Is FinOps Ready to Handle AI Costs?
Is our practice fully developed to handle the AI revolution?
Victor Garcia
August 24, 2025
Together with Amnic
20 Minutes. 100% Clarity of your FinOps Processes
Everything you need to provide you a 360º view of your FinOps processes.
The Enterprise FinOps Acceleration Kit is a 20 minute assessment that gives you a clear view of your roadmap, governance efforts, maturity state and AI adoption readiness.
AI
Can FinOps Control AI Costs?
AI is eating the world, and budgets along with it.
AI costs work differently than regular cloud costs. Traditional cloud services are easy to predict. You can plan how much a database will cost or how much storage you need. AI doesn't follow these simple rules.
Here's what makes AI costs tricky:
Training costs jump wildly with bigger datasets, and bills spike when more people use the system.
Expensive GPUs often sit idle at only 15-30% usage because they're not matched well to the work.
New AI models keep changing what hardware you need, breaking cost predictions.
Things like tokens and context windows drive costs in ways that old measuring tools don't catch.
Teams still struggle to understand AI cost signals in the same way.
FinOps for AI isn't just marketing hype. The FinOps practice has already developed real resources that explain how cost practices need to change for AI work.
The big question remains: Will FinOps for AI become the practice that keeps innovation affordable, or will it turn into fear-based rules that slow us down before we find real answers?
DATA
Databricks Free Healthcheck Dashboard
Capital One Software has created a free dashboard with over 20 SQL queries to help teams check the health of their Databricks systems. The 20+ queries fall into four main groups:
General Usage Questions
Who uses the system the most?
Which clusters cost the most money?
How many clusters are running right now?
Cost Cutting Opportunities
The queries help you find jobs that use too much computer power, spot instances that could save you up to 90% on costs, and track which expenses are growing week by week.
Speed and Performance
When you fix these, everything runs faster.
Data Management
The final group shows you which datasets get used the most and how data flows through your system.
The dashboard comes ready to use with one click. You just paste the SQL queries into your Databricks system or download the complete dashboard.
FINOPS EVENTS
Mastering AI Economics
AI workloads are powerful but they’re also expensive. In this webinar, we’ll break down practical strategies to manage, develop, and reduce unit costs across your AI stack.
Learn how top engineering and FinOps teams are aligning performance with budget by optimizing architecture, tracking true cost per model, and using practical insights to stay ahead of runaway spend.
What we’ll cover:
Understanding unit costs: What they are, why they matter, and how to track them
Cost-efficient architecture: Design patterns and trade-offs that lower compute and storage bills
Data & model strategy: How to optimize what you train, when, and where
FinOps for AI: Building transparency and accountability into fast-moving AI teams
Speakers
Vaibhav Sharma
David Gross
Alon Savo
Host: Victor Garcia
September 9th - 6:00 PM CEST / 10AM EST
AWS
Maximize AWS Savings Plans at Payer vs Linked Levels
AWS Savings Plans can save you up to 72% on compute costs compared to regular on-demand prices. But if you manage multiple AWS accounts, you might notice something confusing. The savings recommendations look different depending on where you check them.
Two Different Views, Two Different Goals
Your main account looks at all your accounts together. It sees the big picture. Maybe your development teams work during the day while your batch processing runs at night. These different patterns create better opportunities to save money when combined.
Individual account recommendations are more careful. They only look at each account by itself. They can't see how unused savings from one account could help another account save money.
How AWS Makes These Recommendations
AWS calculates everything hour by hour, not daily or monthly. Your usage can change a lot within a single day, even if it looks steady when you zoom out.
For the main account view, AWS looks at all eligible spending across your organization. It finds the best way to save money by combining everyone's usage patterns.
For individual accounts, AWS only looks at each account separately. It calculates savings without considering how accounts could share benefits with each other.
Key Tips for Success
Pick one approach and stick with it. Mixing strategies without a clear plan creates confusion and wastes money.
Choose your lookback period carefully. If you just started new workloads, use a shorter period. For stable environments, longer periods work better.
Consider using an account with minimal workloads to buy savings plans. This ensures the savings go where they're needed most.
CLOUD PROVIDERS
AWS OneGov, Microsoft Wins Quadrant and GCP AI Inference Carbon Impact
AWS
The Billing and Cost Management console now shows a "Recommended actions" widget on your homepage, offering personalized cost-saving tips like shutting down idle resources or buying Savings Plans.
