Last year, I talked about how Microsoft is empowering stakeholders across the organization with FinOps resources, tools, and services. It’s been an amazing year with some unbelievable AI-powered advancements that set the stage for an even more amazing future. AI has become a central theme in nearly every corner of business and technology. It’s no surprise that FinOps X 2025 was loaded with news about how AI is changing everything.

Managing AI costs with FinOps

If you’re new to AI or new to FinOps, a great place to start is the FinOps Framework for AI walkthrough. This quick guide offers a high-level overview of the FinOps Framework for AI workloads. From there, check out Understanding AI workload cost considerations for an introduction in some of the factors you’ll need to define the right architectural balance of cost and other metrics to meet your business needs. With these, you should have the foundation for what you need to price out your AI workloads with Azure OpenAI pricing offers.

Regardless of where you are on your AI journey, transparency is key to making smart decisions. And making insights available to every stakeholder is critical to ensuring everyone takes shared ownership for their technology usage. Whether you’re an engineer managing Azure OpenAI costs in the Azure portal or on a business or finance team working out of Power BI or Microsoft Fabric to manage costs and report on unit economics based on token usage (guest perspective), the year has been packed with useful tools and resources to monitor and control your AI spend.

But I have to say, some of the most impactful changes when it comes to FinOps for AI are the introduction of provisioned throughput unit (PTU) reservations and Azure AI Foundry Observability. My favorite part about PTU reservations is that they’re built around you and offer maximum flexibility with 1-month or 1-year reservations. What better way to plan ahead and save on even the newest AI workloads! And with Azure AI Foundry Observability, you have multiple tools all integrated to support monitoring, optimization, and governance, including weighing the cost vs. quality vs. performance of AI models.

And if you’re looking to automate model selection, check out the Azure AI Foundry model router, which can select the best large language model (LLM) to respond to a given prompt in real time by evaluating cost, performance, and query complexity. It’s a brave new world!

Transforming FinOps with AI

And while maximizing ROI for your AI workloads is important, I have to say I’m most excited about advancements in “AI for FinOps”, which is about AI-powered solutions that help you accomplish your FinOps goals. It all begins in the Azure portal, where Copilot in Azure is now generally available! Microsoft Cost Management makes it easier than ever to engage with Copilot with integrated prompts covering new scenarios, like cost estimation and simulation, and the ability to create custom views in Cost analysis. To learn more about what you can do with Copilot in Azure, see Discover cost management opportunities using tailored prompts.

But this is just the beginning. You can also use GitHub Copilot for Azure to answer those same questions directly from Visual Studio Code (or your favorite code editor), bringing FinOps insights right into your development workflow. Additionally, you can simplify and accelerate your journey to the cloud GitHub Copilot App Modernization extension packs for Java and .NET, reducing up-front complexity and costs of migration. And for the power users, GitHub Copilot also enables an innovative Agent mode, which not only answers questions, but can take action for you! Take this even further by connecting GitHub Copilot to FinOps hubs and unleash agentic AI for FinOps in a workspace built for collaboration! Run advanced queries, generate complex reports, and orchestrate data-driven workflows. Forecast your costs, identify anomalous usage and spend patterns, generate code and scripts to automate FinOps processes using the GitHub Copilot coding agent, and more! The sky truly is the limit!

And if you’re looking to bring AI to the masses, ingest your data into Microsoft Fabric, a unified platform that democratizes data engineering for the masses with AI-powered tools, like Copilot, integrated within every experience. And with custom data agents, you can tune that experience to meet your organizational needs and align to your governance policies.

Elevating your FinOps journey with FOCUS

When Microsoft joined the FinOps Foundation, I spoke about the importance of engaging with the community. Over the years, Microsoft has been an engaged member of the FinOps Foundation Governing Board, Technical Advisory Council, several working groups and special interest groups, the FinOps Open Cost and Usage Specification (FOCUS), as well as engaging directly with the FinOps community in Slack. The vibrant, welcoming community is what makes FinOps such a success. If I had to use one word to describe the community, that word would be “family.” And FinOps X is our family reunion where we get to share and celebrate the amazing progress we’ve made both for and with this amazing community.

