01
Cost visibility
Tags, account and project structure, and FOCUS-format cost exports, so every line of the bill has an owner.
FIG 01 · Cloud Cost Optimization
Cloud cost optimization that keeps paying back
FinOps run by engineers. We change the infrastructure code, not just the spreadsheet, and leave your team with tags, budgets and alerts that keep working after we’ve gone.
See how it works
FIG 02 · Quick answer
What are cloud cost optimization services?
Cloud cost optimization services cut what you pay AWS, Azure or Google Cloud without slowing your product. Elgrith follows the FinOps Foundation’s Inform, Optimize and Operate phases: we tag and allocate spend, rightsize or shut down idle resources, plan Savings Plans or committed use discounts, and set budgets and alerts your team keeps.
- AWS, Azure and Google Cloud, reviewed with the same method
- 3 clouds
- Projects delivered since 2015
- 1,000+
- Clients across the USA, UK and India
- 250+
- Your billing accounts and savings plans
- 100%
FIG 03 · Why it matters
Cloud waste went up again. AI is part of why.
From Flexera’s 2026 State of the Cloud and the FinOps Foundation’s State of FinOps 2026, checked 24 Oct 2026. These are industry figures, not Elgrith results.
01
29%
Of cloud spend estimated as wasted
Flexera’s 2026 survey puts wasted cloud spend at 29%, an increase for the first time in five years, with AI workloads named as a driver.
Source: Flexera, 2026 State of the Cloud press release (18 Mar 2026)(opens in a new tab)02
85%
Say managing cloud spend is a top challenge
In the same survey, 85% named managing cloud spend as a top challenge, and 63% said they have established FinOps teams.
Source: Flexera, 2026 State of the Cloud press release (18 Mar 2026)(opens in a new tab)03
98%
Of FinOps teams now manage AI spend
State of FinOps 2026 (1,192 respondents) found 98% manage AI spend, up from 31% two years earlier.
Source: FinOps Foundation, State of FinOps 2026 (checked 24 Oct 2026)(opens in a new tab)
FIG 04 · Definition
What is cloud cost optimization?
Cloud cost optimization is the ongoing work of matching what you pay for cloud resources to what your product actually needs: seeing who spends what, removing waste, buying commitments wisely and keeping it that way. The FinOps Foundation frames it in three phases.
| Phase | What happens | What we deliver | What it needs from you |
|---|---|---|---|
| Inform | Make spend visible and allocate it to teams, products and customers | Tagging policy, account structure, FOCUS-format cost exports, showback reports | Read access to billing and an owner per product |
| Optimize | Remove waste and buy commitments for steady usage | Rightsizing, idle clean-up, storage tiers, Savings Plans or CUD plan, all as code changes | Engineers to review and approve changes |
| Operate | Keep it that way as the product grows | Budgets, anomaly alerts, policies in infrastructure as code, monthly review | A named person who owns the review |
The phases repeat: every new service, team or AI feature starts the loop again.
Moving workloads first? Read the cloud migration checklistFIG 05 · What you get
What our cloud cost optimization services cover
From the first bill review to guardrails in your pipelines. Every change we recommend comes with the code to make it.
02
Showback and chargeback
Shared costs split fairly across teams, products or customers, with reports people can read.
03
Rightsizing
Instances, databases and containers resized from real utilization, with performance checks before and after.
04
Idle and orphaned resources
Unattached disks, old snapshots, idle load balancers and forgotten test environments found and removed safely.
05
Storage tiers
Lifecycle rules that move cold data to cheaper tiers and delete what nobody needs to keep.
06
Data transfer
Cross-region, cross-zone and egress traffic mapped and reduced through architecture changes, caching and CDNs.
07
Commitment planning
Savings Plans, Reserved Instances, Azure reservations or Google committed use discounts, sized to your steady baseline.
08
Kubernetes allocation
Requests and limits, node pools and spot capacity tuned, with costs split by namespace and team.
09
AI and GPU spend
GPU scheduling, model choice and token budgets for model APIs, tracked per feature.
10
Anomaly alerts and unit economics
Alerts when spend jumps, and cost per customer, order or request so growth and spend can be compared.
Deliverables
11
deliverables, in writing
Included in every build, written into your fixed quote before work starts.
FIG 06 · Levers
AWS, Azure and Google Cloud levers
Each cloud has its own discounts and recommendation tools. We use them, check their limits and turn the results into code changes. Vendor figures are the vendors’ own wording, not Elgrith savings claims.
