FinOps X Madness - AI and FOCUS 

About the Author: Sanjna Srivatsa - Product Manager & Data Scientist, CloudHealth | FOCUS Contributor, FinOps Foundation

Now that the whirlwind of FinOps X’26  in San Diego has finally settled, I’ve had a moment to look back on what an incredible week it was and how crazy the weeks leading up to it were. 

In San Diego, I had the honor and privilege to speak with Sangini from Applied Materials about the Evolution of FinOps from Mainframes to now Agentic AI. Adding to the nerves of presenting to the FinOps Community, was the release of two highly anticipated features:our AI & Tokenomics dashboard and FOCUS Dataset.

Before joining CloudHealth as a Product Manager, I was a practitioner and a data scientist. 80% of my job was acting as a "data janitor." I spent my days writing custom R scripts, normalizing mismatched strings, and desperately trying to get AWS, Azure, and Google Cloud billing data to speak the same language. Before FOCUS became a formal thing, I spoke at FinOps Foundations’ April Summit in 2023 about these very real struggles of data normalization. Since then, I’ve been ecstatic to be an active contributor to the FOCUS. It’s been satisfying to see my work solving the issues that I had dealt with on a daily basis. 

Merely three years later, things look entirely different. At FinOps X, we launched the AI & Tokenomics Dashboard, Multicloud FOCUS v1.2 Dataset and Standard Report as part of the new CloudHealth experience, and it feels like the ultimate full-circle moment for me.

Standardization with FOCUS v1.2

With our new report, CloudHealth acts as a Data Generator to normalize your billing data directly. We do not just accept FOCUS exports from cloud providers, but generate and maintain the translation layer in compliance with the version 1.2 spec (with known gaps). 

The FOCUS dataset comes with a  four-level service categorization structure:

• Level 1: Service Category (e.g., Compute)

• Level 2: Service Subcategory (e.g., Serverless compute)

• Level 3: Service Name (e.g., Lambda)

• Level 4: Service Items (e.g., Lambda - Requests—a CloudHealth-generated column for deeper drill-downs)

For our CloudHealth users, if you previously used the v0.5 FOCUS templates via Flex Reports in the classic experience, we strongly encourage you to transition to this new standard report. To get to the new report, navigate to Reports > Standard and search "FOCUS" to get started.

AI Cost with the Tokenomics Dashboard

I'll admit, when the FinOps Foundation first started popularizing the phrase "Tokenomics”, it felt like another buzzword. But seeing our customers’ constantly struggling with growing AI costs over the previous few months, I came around to it. We aren't just managing predictable VM hours anymore; we are tracking non-deterministic token consumption and GPU cluster spend across multiple LLM providers.

The AI & Tokenomics Dashboard tackles this shift objectively by treating AI consumption like the variable financial model it actually is:

• Holistic Multi-cloud AI Visibility: One single view tracking model costs, infrastructure costs (like those incredibly pricey GPUs), and token consumption across providers.

• Multi-Model Cost Tracking: It allows you to track and compare your effective cost per one million tokens across various models.

• Caching and Batching details: Allows you to see batching and caching practices across your organization.

All widgets on the dashboard are backed by a report. You can add perspectives, change categorizations or time granularities in the report.

Getting Started 

FinOps X might be over, but the work is just starting. 

Whether it be managing AI spend or standardizing your organization on FOCUS, CloudHealth has got you covered. Sign up for a Free Trial of CloudHealth and get started putting the things you learned at FinOps X‘26 in practice!