Presented by

This 2-minute benchmark asks five questions about how you govern AI spend, then shows exactly where companies like yours land against 260 senior finance leaders.

What you'll see:

Google Cloud just released a new open-source tool called the FinOps Agent. It lives in Google's professional-services GitHub repo, ready for teams to use and customize.

Thousands of rows, strange codes, and confusing charts that never explain the real problem. This tool aims to fix that.

- It lets anyone ask plain questions like "what are our top spending projects" and get a clear answer back.

- It reads your BigQuery billing data automatically, no matter how your tables are set up.

- It checks spending against your own company policies, so it flags things like against-the-rules storage setups.

- It builds charts on its own when you ask for them. It pulls everything together into one readable report with clear next steps.

Behind the scenes, it uses several smaller AI helpers working together. One reads your billing data. One reads your policy documents. One builds charts. One writes the final summary.

Google even included sample billing data and sample policy files, so teams can test it before connecting real data.

Read More

How to Choose the Right AI Model on GCP Without Blowing Your Budget. We sit down with Vera to break down how to select the best AI model on Google Cloud's Vertex AI, balancing pricing, performance, and business use cases.

Join FinOps professionals at the FinOps Weekly Summit 2026 and discover how to:

Transform FinOps from reactive cost control into strategic business value

Build scalable unit economics for the AI era

Learn proven strategies from organizations managing billions in cloud spend

📅 October 20 & 21, 2026

Kubernetes has become the default way to run containers. But running it well is a different story. This article shows how much money companies waste without knowing it.

The industry report found CPU usage across clusters sits at just 8 percent. Memory usage is around 20 percent. GPU usage, the most expensive resource, is only 5 percent.

That means most of what companies pay for sits idle. Worse, these numbers are getting worse every year, not better.

The author shares a real case from managing 15 AKS clusters. Nobody could see cost or usage data at all. Some clusters were even running on a pricing tier too low to unlock cost tools. Billing showed the total spend but not which team or app caused it.

This created tension between FinOps teams wanting visibility and cloud engineers weighing the cost of getting it. The fix involves turning on built-in cost tools and monitoring, then testing it on one cluster first.

You cannot control cloud costs you cannot see. FinOps teams should push for basic visibility before scaling any cluster further.

Read More

AWS Lambda promises to bill only for what runs. In real life, most teams overpay because settings from years ago never get revisited.

This guide breaks down a practical way to fix that, and it holds real lessons for FinOps teams.

AWS has a free tool called Power Tuning that tests different memory settings and tells you the cheapest option.

Switching to Graviton2 processors can cut costs by 20% with little to no code changes. Savings Plans can trim costs further, but only if bought based on real usage data, not a single busy month.

Tagging every function matters just as much as tuning it. Without tags, finance teams cannot tell which team or product is driving the bill.

The article also shows how to build this into a GitHub Actions pipeline, so cost checks happen automatically before code ships, instead of after a shocking invoice arrives.

Budgets and alerts close the loop, catching cost creep from traffic spikes rather than code changes.

Cloud cost control works best as an ongoing habit, not a one-time fix.

Read More

Join this FinOps for AI webinar to learn how to control AI costs, improve visibility, implement governance, and justify AI investments with confidence.

📅 August 20, 2026
🕚 6:00 PM Spain / 12:00 PM ET

Billing & Cost Management Dashboards now show a Cost Efficiency widget, making it easier for teams to track savings progress in one place.

AWS Data Exports now include standardized Bedrock product metadata, so teams can sort and track AI model costs more easily.

S3 removed the 30-day minimum for moving data to cheaper storage tiers, so cold data can shift to lower-cost storage right away.

Read All AWS Updates

Google Cloud Batch now supports instance flexibility in Preview, letting jobs list multiple machine types to avoid costly overprovisioning and reduce retries.

Gemini 3.6 Flash is now available in the global region, giving teams a newer, more efficient model option as older Flash models head toward deprecation.

Read All GCP Updates

No major FinOps updates this week.

Read All Azure Updates

Cutting cloud costs by scaling Kubernetes nodes down to zero sounds great on paper. But it comes with a hidden cost: delay.

When traffic hits, spinning up a new node can take 30 seconds to several minutes before a user or app can actually run. For teams watching cloud spend, this tradeoff matters.

Slow starts can hurt customer experience, slow down developer pipelines, and cause missed capacity during big traffic spikes like product launches or sales events.

- Reservation placeholders hold a fixed number of nodes ready, best for planned events with known size and timing.

- Overprovisioning placeholders keep smaller backup nodes ready and get bumped when real work needs the space, best for unpredictable demand like per customer sandboxes.

Both methods still cost money, so teams need to balance speed, user tolerance for delay, and budget.

The team also built an open source tool called Keeper that lets you schedule these placeholders, so you only pay for extra capacity during busy hours and save the rest of the time.

Read More

This article makes a strong case for building cost thinking into products from day one, instead of fixing cloud bills after the fact. The author led a FinOps transformation without a big central team.

Instead, one architect worked with engineering, finance, and business teams to make cost part of everyday decisions. The main idea is simple.

- Cost should sit next to reliability and security as a core engineering goal, not something finance cleans up later.

- Cloud spending is different from old hardware costs.

- You pay only for what you use, but that flexibility can quietly turn into overspending if nobody watches it closely.

For FinOps leaders, this piece offers a practical model. Cost control works best when it starts at the whiteboard, not the invoice.

Read More

Visit FinOps Weekly Hub

Find all Jobs in our FinOps Jobs directory or join FinOps Weekly Jobs Newsletter below to have a weekly email with the most relevant updates every Monday

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_20 to get an instant 20% discount on the most prestigious certification bundles:

Get Certified for Less

Join the Slack community where the real FinOps conversations happen