Presented by

Want to appear here? Talk with us

What changes when FinOps evolves?

27% wasted spend
AI cost spikes everywhere
Control still comes too late

From reporting → estimate before deploy
From visibility → act with governance
From cost → prove realized value

FinOpsly calls this Value-Control.
This guide breaks down this new FinOps model

Our friend Sam just released an open-source library of specialized AI agents designed specifically for cloud cost work.

The library covers seven main areas. Cloud cost analysis has 10 agents including ones for AWS, GCP, and Azure billing systems. Commitment management has 5 agents for Reserved Instances and Savings Plans. Kubernetes gets 3 agents for container-level cost tracking. Data platforms has 3 agents for building the pipelines behind cost dashboards. Governance has 11 agents covering tagging, chargeback, policy, and team coordination. Waste detection has 5 agents that hunt down idle resources and zombie infrastructure. Specialized work has 5 agents for ML costs, serverless, and carbon accounting.

Each agent knows the FinOps Framework inside and out. They reference which Domain, Capability, and Maturity level they support. The project covers all 22 Framework Capabilities at 100 percent.

The agents work with most modern coding tools. Claude Code, GitHub Copilot, Cursor, Windsurf, and Aider all support them. You can also use them in ChatGPT Custom GPTs or Claude Projects on the web.

Installation is simple. You clone the repo and run a script that copies files to your tool's config folder. No network calls, no special permissions needed.

The library also includes 9 playbooks for common problems. These are named patterns like "Zombie NAT Gateway" or "Month-Length Illusion" that you can reference by name when you hit those exact issues.

Read More

Join us as we're getting into the real talk behind one of the most common — and quietly costly — calls that engineering and finance leaders make: build or buy your FinOps tooling.

11:00 AM | Erik Peterson - CTO @ CloudZero on Build vs Buy in AI era

11:20 AM | Panel: Erik Peterson, Ben de Mora (FinOps OG) and Victor Garcia

11:40 AM | Q&A with attendees

📅​ Date: May 7th
🕗​ Time: 11:00 AM EST / 17:00 CEST
📍 Online

Subscribe to our Events

Brought by CloudZero Cloud Cost Playbook

Cost Optimization Hub now features one-click CSV exports. This allows to instantly download filtered recommendations for offline analysis by passing manual S3 exports.

Compute Optimizer has expanded support to include 162 new EC2 and 32 RDS instance types. By incorporating the latest price/performance data for current-generation families, identifies modernization opportunities and potential savings.

Redshift Serverless now defaults to AI-driven scaling for new workgroups, supporting a minimum footprint of 8 RPU. Rightsizes resources in real-time, reducing over-provisioned capacity.

Broadened regional availability for C8gn (Graviton4) and the launch of 6th-gen Intel Xeon (R8in/M8in) provide up to 43% better performance and 600 Gbps throughput.

Read All AWS Updates

No updates this week ;(

Read All GCP Updates

No updates this week ;(

Read All Azure Updates

Discover how to optimize cloud spend without losing agility. Real-world examples on how to tackle key challenges in data platforms like Databricks, Microsoft Fabric, and Snowflake, including tagging, capacity management, and optimization in multi-tenant environments with Sonal Gupta from Carlsberg

Listen on Apple Podcast

Listen on Spotify

The new IAM Principal-Based Cost Allocation feature automatically tracks which user or role made each Bedrock API call. This information shows up directly in your cost reports without any extra work.

See which teams use the most AI tokens on models like Claude.Compare costs between different applications, like a customer chatbot versus an internal tool. Charge the right department for their actual AI usage. Find ways to save money by looking at who uses what.

When you turn on this feature, AWS adds a new column to your cost reports. This column shows the IAM identity for every Bedrock call. If you tag your IAM users and roles with labels like "department" or "cost center," those tags appear in your cost data too. You can then filter and group costs in Cost Explorer by these tags. Want to see how much the engineering department spent on AI last month? Just filter by the department tag.

Setup: Add tags to the IAM users and roles that call Bedrock. Use simple labels like "department=engineering" or "project=chatbot." Turn on these tags in the Billing console so they show up in reports. This takes up to 24 hours. Enable IAM principal data in your Cost and Usage Report settings.

One important note: This feature makes your cost report files bigger. Each cost line that used to be one row now splits into multiple rows, one for each person or role that contributed. If you have lots of different users, your storage costs for these reports will go up.

Read More

Martin Nanchev built a FinOps agent using Strands SDK and Amazon Bedrock that scans AWS accounts and finds waste automatically.

The agent checks your AWS account in layers, starting with the basics and going deeper: Cost trends over the last six months, with alerts when spending jumps more than $5 or 25 percent in a week.

AWS charges $0.01 every time you call the Cost Explorer API. If every tool made its own calls, the cost of running the agent would add up fast. The code wraps all Cost Explorer requests in two helper functions that strip out credits and refunds so the spending trends stay accurate.

The agent won't tell you to buy Reserved Instances or Savings Plans if your compute spending is below $50 per month.It also checks whether you have Cost Anomaly Detection or Budgets turned on, because most accounts don't, and that itself is a finding worth reporting.

The final output is a PDF report that summarizes everything in one place. This is the kind of tool that pays for itself the first time it finds a $200 per month EBS volume that nobody knew existed.

Read More

A FinOps practitioner just saved their company $16,000 per month by doing something simple: turning off resources nobody was using.

They ran a basic script to check CPU usage on all EC2 instances over 30 days. 23 instances were running at less than 5% CPU. The worst offender was a t2.micro instance that had been running for 14 months at 0.2% CPU usage.

Then they checked for unattached EBS storage volumes. They found 87 volumes just sitting there, costing money every month.Some were from instances that had been deleted two years earlier.

Next up: RDS database snapshots older than 60 days. Over 400 snapshots were piling up, each one adding to the monthly bill.

The fix took two hours. They wrote a 50-line Python script using boto3 that tags obsolete resources and sends alerts to Slack. Now it runs automatically every week.

Other FinOps professionals in the discussion confirmed this is common. One freelancer said they do this exact work for companies and get paid a fraction of what they save. The issue is that most engineering teams are too busy building products to hunt down $10 EBS volumes. So these zombie resources stay active for years.

Read More

The job market is hungry for certified professionals who can prove results. Don't let your company's budget leak due to a lack of specialization.

Use code: FINOPSWEEKLY to get an instant 20% discount on the most prestigious certification bundles:

If you want to be in the room where the big decisions are made, you need to be at FinOps X. We’ve secured preferred access for our community:

Cloud Value Leader | FinOps & Cloud Optimization Executive | Driving Digital Transformation, Scalable Cloud Strategy, Governance & $250M+ Value Creation

Connect with Professional