the files into your own data lakehouse or data warehouse becomes an ordinary data engineering project. My rule of thumb: building in-house pays off when costs must be joined with data only your company owns (customers, products, cost centers) and a data team already exists; a platform pays off when you need alerts, recommendations and commitment management right away, or when many of your vendors do not export FOCUS yet. Either way the open format reduces lock-in: cost history survives a change of tools.

An enterprise example

A manufacturing group runs AWS for its plant systems, Azure for its business applications and Snowflake for analytics, and every month the controller's team stitches three bills together in a spreadsheet. The path I would propose: turn on native FOCUS exports, load them into a single table on the data platform already in use and map them to the company's cost centers; then, once vendors offer them, add the 1.4 invoice datasets to check every line before payment. The result is a single, verifiable number to start optimizing from. A commercial platform comes later, judged on what the in-house solution cannot do.

Why it matters for decision makers

FOCUS moves part of the problem out of engineering and into contracts. Asking every new cloud, SaaS or AI vendor for a FOCUS export costs one line in the RFP and spares you a custom connector for each; the FinOps Foundation itself urges customers to request it, along with the 1.4 invoice datasets. And since the specification has moved past the version the major vendors export, write the required version into the contract; the FOCUS logo on the vendor's website is not enough.