Presented by

Want to appear here? Talk with us

The September AI Economics Pulse tracks 430 CloudZero customers' real cloud bills. AI now takes 2.66% of the median company's bill, roughly four times its level a year ago, and the top quartile just crossed 11%.

Inside this month's Pulse:

A FinOps professional found a hidden cost trap that many teams miss when they try to save money on cloud storage.

He was auditing S3 buckets and found one costing $12,000 a month. It stored old logs for a tool called Coralogix.

The data looked like a quiet archive, but it was getting hit with 44 million file requests a day.

Moving that data to a cheaper storage tier, like Glacier, would normally cut costs.

But because users could run loose searches across months of data, cheaper storage would have caused huge retrieval fees instead.

The real fix was not just a storage change. He worked with teams across the company to set a 180 day limit on how far back most users could search.

Only a small group of admins kept full access for real emergencies.

Once that limit was in place, older data could safely move to Glacier without risk of surprise bills.

The result was a savings of $114,000 a year, with almost no change to daily work for most employees.

Cost problems often sit between two systems, not inside just one. A high number of requests rarely means real business need.

It usually means no one set a limit. Finding that missing limit is often where the real savings hide.

Read More

A live conversation with me on the cloud waste your cost tooling can't see. I break down audience-submitted waste live in the closing segment.

📅 September 23rd, 2026
🕚 5:00 PM Spain / 11:00 PM ET

CloudFront flat-rate plans now support API management, letting teams automate flat monthly pricing setups for more predictable delivery costs.

AWS Marketplace private offers can auto-renew, keeping negotiated prices and terms in place while giving buyers opt-out controls to avoid surprise renewal costs.

Redshift RG.Large now supports single-node clusters, giving smaller workloads a 30% lower price per vCPU for testing and proofs of concept.

Read All AWS Updates

Reservation sharing changes are now GA, letting teams share or restrict compute reservations to cut down on unused capacity spend.

Cloud Run jobs can now delay their start for lower cost, letting non-urgent jobs wait up to 12 hours to use cheaper pricing.

Read All GCP Updates

Microsoft shares new tips to cut AI agent costs, showing how smarter context handling can lower token usage without hurting quality.

Copilot Observability Agent now works with cheaper log tiers, so teams can use lower-cost Basic and Auxiliary log plans and still get AI-powered insights.

Read All Azure Updates

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 - 9AM ET / 3PM CEST

We explore Breaking Silos in FinOps and how organizations can bring Engineering, Product, Finance, FinOps and Leadership together to improve cloud cost management.

FinOps did not start with the cloud. Companies have tracked technology costs since the mainframe days.

What changed is how fast tech choices turn into money spent. Cloud let developers spin up resources in minutes, so costs moved just as fast as decisions.

AI is now speeding things up again. A single AI request can trigger many hidden steps, like model calls, searches, and retries, each with its own cost.

Software can now decide on its own to run another search or call another tool, which means the system itself is making spending choices.

This means teams need budgets and limits built right into AI workflows, not just monthly cost reports.

The better question is cost per outcome, like cost per case solved or cost per fraud caught.

That connects spending directly to business value.

FinOps has always been about understanding technology costs.

The next challenge is understanding the economics of machines that can think and act on their own.

Read More

Every FinOps team knows this story.

A test VM gets spun up for a quick check. Then everyone forgets about it. The bill arrives weeks later, way higher than expected.

Instead of waiting for someone to check a dashboard, the system checks itself on a schedule.

It pulls current spend straight from the Cost Management API. If spend crosses a set limit, two things happen automatically.

First, an alert goes out to notify the team.

Second, the idle VM gets powered off, no human needed.

The setup relies on managed identity, a way for the Logic App to prove who it is and act only where it is allowed.

This keeps things secure without hardcoded passwords or keys floating around.

Fixed thresholds and one hardcoded VM work fine for testing.

But a live system would need smarter thresholds and tag-based rules to scale across many resources.

Cost control works best when it happens automatically, not when someone has to remember to check.

Automating the boring stuff is often the smartest way to protect a budget.

Read More

A production Azure workload was costing about ₹97,000 a month. The team found waste hiding in plain sight.

Cosmos DB was using provisioned throughput but barely getting used.

A virtual machine was much bigger than needed. A Premium SSD disk was paying for speed nobody used.

And there were no budgets or alerts to catch any of this early.

The fix was moving Cosmos DB to serverless mode. This dropped that single cost from ₹38,900 to about ₹109 a month.

The disk was switched to a cheaper tier too. Total monthly cost fell to about ₹57,100, a 41 percent savings.

The team set clear pass or fail rules before touching production.

- They rehearsed the data migration first and caught a throttling problem before it could cause a real outage.

- They checked that every document, index, and count matched exactly before trusting the new system.

- They kept the old database running for 72 hours just in case.

- They also added budget alerts so future cost creep gets caught early instead of showing up as a surprise bill.

Big savings are possible without breaking anything. But only if you measure first, test the risky steps ahead of time, and build a safe way back if something goes wrong.

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

Use code: FINOPSWEEKLY_20 to get an instant 20% discount on the most prestigious certification bundles:

FinOps Certified Practitioner

FinOps Certified Engineer

FinOps Certified FOCUS Analyst

FinOps for AI

Get Certified for Less

Join the Slack community where the real FinOps conversations happen