👉 Brought to you by:- OpenOps - the open source FinOps & CloudOps automation platform. Automate common cloud workflows with human-in-the-loop approvals and first-class AI support.- WorkOS: Modern APIs for auth, identity, and enterprise features like SSO, SCIM, RBAC, and more. FinOpsRelated terms: Cloud Cost Management, Cloud Cost Optimization. ✅ Who is this for?
- Engineering leaders (CTOs, VPs of Engineering) dealing with unpredictable cloud bills
- Platform engineers and DevOps teams who need to optimize cloud spending while maintaining performance
- Finance and procurement professionals working with engineering teams on cloud budgets TL;DR:- Problem: As cloud costs grow, they become hard to track, attribute and act upon across teams.
- Solution: FinOps platforms provide cost visibility, optimization recommendations, and budget controls through tagging and automation.
- In Sum: FinOps is the tooling and automation layer that makes cloud costs visible and controllable, more broadly, a discipline and framework promoted by the FinOps Foundation (a Linux Foundation program).
How does it work? 💡The FinOps Foundation has a framework with phases and maturity models, and recommends a FinOps Practitioner to run things - but the platforms do the heavy lifting. Imagine this, you're scaling infrastructure, launching features, and every deployment adds costs you're not tracking. FinOps platforms catch this through tagging: tag resources with team, project, environment, and these tags feed dashboards showing who's spending what. Platform team spins up a GPU instance? Shows up under their budget immediately. New services, databases, environments - all tracked as they happen instead of adding to your bill invisibly. Platforms give you visibility - "this API call costs $0.003" or "this namespace burned $2K yesterday." They spot waste automatically: idle environments, oversized databases, unattached storage. Alert: "this database has been at 5% CPU for 30 days - downsize and save $1500/month." They also fix things like: - Budget alerts: Get notified at 80% spend before you overshoot
- Anomaly detection: Cluster suddenly 3x more expensive? Alert within hours
- Commitment discounts: Buy Reserved Instances or Savings Plans for 40-75% savings
- Automated recommendations: "Delete these 47 volumes ($1,200/month)" or "Buy RIs for these 12 instances (save $28K/year)."
One caveat: recommendations only create value if teams follow through. FinOps compounds when optimization is continuous, not a one-time cleanup. Three things every org must get right: (1) Tagging discipline - you can't attribute costs you can't identify, (2) Visibility - dashboards that show who's spending what, (3) Actionability - recommendations that teams actually follow through on. Questions ❔- FinOps vs procurement or finance?
- Procurement negotiates vendor contracts before you buy. Finance tracks expenses after you spend. FinOps is an engineering-adjacent practice - teams own their cloud costs in real-time and optimize for value, not just cost reduction.
- FinOps vs cost cutting?
- FinOps is about maximizing value, not just cutting spending. If faster CI/CD infrastructure costs $10K/month but ships features 2 days faster and drives $100K in revenue, that's a good investment.
- Reserved Instances vs Savings Plans?
Why? 🤔- Massive cost savings: Organizations typically save 20-60% of cloud costs - Lyft cut costs per ride by 40% in six months, a European bank saved $406K in 6 months. Companies consistently find millions in waste: idle VMs, over-provisioned databases, unattached storage volumes, and so much more.
- Automated optimization at scale: Platforms scan your infrastructure 24/7 finding waste you'd never catch manually - unattached volumes, idle load balancers, oversized databases, orphaned snapshots and more.
- Granular cost attribution: Tag-based allocation down to teams, services, and API calls. See exactly what the Platform team spends vs what the Frontend team spends, or what "the user authentication service" costs vs "the recommendation engine" - which helps you make data-driven decisions about where to optimize.
- Multi-cloud management: Unified visibility across AWS, Azure, GCP, OCI, Alibaba Cloud, and others. Standardized metrics via FOCUS enable cross-cloud cost comparison, so you can pick the best cloud for each workload instead of staying locked in by billing confusion.
Why not? 🙅- Why can’t the CTO do this? Seriously, do you need an entire framework, FinOps Practitioner certifications, and expensive platforms when a finance analyst with a spreadsheet could handle this? For smaller companies, the CTO could just spend 30 minutes reviewing the cloud bill each month instead of building out a whole FinOps practice.
- Premature optimization: Your cloud bill is $5K/month and you're implementing FinOps? You're wasting engineering time that should be spent shipping features that make money. Below $1M/year cloud spend, you're bikeshedding cost optimization when you should be focused on growth.
- One-time fix, recurring fee: Many FinOps platforms charge 1-3% of cloud spend annually (pricing models vary) - that's $300K/year on a $10M bill. Once you've cleaned up idle resources and bought Reserved Instances, what are you paying for? To be fair: large orgs with dynamic infrastructure do need ongoing optimization. New teams spin up resources constantly, workloads shift. But you can read this either way - necessary practice or over-engineered justification for recurring fees.
- Tool sprawl and integration complexity: You already have native cloud cost tools. Now add third-party platforms for Kubernetes, multi-cloud visibility, and infrastructure-as-code cost estimates, and suddenly you're managing 4+ cost tools with overlapping data and conflicting recommendations. Which platform's "savings suggestion" do you trust? Getting them to agree on a single source of truth is genuinely hard. (P.S. check out OpenOps in that regard - a Zapier-like platform for FinOps automation - (disclaimer -> and sponsor!)).
- Data quality dependency: Traditional FinOps platforms rely heavily on tagging discipline - if 40% of your resources aren't tagged properly, cost attribution suffers. Modern platforms can allocate spend without modifying tags, and K8s tools use namespace/label-based allocation. Still, most enterprise platforms assume solid tagging governance, and retrofitting tags across existing infrastructure is painful.
