Presented by

Want to appear here? Talk with us

This 2026 buyer’s guide breaks down how to compare FinOps platforms that actually deliver results.

Learn:

CIOs are about to face a big surprise with their AI budgets.

Research firm IDC predicts that large companies will underestimate their AI infrastructure costs by 30% through 2027. The problem is simple but serious. AI projects cost money in ways that are hard to predict. GPU prices keep changing based on supply and demand. You pay for tokens, inference, networking, and storage in ways that are hard to track. Security, governance, and training costs add up fast.

5 Things to do about it:

Expand your FinOps team. IDC says most FinOps teams will report directly to the CIO. These teams need to track AI spending constantly, not just once a quarter.

Start with easy wins. Pick AI projects that clearly save money or make money. If your competitor is losing money on an AI project, let them keep losing.

Use smaller AI models when possible. You don't need the biggest, most powerful model for every task. Right-size your AI just like you right-size your servers.

Watch your GPU usage. Many companies use only a fraction of their GPU capacity because of poor scheduling. Better management here can save real money.

Avoid getting locked into one vendor. Use hybrid approaches so you can move workloads to cheaper options when needed.

Read More

FinOps teams can now allocate costs by user attributes (like cost center or department) via IAM Identity Center, enabling sharper chargeback models. Reporting workflows also get a boost with new PDF and CSV dashboard exports for easier stakeholder sharing.

Payment Cryptography pricing has been slashed by up to 63%.

Graviton4 instances (M8gn/M8gb, C8gb) are generally available, delivering ~30% better performance.

More updates: EMR Managed Scaling is now available globally for automated cluster right-sizing. OpenSearch introduces a writeable warm tier to lower storage costs.

Read All AWS Updates

Cloud SQL Enhanced Backups reached General Availability, centralizing retention policies to prevent surprise storage costs.

For capacity planning, Calendar Mode now allows teams to reserve high-demand GPUs/TPUs up to 90 days in advance.

AlloyDB also introduced the C4 machine series, scaling to 288 vCPUs for massive database workloads.

Read All GCP Updates

None :(

Read All Azure Updates

Your voice defines this community. 2025 is wrapping up. It is time to look back. We shared all our achievements and plans for 2026

Access slides

Join us as Katya explores FinOps, sharing tips on how to explain it to teams, adjust practices for different people, and make the best use of cloud resources.

Find out about common misunderstandings, why clear communication matters, and how to successfully teach FinOps in a company.

Learn how to inspire team members, steer clear of common mistakes, and see real growth in FinOps culture.

The House passed a bill on Monday that could save federal agencies up to $750 million each year by fixing how they buy software.

The goal is to stop agencies from buying the same software twice or paying for licenses they don't use. Agencies would need to send reports about their IT purchases to the Office of Management and Budget, the General Services Administration, and Congress.

Think of it like this: imagine your company paying for 100 software seats when only 50 people actually use the program. Or buying three different tools that all do the same job. That's what happens across federal agencies right now. This bill would force them to track what they have and stop the waste.

The bill now sits in the Senate, where six senators introduced it back in June. It hasn't moved since then. The Senate Homeland Security and Governmental Affairs Committee needs to take action before this congressional term ends.

Read More

Glassity has released a five-step framework for FinOps teams who want to move past manual cost management and reach what they call "algorithmic cost efficiency."

The framework tackles a common problem: cloud costs keep growing while it gets harder to connect those costs back to actual business value. Here's how the 5 steps work:

Step 1: Opportunity Discovery: You build a pipeline of cost-saving ideas. This includes automatic detection of waste in your current setup, plus reactive fixes based on how people actually use your systems. You also capture bigger improvements like adding guardrails or building governance into your development process earlier.

Step 2: Opportunity Assessment: You rank every opportunity by risk, effort, and value. The key here is thinking about opportunity cost - what else could your engineers be doing with their time?Each item gets clear context: what's broken, what it costs, how to fix it, and what happens if you ignore it.

Step 3: Stakeholder and Planning Alignment: You assign clear ownership for each optimization. Every opportunity gets a specific person responsible and a concrete remediation plan. The goal is to remove all uncertainty before anyone starts working on fixes.

Step 4: Execution: Because you did the prep work, execution becomes straightforward. Most fixes happen through a pull request - someone reviews the code, it gets merged, and the optimization is done.

Step 5: Feedback Loop and Celebration: You analyze what worked and what didn't. Those lessons create new opportunities that feed back into Step 1. You also celebrate wins publicly - through leaderboards or recognition programs - to keep teams motivated and engaged.

The framework is designed to fit the "Run" phase of the FinOps maturity model, which means it's for teams ready to automate instead of doing everything manually.

Read More

Great article by the Alvin Team. FinOps current approach goes like this: look at a report, talk about what happened, create a ticket, wait for someone to fix it. By the time you act, the money is already gone.

Instead of just showing you what went wrong last month, it should automatically make decisions in real time based on rules you set up front.

Instead of generic cost reports, you'd set specific goals like "dashboard queries must load in under 2 seconds" and "monthly spend stays under $250,000." The system would then automatically route work to the cheapest option that still meets those goals. If ad-hoc queries start eating up too much capacity, the system moves them to cheaper tiers without breaking your dashboard speed promise.

No tickets. No meetings. No waiting.

Here's the uncomfortable truth: cloud vendors make money when you overspend. They profit from the gap between on-demand pricing and committed discounts. They'll give you tools to see your spending, but they won't build systems that automatically minimize it. That would hurt their margins.

Let the system choose the best execution path and prove it worked.Keep humans in charge of setting the rules, but let machines handle the thousands of micro-decisions that happen every hour.

FinOps started by making cloud costs visible, which was necessary but not sufficient.The next step is making those costs self-regulating through policy-driven automation that runs continuously without human intervention.

Read More

Participate by answering our questions. There will be surprise gifts to winners every month!

Adding Music to FinOps!

I met Jill the other day in Madrid and had a wonderful chat in FinOps. Coming from Soundcloud and sharing lots of valuable content. A great new member of our community!

Connect with the Professional of the Week

Azure FinOps Cost Optimization
(Starting February 2026)

Enrollment ends today.

Alfonso is creating a masterpiece.

The biggest educational we’ve done till date.

Enrollment for super early users closes the 21st.

Learn all you need to optimize resources in Azure.

Enrollment Calendar

199€ Super Early Bird – Ends December 21st

299€ Early Bird: Opens XXX - Ends XXX

499€ Launch Offer: Opens XXX - Ends XXX

699€ Starting Offe: Opens XXX - Ends XXX

999€ Final Price

The Perfect Training for Azure Practitioners