- AWS
- Azure
- Google Cloud
Tagged, allocated spend
- Budgets
- Alerts
- Unit costs
AWS cost optimization
- Lever
- Savings Plans: AWS says you “can save up to 72%” on compute, on 1- or 3-year terms
- We add
- Commitments sized to your steady baseline, plus rightsizing from Compute Optimizer
Azure cost optimization
- Lever
- Azure reservations and savings plans, and Advisor rightsizing and shutdown recommendations
- We add
- Advisor findings checked against your existing reservations before anything is resized
Google Cloud cost optimization
- Lever
- Committed use discounts: spend-based on 1- or 3-year terms; resource-based for Compute Engine only
- We add
- The right CUD type per service, with BigQuery and other products planned separately
Kubernetes on any cloud
- Lever
- Requests and limits, cluster autoscaling, spot node pools and OpenCost allocation
- We add
- Cost per namespace and team, and limits that stop one service starving the rest
Data platforms
- Lever
- Warehouse auto-suspend, sizing and query tags on Snowflake, Databricks or BigQuery
- We add
- Guardrails from our data warehouse consulting team
Hosting and deployment
- Lever
- Managed hosting, CDNs and right-sized environments for web apps
- We add
- Staging environments that switch off out of hours
FIG 07 · FinOps
FinOps consulting, run by engineers
A FinOps consultant connects finance, engineering and product around one view of cloud spend. We follow the FinOps Foundation’s framework and its open billing standard, FOCUS.
04principles
FOCUS 1.4 cost data
FOCUS, the FinOps Open Cost and Usage Specification, gives every cloud’s bill the same columns. Version 1.4 was ratified on 4 Jun 2026 and adds Invoice Detail and Billing Period datasets (FinOps Foundation, 10 Jun 2026).
Shared ownership
Engineers see the cost of what they ship, finance sees why it changed, and product sees cost per customer.
Unit economics
Cost per customer, order, request or model call, so a rising bill can be read against growth.
FinOps for AI
Model API and GPU spend tagged per feature, with budgets and limits, because AI spend is now part of almost every FinOps practice.
FIG 08 · Tools
Native cloud cost optimization tools and where they stop
The clouds’ own tools are free or cheap and worth using. Their defaults have limits you should know before acting on them. Checked 24 Oct 2026.
04principles
Short lookback windows
AWS Compute Optimizer analyzes 14 days by default, up to 93 days with its paid enhanced infrastructure metrics. Azure Advisor looks at 7 days by default, configurable up to 90.
Estimates that ignore commitments
Azure Advisor’s savings estimates use retail rates and don’t account for Reserved Instances or savings plans you already own, so listed savings can be higher than possible (Microsoft Learn, 18 Feb 2026).
Narrow scope
Google’s resource-based committed use discounts apply only to Compute Engine (Google Cloud docs, 21 Oct 2026); other services need spend-based commitments.
Recommendations aren’t changes
A tool can say an instance is too big. Someone still has to change the code, test it and ship it. That is the part we do.
FIG 09 · Workloads
Kubernetes and AI workload costs
Containers and AI are where bills hide: shared clusters blur ownership and GPUs are expensive to leave idle.
05principles
Requests and limits
Set from measured usage, so pods don’t reserve capacity they never use.
Node pools and spot
Separate pools for steady and bursty work, with spot capacity for jobs that can be interrupted.
Shared cost allocation
OpenCost or the cloud’s own allocation to split cluster costs by namespace, label and team.
GPU scheduling
GPUs shared or scheduled for training and inference, and released when jobs finish.
Model API spend
Model choice, caching and token budgets per feature, so an AI feature’s cost is known before it scales.
FIG 10 · Operate
Keeping costs down after we leave
Savings that depend on a consultant checking every month don’t last. We build the habits into your tools.
05principles
Budgets and alerts
Budgets per account, team and environment, with alerts routed to the people who can act.
Guardrails in infrastructure as code
Terraform modules and policies that require tags, block oversized instances and set lifecycle rules by default.
Monthly review
A short, repeatable review of anomalies, commitments and unit costs that your team runs.
Cost checks in CI/CD
Cost estimates on infrastructure pull requests, set up with our DevOps consulting team.
DevOps consultingBuilt by the team behind our flagship
Kenerate AI, our flagship build, designed and engineered by Elgrith. It is an AI image and video studio, not a cloud cost optimization project.