Last year, we announced FOCUS 1.0 exports in Cost Management. Since then, we’ve seen amazing adoption – more than one-third of all cost exports are FOCUS exports. While adoption has moved much faster than we expected, it’s not surprising, given FOCUS is the best version of our cost data. And with our continued investment in each release, it will only get better. With that, I’m happy to share that a preview of FOCUS 1.2 is now available in Cost Management exports!

If you haven’t tried FOCUS, now is a great time to see what you’re missing. To learn more about FOCUS and its benefits, see What is FOCUS? and explore the many resources to help you understand and adopt FOCUS. For an in-depth walkthrough, check out the Learning FOCUS blog series, which covers every nook and cranny of the specification with useful tips and tricks. And to get started with exports, see Create and manage Cost Management exports.

If you’re already using or plan to use FOCUS, I would encourage you to join the project, share your direct feedback, and help us improve the specification to better meet your needs. Let’s do this together!

Strengthening FinOps fundamentals

There’s a misconception that I still hear today that to adopt FinOps, you need to implement every capability or you must have a central FinOps team. There’s nothing further from the truth. FinOps is about maximizing the ROI you get from technology and the FinOps Framework lays out a map that helps you plot the course for your journey. And with that, we continue to invest in guidance to help educate and enable stakeholders across the organization. From learning modules and documentation to new and evolving FinOps resources that we continually update to align to the latest version of the FinOps Framework, Well-Architected Framework, and Cloud Adoption Framework. And if you need assistance, we can also help you adopt and implement FinOps with specialized offerings from Microsoft Services that cover every aspect of the FinOps Framework – from education and assessments to foundational implementation, operations, and strategic planning. We’re here to help. Reach out to your Microsoft account rep to learn more.

When it comes to FinOps tooling, native provider tools are usually the first place you go for quick and easy answers. I already talked about how Copilot makes it easier and faster than ever to get quick answers with Copilot in Azure—now generally available. If you aren’t using Copilot to answer common questions, you’re missing out. Copilot is continually expanded to support new capabilities, like creating new custom views in Cost analysis, simulating Azure OpenAI costs, troubleshooting disk performance, and more. Cost Management brings Copilot one step closer with embedded Copilot prompts that make it easier to start the right conversation directly from within Cost Management. Add on data quality and completeness improvements, new smart views, like Azure OpenAI costs in Cost analysis, and pricing updates spanning virtual machines, Azure AI, Microsoft Fabric, and more, and there’s plenty to explore!

Looking beyond Cost Management, Azure Carbon Optimization is now generally available, giving you one place to view carbon emissions across services in the Azure portal. Paired with this is new sustainability features in Azure Migrate, which helps you plan your on-premises to cloud migration. With application-aware assessments, integrated sustainability metrics, and a new user experience with enhanced guidance, you’ll want to see how Azure Migrate can help you quantify the value of migrating your workloads to the cloud. To learn more, see Azure Migrate Build 2025 updates.

But that just covers some of the core FinOps tools. We’d need an entirely separate blog post to cover all the updates from all Microsoft Cloud services. I’ll mention just a few and encourage you to follow the Cost Management updates for more.

• Azure Backup reduced protected instance fees for SAP HANA backups.


• Azure Cache for Redis added a smaller enterprise tier cache instance.


• Azure Cosmos DB added dynamic scaling and autoscale for vCore-based MangoDB.


• Azure Database for PostgreSQL added new Advisor recommendations and V5 reservations.


• Azure Databricks added serverless jobs to reduce unused capacity costs.


• Azure DevTest Labs added VM hibernation to reduce costs by pausing VMs while preserving state.


• Azure Kubernetes Service added new Advisor recommendations for rightsizing, autoscaling, and more.


• Azure Monitor Log Analytics added summary rules to aggregate data, minimize cost, and speed up query performance.


• Azure NetApp Files added support for 50 GiB minimum volume sizes, storage with cool access for all service levels, and reservations.


• Azure SQL Managed Instance improved the free offer.