- Trend vs need: Are you doing FinOps because your cloud bill is actually out of control, or because it's the hot topic? The FinOps market is projected to grow from $9.79B to $20.87B, but you should only do FinOps if you feel like your cloud bill is actually out of control.
- CloudHealth by Broadcom - Multi-cloud management with policy-driven governance (originally VMware, now Broadcom).
- Apptio Cloudability - Enterprise FinOps platform acquired by IBM for $4.6B in 2023.
- Flexera - Enterprise IT management and FinOps platform that acquired Spot by NetApp in 2025.
- Kubecost - Kubernetes-specific cost management acquired by IBM in Sept 2024.
- Vantage - Multi-cloud cost visibility and optimization.
- Ternary - Multi-cloud with strong Google Cloud support. Offers self-hosted option.
- CloudZero - SaaS-focused with unit economics and 100% cost allocation.
- PointFive - CEPM platform focused on waste detection over cost tracking.
- Finout - Multi-cloud FinOps platform with unit cost tracking.
- ProsperOps - Automated commitment management and rate optimization.
- CloudFix - Automated AWS optimization and cost reduction.
- AWS Cost Management - Native AWS tools (Cost Explorer, Budgets, Savings Plans). Free but AWS-only.
- Archera - Cloud commitment management and optimization.
- Open source: OpenOps (no-code automation, Apache 2.0), OpenCost (CNCF project), Komiser
- More: nOps, Anodot, Infracost (Terraform cost estimates), Vega Cloud.
🤠 My opinion:For most teams, start with cloud-native tools (they're free) to get visibility, then add a third-party platform once you're past $1M/year cloud spend. If you're heavy on Kubernetes, invest in Kubecost or OpenCost early - container cost allocation is genuinely hard without specialized tooling. One non-negotiable: implement a proper tagging strategy from day one - it's painful to retrofit and your cost attribution will be garbage without it Forecast 🧞- AI x FinOps: GPU costs are 10-100x more expensive than traditional compute, making AI workloads training and inference bills a huge concern. Key challenges: workload placement (GPU-aware scheduling, Multi-Instance GPU partitioning), scheduling efficiency (time-slicing, spot instances for training), and model lifecycle governance (intelligent routing to cheaper models for simpler tasks, token optimization). Expect dedicated "AI FinOps" and LLMOps specialization by 2026 - platforms like Apptio are already adding AI-specific features.
- Market consolidation continues: IBM's $4.6B Apptio acquisition and Kubecost purchase signal ongoing consolidation. The $9.79B market growing to $20.87B by 2030 (13.5% CAGR) will attract more serious players and acquisitions.
- Autonomous optimization becomes standard: We're shifting from recommendations to automated actions with guardrails. Tools are enabling autonomous optimization with human-in-the-loop approvals.
- FinOps extends beyond cloud to ITAM: The FinOps Foundation partnered with ITAM Forum in 2025, expanding from cloud to IT Asset Management - SaaS, software licenses, hybrid infrastructure. "FinOps" is evolving from "cloud cost management" to managing spend across all IT assets.
- Carbon metrics become as important as cost: Sustainability reporting requirements will drive "GreenOps" integration, with platforms like Greenpixie leading the charge. By 2026, expect FinOps platforms to track carbon footprint alongside costs as standard.
- FinOps becomes a distinct career path: The FinOps Practitioner role is already established with certification programs (Practitioner and Professional levels). The growing industry means companies are scrambling to hire, with more certified practitioners expected.
- (Startup idea) I believe a company will emerge for SMBs solving the surprise bill problem: Startups don't need fancy FinOps dashboards and cost attribution - they need hard spending limits that actually stop resources when you hit $5K/month. The dreaded surprise bill kills startups - $2.3M from a misconfigured S3 bucket, $1M from auto-scaling gone wrong, crypto-mining malware racking up six-figure bills. I talked to a VC in Israel about this problem 8 years ago and believe it even more now. Some providers are adding budget enforcement here and there, but AWS won't do it - there's no incentive to let customers cap their spend. A company that sits between your AWS account and actually shuts down resources at hard limits could save startups from going under - I'd love to invest if you're building this.
- Edge platforms won't disrupt the hyperscalers: Vercel and Cloudflare are streamlined and developer-friendly, but Cloudflare does ~$1.67B/year while AWS does ~$100B/year. There's a reason you see BlackDuck and JFrog Artifactory more than Vercel or Cloudflare Pages in many companies - some platforms are built with the enterprise in mind first, others with bottom-up growth. Both are valid approaches, but big companies tend to use boring technology.
Who uses it? 🎡For more case studies, check the FinOps Foundation YouTube - hundreds of talks from FinOps X conference. Additional resources for diving deeper: Thanks 🙏I wanted to thank @TomGranot, the best marketer I know, who edits every issue since I started this newsletter. Full disclosure - he works for OpenOps, which is mentioned in and has sponsored this issue. Thanks also to Andrew Mallaband — Founder of Breakthrough Moments. With over three decades helping technology companies bridge engineering, sales, and marketing to unlock growth, Andrew works with organisations across North America and EMEA to refine messaging, accelerate revenue, andscale sustainably. He also regularly publishes analyst-grade content across multiple sectors, including Observability, FinOps, Middleware Messaging & Streaming, Cyber Security, and Digital Adoption. EOF(Where I tend to share unrelated things). Sorry for not posting for so long! I'm back at it. Would love to hear about any trends you'd like me to cover next, just reply back here.
|
