TOGETHER WITH KION
Is Your Cloud Spend Working as Hard as You Are?
Cloud, AI and SaaS costs don't have to be a mystery.
The GigaOm Radar for Cloud FinOps evaluates 24 solutions across key criteria: from cost optimization and forecasting to governance and AI-driven insights.
So you can cut through the noise and make smarter decisions.
Whether you're just starting your FinOps journey or scaling a mature practice, this independent research helps you benchmark what good looks like.
ITAM
FinOps & ITAM: How to make them work together
Why FinOps & ITAM Must Team Up?
Well, vendors are aggressively blurring the lines between static software licenses and pay-as-you-go cloud usage.
Operating in silos guarantees wasted spend. To regain control, FinOps and IT Asset Management (ITAM) teams must actively share data. Here is how to make this partnership work:
Planning & Procurement: Stop shadow IT early. Model costs together before signing contracts and set strict rules for marketplace purchases so they don't conflict with existing Enterprise Agreements.
Deliver & Govern: Prevent surprise SaaS overages. Track SKU consumption weekly, enforce tagging from day one, and use real-time inventory controls to shut down forgotten resources.
Optimize Cost, Risk, & Value: Look at the exact same numbers. Connect your configuration databases directly to your cloud billing platforms to instantly spot duplicate purchases and unused bundles.
Want to run FinOps Events?
Apply here to become a FW Regional leader
CLOUD PROVIDERS
AWS Sustainability Console & GCP CUD Modeling
AWS
Sustainability Console: Exportable regional/service carbon estimates via API/CSV for easier ESG reporting and chargebacks.
CloudWatch Logs: Centralize logs by data source type (e.g., VPC Flow Logs) to reduce mapping overhead and aggregation costs.
Athena Capacity Reservations: Expanded regional availability offers dedicated capacity to isolate workloads and forecast spend.
Marketplace Self-Service: Sellers can now issue refunds and cancel agreements directly, streamlining billing reconciliation.
Google Cloud
CUD Scenario Modeling: New tool to simulate spend-based and resource-based commitments to maximize purchase-driven savings.
TPU7x (Ironwood): Generally available next-gen AI accelerators providing high-performance alternatives for large-scale training.
Azure
No Updates this Week :(
ONLINE EVENTS
Summit: FinOps Governance Playbook for Cloud & AI Infrastructure
Standard cloud optimizations: reserved instances and basic rightsizing, are no longer enough. With the rapid rise of AI infrastructure and GPUs, cloud spend is outpacing traditional playbooks.
Join us to discover how leading teams are moving from reaction to prevention
📅 Date: April 16th
🕗 Time: 11:00 AM EST / 17:00 CEST
📍 Online
MEETUPS
FinOps Weekly LA Meetup April 9th
Join us at the heart of Hawthorne as we kick off our first-ever LA FinOps Weekly Meetup at the lively Common Space Brewery!
Hosted by FinOps Weekly Regional Leader Diana Molski, this event is your chance to dive into the world of FinOps while enjoying good company and great brews.
We promise an evening of engaging conversations and a relaxed vibe.
📅 Date: April 9, 2026
🕗 Time: 5:00pm PT
📍 Common Space Brewery | Hawthorne, California
FINOPS
How AI Will Change FinOps?
We analyze how GenAI is changing the rules of FinOps. We break down the new skills required for the FinOps Persona and how to decouple experimentation spend from production costs.
KUBERNETES
Rightsizing vs Autoscaling
Amazon EKS bills you for the nodes you run, not the work those nodes actually do. Two tools promise to fix this: right-sizing and auto-scaling. Most teams turn on Cluster Autoscaler, watch their node count drop a bit, and then wonder why the bill barely moves. The reason is simple: these tools solve different problems, and using just one leaves most of your savings behind.
The order matters more than the tools. Right-sizing should come first. Start by running VPA in recommendation mode for at least a week to capture your real usage patterns. Apply those recommendations with some extra room for bursts. Cluster Autoscaler can finally remove nodes that are truly underused. Then tune your HPA thresholds because they work as percentages of requests, and your requests just changed.
Fargate changes the math. EKS Fargate bills per vCPU-second and GB-second of your actual pod requests. There are no nodes to manage, so Cluster Autoscaler does not apply. Right-sizing requests directly cuts your bill since you pay for what you request. HPA and KEDA still work to match pod count to demand.
Right-sizing gives you immediate, permanent cost reduction by fixing structural waste in how pods claim resources, while auto-scaling gives you proportional savings when your traffic actually varies throughout the day.
AWS
Cut AWS Waste Across Accounts
An SRE at a Korean tech company found $12,000 per year in AWS waste across four accounts in just two weeks. None of the work touched production systems.
The company had four AWS accounts that nobody had audited since they were created. Old projects were still running. Logs from deleted systems were still collecting. Six WorkSpaces that nobody used were costing $525 every month. The biggest wins came from three areas.
Abandoned infrastructure was the gold mine: A cancelled food project left behind an entire VPC with NAT Gateways, Elastic IPs, and EC2 instances. Nobody was using any of it. Cost: $64 per month.
CloudWatch logs that never expire: AWS sets log retention to "never expire" by default. Logs grow forever unless someone sets a policy.
S3 storage without lifecycle rules: Three buckets held 1,883 GB of data. Most objects were older than 90 days and rarely accessed.
🎖️ MENTION OF HONOUR
Solving the FinOps January Problem
Cloud vendor credits can make your financial reports look like a roller coaster, and that's a problem when your engineering team starts asking why the numbers vanished. John Grubb calls this "the January problem" though it can hit any month depending on your contract terms.
Here's what happens: Cloud vendors like AWS and GCP give you big promotional credits as part of long-term deals, and these credits can zero out your bill for weeks at a time.
The core issue: When a vendor drops $600k in credits on your account in February, your bill looks artificially low that month. The other 10-11 months look inflated by about 5% compared to what you're actually paying.
The fix: Amortize those credits: Calculate a 12-month moving average of your promotional credits. Divide that average by your monthly costs to get a discount percentage. Apply that percentage every month instead of taking the full credit hit in one billing cycle.
What this looks like in practice: Say you spend $12M annually and get $600k in credits. Instead of showing two weeks of near-zero costs, you reduce every month's bill by roughly 5%. This gives you what Grubb calls "effective cost" - the smoothed-out number that reflects what you're really paying over time.
The technical approach: The article includes detailed SQL for GCP billing data that separates promotional credits from other discounts like committed use discounts.
PROFESSIONAL SPOTLIGHT
Mike Fuller
CTO and Founding GB Member, FinOps Foundation
Become a Tagging Expert
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.
