TOGETHER WITH CLOUDZERO
AI Spend Limits And Capitalization: FY2027 AI Budgets Webinar Recap
CloudZero's webinar recap follows Emily Allen, CloudZero's Director of FP&A, through a year in which AI went from one software line item to a bill every team in the company runs up. She had approved an engineering budget bump that same morning, two days before quarter end. It took five minutes.
What she and Brady Lenahan covered:
Hard caps backfire twice. People who hit them stop working, and people whose cap is too low move the work to personal accounts finance can't see
Smart caps instead: a baseline by role, Slack alerts at 75%, and increases tied to what the team is shipping
Capitalizing engineering AI spend by accounting policy, down to the feature, with a screenshot the auditors accept
Run-rate budgeting is out. AI is now budgeted bottom-up by project, each one held to an outcome the board can see
FINOPS TOOLING & AUTOMATION
Tired of AWS Billing Surprises? How I Built a Multi-Account Cost Reporter in Python
AWS’s advice for FinOps teams: use AI to speed up the work, not replace human checks. A 25% cost rise could mean waste, or a planned move to the cloud.
Acting before you know the cause can waste time and weaken trust.
AWS reports that one finance team cut cost variance analysis from 6–8 hours to about five minutes. People still checked the results and added team feedback. To put this into practice:
Give each AI project a clear spend owner and set limits early.
Cut low-value alerts so teams can focus on real issues.
Check forecasts for one-time costs that could skew the results.
Get the right approvals before letting tools change cloud resources.
Rank savings ideas by dollars saved and staff time needed.
Track business results, such as bugs fixed or features shipped—not just AI use.
AI also adds costs, from model use to GPU time. Those costs need the same care as other cloud spending.
CLOUD PROVIDERS
AWS cuts EMR on EKS fees 30%, Bedrock tokens 50% cheaper, GCP adds Hyperdisk CUDs
AWS
Amazon EMR on EKS cuts its charge by 30%, lowering both the per-vCPU-hour and per-GB-hour fees. EKS pricing stays the same.
OpenAI GPT-6.1 Terra is now on Amazon Bedrock, with token prices at least 50% lower than similar older models.
Amazon ECS now shares one GPU across many tasks with NVIDIA MPS, so you waste less GPU power and pay less.
Read All AWS Updates
Google Cloud
Committed use discounts now cover Hyperdisk storage (Balanced and Throughput Exapool), so you pay less for storage you know you will use.
Agent Platform Workbench now supports custom organization policies, letting teams allow or block instance creation to stop sprawl and extra spend.
Read All GCP Updates
Azure
Claude model costs now show up in Azure Cost Management for Microsoft Foundry, with tags and attribution for clear chargeback and showback.
Network Watcher VNet flow log filtering is now GA, so you log only what you need (by IP, port, protocol and more) and store less data.
Read All Azure Updates
EVENTS
FinOps Weekly Summit 2026
The first sessions from FinOps Weekly Summit 2026 are now available!
Catch the key insights, expert perspectives, and actionable strategies from our speakers.
Watch the sessions now and stay ahead in the evolving world of FinOps.
EVENTS
100th Edition FinOps Weekly Event
We’re bringing the FinOps community together for a special end-of-year event filled with inspiring talks, meaningful conversations, and new connections.
This is your chance to connect with fellow FinOps professionals, share ideas, exchange experiences, and celebrate everything our community has achieved.
Be part of this special milestone and help us make our 100th edition an unforgettable experience!
CLOUD FINANCIAL MANAGEMENT
AU Gov Cloud Guidance Backs FinOps: 5 Takeaways 2026
A lower cloud rate will not help your budget if usage keeps rising.
Australia’s Digital Transformation Agency puts that problem at the heart of its cloud guidance, released on 21 October 2026.
The guidance supports a policy that already requires non-corporate Commonwealth agencies to manage cloud costs. It recommends FinOps practices, but does not require a specific tool or framework. The main actions are clear:
Give each major cloud bill an owner and link spending to a business unit.
Compare actual costs with forecasts and set clear limits for when leaders must step in.
Track AI use early, since a small trial can become costly when thousands of staff start using it.
Plan when to shut down old systems so the agency does not keep paying for both.
Include data export, moving costs and exit tests in provider plans.
The policy is mandatory; the guidance offers ways to meet it. Teams can choose methods that fit their size and spending.
CLOUD COST OPTIMIZATION
The Easy Button for AWS Account Details and Cost Savings
Your AWS resource list may look lean while your bill tells a different story. David Ernst tested an AI agent that compared account details with billing data to find costs worth checking.
The account’s monthly bill rose from under $4 to $335. QuickSight and Neptune made up 98% of the charges, yet neither appeared in the first resource scan.
QuickSight charges included a $40 author seat and a $250 Amazon Q add-on, with no billed SPICE usage.
Neptune had already been deleted, but past bills still showed its charges.
Eleven log groups had no expiry date, though their small size made them a low priority.
The access rules matter as much as the findings. Ernst used a limited AWS role, blocked access to data and secrets, required human approval for tool calls, and checked activity through CloudTrail. He also found that some tool setup guides included write permissions.
🎖️ MENTION OF HONOUR
Showback, Chargeback, and Allocation: The FinOps Ladder Nobody Gets to Skip
Shannon’s guide makes the case for fixing cost allocation before moving money. Allocation links costs to teams or products. Showback lets teams see those costs without changing their budgets.
Chargeback puts those costs on each team’s budget, but it is optional—not a sign of a stronger FinOps practice.
Start with these steps:
Run showback first so teams can flag errors before finance moves money.
Split shared costs using actual use, such as GPU-hours or AI tokens, rather than equal shares.
Write down how you handle shared support costs and cloud discounts.
Track allocation coverage: the share of spend assigned with confidence, including costs kept central under a clear policy.
Tag fixes left coverage near 83%, while adding usage data for a shared AI service pushed it above 90%.
Give someone ownership of costs with no clear team, and require tags when resources are created.
Build trust in the cost report before turning it into a bill.
We Are More than a Newsletter
FinOps Learnings (30% off only for members)
Hands-on on-demand workshops led by FinOps Experts, with a limited discount for members. Learn about:
Discounts on FinOps Certifications
Use code: FINOPSWEEKLY_20 to get an instant 20% discount on the most prestigious certification bundles:
FinOps Certified Practitioner
FinOps AI Value
FinOps Technology Value
FinOps Certified Engineer
FinOps Certified FOCUS Analyst
FinOps Weekly Community
Join the Slack community where the real FinOps conversations happen
FinOps Jobs
Join our FinOps Weekly Jobs Newsletter to have a weekly email with the most relevant updates every Monday
FinOps Hub
Find the right FinOps solution for your company in our Hub
Want to appear in FinOps Weekly? Talk with us