AWS Batch got an upgrade too - it automatically picks from more EC2 instance types, including older, cheaper options that fit your workloads better without manual tweaking.
The US GSA's new "OneGov" agreement with AWS could save agencies up to $1 billion through better discounts and streamlined procurement.
Microsoft Azure
Azure Kubernetes Service (AKS) detailed cost analysis is now live, giving teams clearer cost breakdowns by cluster and namespace. This helps accurately track and allocate container costs.
Microsoft also earned recognition as a Leader in Gartner's 2025 Magic Quadrant for Cloud-Native Application Platforms.
Google Cloud
Google Cloud launched carbon emissions data for AI inference workloads, letting teams track environmental impact alongside costs - perfect for emerging "GreenOps" practices.
AZURE
Finding Azure VM Disk Space Tricks
A customer asked what seemed like a simple question: "How do I find out how much actual used disk space our Azure VMs have?"
The author thought Azure Resource Graph would have all the answers. Here's the trick - Azure Resource Graph only shows how much disk space you bought, not how much you're actually using. Here’s how to solve this:
Option 1: KQL in Log Analytics
If you have VM Insights turned on, you can write queries to see real disk usage. The author shared code that calculates used space across all drives. This works great if all your VMs report to one Log Analytics workspace, or you can modify it to check multiple workspaces.
Option 2: PowerShell Scripts
Sometimes Log Analytics isn't set up everywhere. The author created a PowerShell script that can pull Windows servers from Active Directory or Azure, then check each one directly. It tests if each server is reachable, queries all local drives, and exports everything to a CSV file.
Option 3: Combine Methods
📺️ VIDEO
How to Make Your Engineers Take Action in FinOps with Larry Advey (FINOPS EXPERT)
Larry Advey, an experienced FinOps practitioner and maintainer of the FOCUS specification, shared game-changing insights on collaboration that challenge conventional wisdom about working with different personas in organizations.
Whether you're struggling to get engineering teams on board with cost optimization or wrestling with multi-cloud cost visibility, this deep dive into FinOps collaboration strategies offers practical solutions you can implement immediately.
🎖️ MENTION OF HONOUR
FinOps Teams Built As Platforms Are Just Better
Many companies struggle with FinOps because they treat it like a regular project instead of building it the right way. This creates problems that could be easily avoided.
The Engineering-Only Approach
When engineers run everything, they build great tools and dashboards. They can automate cost savings and create impressive reports.
But they miss the bigger picture. They don't connect their work to business goals or help finance teams understand what the savings actually mean. This leads to random cost cutting that might hurt important projects.
The Finance-Only Approach
When finance teams control FinOps, they're good at setting budgets and tracking spending. They know how to tie costs to business results.
But they can't actually fix the technical problems that cause waste. Engineers get frustrated because they receive cost targets without any context about why or how to meet them.
The Better Way: Platform Engineering Model
Smart companies treat FinOps like a platform that serves everyone. A central team builds tools and systems that both engineers and finance teams can use.
This approach works because it stops making FinOps an "us versus them" situation. Engineers get clear cost information built into their daily work. Finance gets accurate, predictable reports.
Professional Spotlight
Shannon Eldridge-Kuehn
FinOps Insights on (Digital) Paper
I just find amazing how Shannon writes about FinOps. The articles are so easy to digest, and really insightful. Someone to must watch in the ecosystem, and great collaborator. Looking forward to her participation in our Summit 🤞
That’s all for this week. See you next Sunday!
Join FinOps professionals at the FinOps Weekly Summit 2025 and discover how to:
Transform from reactive fire-fighting to strategic leadership — Learn the proven frameworks that top practitioners use to turn cloud cost management into a competitive advantage
Master AI-powered cost optimization — Get exclusive access to the latest automation tools and techniques that can reduce your cloud spend by up to 40% while accelerating innovation
Build bulletproof unit economics — Walk away with actionable playbooks for calculating true cloud ROI and proving business value to executives who control your budget
"But I don't have time for another webinar..."
This isn't another generic webinar. This is the only major FinOps event left in 2025, featuring battle-tested strategies from companies managing billions in cloud spend.
October 23rd, 2025 | 4:00 PM - 8:00 PM CEST
Limited seats available