New FinOps FOCUS Specification Enhances AI Cost Accountability

The introduction of the FinOps FOCUS specification marks a significant shift in how enterprises manage AI costs. As organizations increasingly depend on cloud services and artificial intelligence, accurately tracking expenses across multiple providers has become a complex challenge. The FOCUS specification aims to establish a unified language for these costs, enhancing accountability in an environment where spending can escalate rapidly.

Karl Kraft, a senior manager of software engineering at Walmart Inc. and a key contributor to the FOCUS initiative, shared insights during a recent interview at the FinOps X 2026 event. Kraft discussed the traditional challenges faced by FinOps teams, recalling his early experiences with inconsistent cost data: "When I joined the FinOps practice back before it was called FinOps, I was handed some Excel files of our cost and usage from our cloud service providers — two different cloud service providers. I looked at it and thought, what am I going to do with this? … I don’t want everybody to have to go through that pain again and again."

The FOCUS specification seeks to alleviate such issues by establishing a common framework that allows for straightforward comparison and understanding of cost data from various providers. This is especially important as enterprises grapple with rising costs linked to AI usage. The specification introduces an 'effective cost' column that breaks down large monthly fees into daily rates, giving teams a clearer view of actual usage and expenses. Kraft explained, "There is the effective cost — it’s a new column. I can see over the month, my $30,000 is a thousand dollars a day and who consumes it each day, rather than just a big upfront cost that … I don’t know what to do with it."

A key feature of the FOCUS specification is its approach to Kubernetes workloads. The recent version 1.3 has implemented provider-defined split cost allocation, allowing cloud providers to deliver pod-level cost data directly within billing feeds. This development removes the need for teams to manually track and instrument open-source projects, simplifying processes and saving valuable engineering time.

As the development cycle for version 1.5 moves forward, the FOCUS working group is concentrating on standardizing how cloud providers communicate pricing options. Currently, even a single provider may offer multiple pricing variants for datasets, which can lead to confusion and inefficiency. The upcoming version aims to introduce a stock-keeping unit pricing dataset to clarify these communications.

https://www.youtube.com/watch?v=0znMoxxtDBM

The urgency to refine the FOCUS specification is heightened by the evolving demands of AI operations. The community is actively working on integrating AI agents into its development workflow to help identify inconsistencies across the extensive specification. Looking ahead, version 1.5 plans to introduce features that allow organizations to break down AI spending by token type and workload. This level of detail will be crucial for practitioners who need to link model inference costs back to the specific teams responsible for them.

In an era where AI continues to transform technology, the FOCUS specification signifies a major step toward improved financial oversight in the cloud. As enterprises manage the complexities of multi-cloud environments, a standardized approach to financial operations will become increasingly essential. With the FOCUS initiative, companies can anticipate enhanced clarity and efficiency in their AI-related expenditures, potentially leading to more strategic resource allocation and long-term cost savings.