If all that isn’t enough, the FinOps blog also has you covered with guidance to control costs with Azure Policy, cut costs with Azure Hybrid Benefit for Linux, optimize storage and VM costs, and more.

Of course, Azure service costs are only part of the equation—SaaS costs are also critical to a holistic FinOps practice, with the added benefit of Marketplace SaaS charges accruing towards your Microsoft Azure Consumption Commitment (MACC), effectively maximizing your rate negotiation power. And now, with flexible billing for private Marketplace offers and the ability to lock in marketplace terms for up to five years, you have more options than ever to strike the right deal with your SaaS vendors that aligns cloud spending with your business needs.

These enhancements help ensure your teams have the clarity and flexibility they need to manage budgets, allocate costs, and drive accountability across the organization. With all these updates, you have more ways than ever to manage, optimize, and report on your cloud investments directly in the Azure portal and other native Microsoft tools.

Unlocking advanced analytics

While the Azure portal makes it easy to get quick answers, sometimes you need more. Whether you need to integrate cost data into an existing system or are simply looking for more flexible reporting, exports offer a reliable, scalable solution that pushes data to you, simplifying data acquisition and ingestion. And with the general availability of Cost Management exports, you can get your data where you need it, in the format you want with support for efficient parquet and compression formats—making it simple to move large datasets into your analytics platform of choice. The new exports platform is available across every Microsoft Cloud environment (including Azure Government and Azure China) supporting Enterprise Agreement, Microsoft Customer Agreement, and Microsoft Partner Agreement accounts and subscriptions. Based on customer feedback, exports also include new AccountId and InvoiceSectionId columns in the latest dataset versions to help allocate costs more precisely across your business.

If you need to manipulate or augment the cost data, Microsoft Fabric provides a unified platform for analytics with all the tools you need to integrate and join with other cloud and business data. Sign up for the Cost Management export to Fabric private preview to push data directly into a Microsoft Fabric lakehouse, removing the need for Azure storage.

If you’re looking for lightweight reporting on data across accounts and subscriptions, the FinOps toolkit now offers Power BI reports that connect to raw exports in storage. This is the quickest way to get up and running in Power BI where you can build more advanced reports with the latest data from Cost Management. FinOps toolkit Power BI reports have a fresh, new look and feel and offer expanded coverage of invoice reconciliation and commitment discount utilization, reservation break-even point, cloud governance, and more.

For more advanced reporting with larger accounts or multi-year reporting needs, FinOps hubs offer an open, extensible platform that scales to meet your needs. FinOps hubs provide scalable analytics that can now be hosted on top of Azure Data Explorer or Microsoft Fabric Real-Time Intelligence. FinOps hubs add network security, a version-agnostic platform that protects you from compatibility issues, multi-tenant, and multi-account support using an extensible platform that can also support cross-cloud reporting needs and supports a broad set of reporting and data connectivity options. FinOps hubs take Cost Management FOCUS exports and perform additional data cleansing to offer the best version of FOCUS data available, using a platform with built-in capabilities like forecasting and anomaly detection. And when you connect FinOps hubs to Microsoft Fabric, you also have a unified platform for self-serve analytics with an AI-powered, SaaS-like experience that streamlines data engineering. Microsoft Fabric offers Copilot integration within every experience, making it effortless to build the solutions and get the answers you need.

You can also leverage GitHub Copilot or your own AI agents to bring agentic AI to your FinOps practice. With FinOps-tuned AI instructions and prompts, GitHub Copilot Agent mode can answer questions about your data, build detailed reports, or orchestrate and automate complex FinOps scenarios.

FinOps hubs is built for the community, by the community to help organizations evolve their FinOps practice and gain the transparency and accountability needed to meet business needs.

Next steps

I hope you’re as thrilled as I am about the transformative changes we’re witnessing in the realm of FinOps, especially with the advent of AI. The evolution of FinOps in the era of AI is set to revolutionize how we manage cloud and technology innovations, bringing unprecedented efficiency and insights. This is just the beginning, and the potential is limitless. Stay tuned for more updates on the FinOps blog as we continue to explore and innovate in this exciting journey.