Presented by

Want to appear here? Talk with us

Get the 2026 FinOps in the AI Era report to see how leading teams are:

Download the Report

Google Cloud just made their committed use discounts much easier to understand. If you've ever squinted at your cloud bill trying to figure out your actual savings from CUDs, you're not alone.

Instead of showing the full price, then adding a commitment fee, then subtracting a credit, you now just see the discounted price you actually pay. Old way: Your bill showed $10 for the service, plus a $5.50 commitment fee, minus a $10 credit, leaving you to calculate that you paid $5.50. New way: Your bill just shows $5.50.

The changes also expand CUD coverage to Cloud Run and new VM types like H3 and M-series machines. Google fixed a reporting problem too. If you used both Flex CUDs and GKE or Cloud Run CUDs together, the system couldn't tell which discount came from where.

The CUD Analysis tool got better. You can now see hourly data for up to 30 days instead of just daily averages. This matters because a CUD might look 80% used for the day, but if you check hourly, you might see it was 100% used during work hours and 0% used at night.

For teams that want to dig deeper, there's a new metadata export you can pull into BigQuery. You can also download CSV files directly from the analysis view without setting up the full export.

The recommendation engine now includes more services when suggesting how much commitment to buy. You can model different scenarios by adjusting coverage levels, filtering out unusual usage days, or looking back up to 180 days.

If you're managing cloud costs, check your billing console to see when your account migrated to the new model. The CUD Analysis tool will show you that your savings stayed the same before and after, just presented more clearly.

Less time doing billing math means more time actually optimizing your cloud spend.

Read More

Network Firewall price reductions now extend NAT Gateway hourly and data processing discounts to service-chained secondary endpoints. Additionally, AWS has eliminated data processing charges for Advanced Inspection (TLS).

Amazon Athena introduced 1-minute Capacity Reservations and lowered the minimum reservation to 4 DPUs.

OpenSearch Serverless Collection Groups allow multiple collections to share OCU capacity while maintaining distinct KMS keys.

Read All AWS Updates

Compute flexible CUDs have expanded to all Cloud Billing accounts. Google automatically migrated accounts to this spend-based model, broadening coverage across Compute Engine, GKE, and Cloud Run.

Capacity Planner (preview) now includes visibility for Cloud Storage egress bandwidth and Spot GPUs.

Read All GCP Updates

Agentic Cloud Operations with Azure Copilot introduces optimization agents that continuously identify and execute improvements across cost, performance, and sustainability.

Read All Azure Updates

Topics we’ll explore:

Speakers & Host:

Date: February 26th, 6PM CET / 12PM EST

We talk about cloud billing & finance management, accounting with a FinOps specialist from British Airways. Magda & I cover cloud billing, invoicing reconciliation, understanding cost allocation, handling cost anomalies, and best practices for creating actionable financial reports.

Listen on Apple Podcast

Listen on Spotify

When your systems go down, the costs start adding up fast.Most companies track the obvious losses like lost sales and angry customers. But there are hidden costs that can hurt just as much or more.

Think about what happens during an outage. Your engineering team drops everything to fix the problem. That means they stop working on new features and improvements. Projects get delayed. Deadlines get missed.

Your support team gets flooded with calls and emails. They need to explain what happened over and over. This takes time away from helping customers with normal issues.

Your reputation takes a hit too. Customers start to wonder if they can trust your service. Some might look at other options. Getting them back costs money in marketing and sales efforts.

Insurance can help cover some downtime costs. But most policies only cover direct losses. They do not pay for the hidden costs like lost productivity or damaged reputation.

The best approach is to prevent downtime before it happens. Track all your costs when outages do occur. This includes both the obvious and hidden expenses. Use this data to make the case for better systems and processes.

Read More

AWS is charging companies up to a million dollars a year for something that should cost nothing. When your private subnets need to talk to AWS services like S3 or DynamoDB, the default setup routes everything through a NAT Gateway. AWS then charges you $0.045 per gigabyte for traffic that never leaves their own network.

The numbers add up fast. A busy production account pushing 2 terabytes per day to AWS services pays about $2,700 per month just in data processing fees. That's $32,000 per year for one NAT Gateway.

Most companies run multiple NAT Gateways across different availability zones and VPCs for reliability. Ten VPCs with three availability zones each, processing 5 terabytes per month, costs $101,000 per year. The fix costs zero dollars for most of this traffic.

AWS offers something called Gateway Endpoints for S3 and DynamoDB. They're completely free. No hourly charge. No per-gigabyte charge. You just add them to your VPC and traffic automatically routes through them instead of the NAT Gateway.

When teams audit their NAT Gateway traffic, they find the same pattern every time. Between 60 and 80 percent of traffic goes to AWS services that support these free endpoints. Read more to fix it in your environment

Read More

Great deep dive by Alberto. Every FinOps team eventually runs into the same problem. You need clean cost data to make good decisions. But cloud billing data is messy, inconsistent, and full of surprises.

That's why FOCUS (FinOps Open Cost and Usage Specification) sounds so good. It promises to normalize your cost data across all cloud providers. One standard format. One shared language. Less time cleaning data, more time on strategy. But there's a catch nobody mentions until you're knee-deep in the work.

FOCUS 1.2 includes 57 columns. That's 57 different data fields you need to understand. Some are mandatory. Some are conditional. Some only matter if other columns have data.

For someone building a FinOps practice, this is overwhelming. You're already learning cloud billing, aligning stakeholders, and trying to show value. Now add 57 columns with complex relationships.

Here's what happened to one practitioner working with AWS data. They wanted to use ServiceSubcategory, a column that organizes services into groups like "Compute" and "Virtual Machines." The FOCUS spec says this column must not be null. AWS lists it as populated in their documentation.

But when they ran the query, every single row was empty. Not null. Just empty strings. Thousands of rows of useless data. The same thing happened with ResourceType. Technically populated. Practically useless. And nowhere in the documentation did AWS mention this behavior.

Think about what happens next. You build a dashboard using ServiceSubcategory. You share it with executives. They make decisions based on those service categories. Then someone notices the data is wrong.

This happens because cloud providers can't or won't fully follow the FOCUS spec. They're supposed to publish gap reports that explain their deviations. But those reports often miss the details that actually matter to practitioners.

FOCUS is still better than working with raw billing data from each provider. Having a common structure helps teams communicate. But for FOCUS to really work, cloud providers need to be honest about their data.

Read More

AWS King of Contracting & Content

No words needed :)

Connect with the Professional of the Week

Special Discount Code only for FinOps Weekly Readers

Naming conventions is essential to avoid a Cloud mess.

Tagging is one of the core skills to govern Cloud Resources.

Cloud Governance is a core skill for FinOps.

Learn both in our course Tagging & Naming conventions with our discount code only for FinOps Weekly Readers.

Use code “FW2026” for 30% off in our course.

P.S. Azure FinOps Course already started. Thanks everyone who joined there.