In my telecom years in Saskatchewan, the worst week of every month was invoice week. The carrier sent a statement. Our own systems said we had used something slightly different. The two numbers were close, never identical, and the job was to find out which one was wrong before accounts payable cut a check. Some months the carrier had added a circuit charge nobody recognized. Some months our internal records were stale. Either way, a person sat with two spreadsheets and reconciled line by line, because there was no shared format that let the usage record and the invoice talk to each other.
Cloud billing has had the same problem since the first AWS statement landed in a finance inbox. The cost and usage export answers “what did we consume?” It does not answer “what is on the invoice, and does it match?” On June 4, 2026, the FinOps Foundation ratified FOCUS 1.4, and for the first time the open billing standard treats the invoice itself as a first-class dataset. That is a bigger deal for finance teams than the version number suggests.
What FOCUS 1.4 actually adds
FOCUS, the FinOps Open Cost and Usage Specification, is the common schema that AWS, Azure, Google Cloud, Oracle, and more than a dozen other providers now use to export billing data in the same shape. Before FOCUS, every provider’s export looked different, and normalizing them across a multi-cloud estate meant custom integrations and a lot of guesswork. The standard updates twice a year. Version 1.4 was authored by Shawn Alpay and Matt Cowsert and adds 2 datasets, 47 columns, 6 attributes, and 17 glossary entries.
The two new datasets are the story:
- Invoice Detail carries the financial record as it appears on the issued invoice: payment currency, payment terms, due date, and purchase order number. This is the data AP actually works from, not the usage telemetry FinOps usually sees.
- Billing Period provides invoice-issuer-aware period boundaries and status, so you can do accurate period-over-period analysis even when a provider’s billing month does not line up with your calendar month.
The new “invoice reconciliation” feature joins Invoice Detail to the Cost and Usage dataset on Invoice ID, and an Invoice Detail ID lets you reconcile at the individual line-item level. There is even a Rounding Variance Tolerance appendix that defines precision formulas, so a reconciliation matches sums within a defined tolerance instead of failing every time a fraction of a cent rounds differently. Anyone who has watched an automated check break over a $0.003 discrepancy will appreciate that detail.
Why “what we used” was never the same as “what we paid”
The gap FOCUS 1.4 targets is not theoretical. It is the reason month-end close on cloud spend is still partly manual at most companies.
Cloud usage data and the invoice diverge for ordinary reasons. Providers revise usage records after the fact, so finance triggers close on Tuesday and discovers on Thursday that the provider added previously unreported charges. Reserved instance and savings plan discounts get applied at the invoice level in ways the raw usage export does not always reflect cleanly. Marketplace charges, support fees, taxes, and credits show up on the bill but not in the consumption telemetry. Each of these is a place where the usage number and the payable number drift apart, and someone has to explain the difference before the books close.
The cost of not catching those differences is real money. Across enterprise accounts payable broadly, teams catch only about 39% of invoice errors, meaning roughly two-thirds of billing mistakes are paid without challenge. Cloud invoices are not exempt from that. Duplicate line items, misapplied discounts, and pricing that does not match the contract all happen, and at six- and seven-figure annual cloud spend, a one or two percent error rate is a budget line of its own. The FinOps Foundation’s own invoicing and chargeback capability names this as a core practice precisely because so many teams never build a reliable bridge between the bill and the usage.
What FOCUS 1.4 gives you is a standard way to build that bridge once instead of rebuilding it per provider. Invoice Detail is the financial truth. Cost and Usage is the operational truth. The Invoice ID join is the link, and it is now part of the spec rather than a bespoke script your one cloud analyst wrote and never documented.
The commitment data that used to live in a vendor portal
The other meaningful change in 1.4 is the expansion of the Contract Commitment dataset from 13 columns to 30, organized across identification, lifecycle and periods, commitment structure, and cost and quantity. The specification also adds the ability to track which charges are eligible for a commitment, not only which charges a commitment covered.
That distinction matters more than it reads. Today, if you want to know whether your AWS savings plans and Azure reservations are sized correctly, you log into each provider’s console, pull commitment details from a portal that looks different on every cloud, and stitch the picture together by hand. With commitment structure expressed in a common schema, you can compare commitment coverage across providers in one query and see eligible-but-uncovered spend, which is exactly the spend you are paying on demand rates for when you could be committing it. For anyone running a multi-cloud commitment strategy, that is the difference between managing coverage from data and managing it from a quarterly screenshot.
What this does to the monthly close
The practical payoff is the close cycle. FOCUS 1.4 is the first release that lets a FinOps team and an accounts payable team work from the same billing facts. Invoice Detail tells AP what is owed and when. The reconciliation join tells finance whether the usage supports it. The Billing Period dataset keeps period boundaries honest when a provider’s billing month runs to the 27th and yours runs to the 30th.
Put together, that shortens the monthly close and removes the manual stitching that used to sit between AP, finance, and the FinOps tooling. The reconciliation that took my telecom team a week becomes a query that flags the line items that do not tie out, so a human only looks at the exceptions. That is the right division of labor: let the standard handle the 95% that matches, and spend human attention on the 5% that does not.
It also changes who can do the work. When the invoice and the usage live in one schema, a finance analyst who knows SQL can run the reconciliation without waiting on a cloud engineer to explain what a “covered charge” is. FOCUS 1.4 standardizes that vocabulary too, with formal “covering charge” and “covered charge” concepts and cleaner provider naming that retires the old overlapping Provider Name and Publisher Name columns.
What it does not fix yet
Standards are only as useful as their adoption, and this is where honesty matters. The spec was ratified on June 4, but the FOCUS Validator’s support for 1.4 conformance testing is not expected until Q3 2026. Until your providers actually emit 1.4-conformant exports, including the Invoice Detail and Billing Period datasets, you have a schema and not yet a feed. More than a dozen providers support FOCUS broadly, including AWS, Google, Microsoft, and Oracle, but support for the newest version always lags the ratification by months, and historically not every provider implements every optional dataset on day one.
So the realistic timeline is this. The capability is defined now. The exports will trail through the back half of 2026. Native AI support, which finance teams arguably need more urgently given that 98% of organizations now report managing some AI cost, is not in 1.4 at all; model identity and token consumption are slated for FOCUS 1.5 in December 2026, alongside a dedicated Price Sheet dataset.
That means FOCUS 1.4 is a finance-and-AP release, not an AI release. If your pain this quarter is reconciling a sprawling cloud invoice against usage and closing the books faster, this is built for you. If your pain is allocating AI token spend by model and team, the standard catches up at the end of the year.
The move to make now is small and concrete: ask each of your cloud providers and your FinOps platform vendor when they will emit FOCUS 1.4 datasets, specifically Invoice Detail. Put it in the next quarterly business review. The teams that get this feed early are the ones who stop paying cloud invoices they have not actually checked, which is a position most finance leaders did not realize they were in until someone reconciled the bill line by line and found the circuit charge nobody recognized.