See Kenerate AI
FIG 11 · Measurement
What we measure during cloud cost optimization
Savings you can trace to the bill, and spend you can explain. That is what we report.
| How we measure it | What we won’t do | |
|---|---|---|
| Allocation | Share of spend tagged and assigned to an owner | Report savings on spend nobody can attribute |
| Waste removed | Idle and oversized resources removed, with the monthly run-rate change from your own bill | Claim a percentage before we have seen your bill |
| Commitments | Coverage and utilization of Savings Plans, reservations or CUDs | Lock you into commitments beyond your steady baseline |
| Performance | Latency and error rates before and after each change | Cut cost by slowing your product |
| Unit costs | Cost per customer, order or request, month on month | Leave reports only we can update |
Savings are confirmed after the assessment, against your own billing data. We agree the measures and reporting rhythm in your written scope.
FIG 12 · Process · 6 steps
See the spend, then change the code
Free call, bill review, fixed written quote, then changes shipped as code in weekly demos, with performance checked before and after.
- Step 01
Free call
Tell us which clouds you use, what worries you about the bill and what you have already tried. We ask about growth plans and commitments first.
- Step 02
Discovery
A cost assessment with read-only billing access: allocation gaps, waste, commitment coverage, Kubernetes and AI spend, ranked by effort and impact.
- Step 03
Fixed written quote
A written scope and a fixed quote, with the timeline, access and hand-over agreed before any work starts. You own the code and IP, and your billing accounts and savings plans stay in your name.
- Step 04
Sprints with weekly demos
Changes shipped as infrastructure code, highest impact first, each shown in a weekly demo with the before-and-after numbers from your bill.
- Step 05
Launch
Budgets, alerts, policies and dashboards switched on, with a hand-over session for engineering and finance.
- Step 06
Support
Monthly reviews, commitment planning or a full hand-over to your team, as agreed in your written scope.
FIG 13 · What you get
Everything included
Written into your fixed quote, so you know exactly what you are paying for before work starts.
11items
- Written scope and fixed quote
- Cost assessment and ranked opportunity list
- Tagging policy and account structure
- FOCUS-format cost exports and showback reports
- Rightsizing and idle clean-up as code changes
- Commitment plan: Savings Plans, reservations or CUDs
- Kubernetes and AI cost allocation
- Budgets, anomaly alerts and policies in Terraform
- Unit-economics dashboard
- Monthly review runbook
- Your billing accounts and savings plans stay in your name
FIG 14 · Tech & tools
The tools behind the work
Proven, well-supported tools, chosen for your project and set up in accounts your business owns.
- AWS Cost Explorer
- AWS Compute Optimizer
- Azure Cost Management
- Azure Advisor
- Google Cloud Billing
- FOCUS
- Terraform
- Kubernetes
- OpenCost
- Karpenter
- Prometheus
- Grafana
- Infracost
- BigQuery
- Python
FIG 15 · Who it’s for
Cloud cost optimization for teams whose bill outgrew the plan.
Same method everywhere: allocate first, change code second, keep the guardrails.
4 audiences
SaaS teams
Gross margin that survives growth.
Cost per customer and per feature, so pricing, architecture and AI features are decided with the real numbers.
- Unit economics per customer
- AI and model API spend
- Kubernetes allocation
- Guardrails in CI/CD
Mid-market
A bill finance can read.
Spend allocated to departments and projects, commitments planned with finance, and monthly reports in plain English.
- Showback by department
- Commitment planning
- Budget alerts
- Monthly review
After migration
Lifted and shifted, now too expensive.
Workloads moved as-is often run oversized. We rightsize and re-architect the costly parts after the move.
- Rightsizing from real usage
- Storage tiers
- Data transfer review
- Managed services where they pay
UK teams
Cloud cost optimisation for UK organisations.
UK engagements run through our London client-partnership centre, with delivery from our New Delhi HQ. Reports use UK English and show spend in the currency you are billed in, and working hours are agreed in your written scope.
- London client partner
- UK English reports
- Fixed written quote in GBP
- Your accounts stay yours
FIG 16 · Working with us from the USA and UK
Built in New Delhi. Run on your hours.
Every project gets a point of contact in New York or London and an engineering team in New Delhi. Here is how today’s working days line up, live.
New York
North America
--:--ET
New Delhi team Your working day- Shared working time
- ≈ 3 hours today
- New Delhi is
- 10.5 h ahead
- Live calls, your time
- 9 a.m. to 12 p.m. ET
London
UK & Europe
--:--UK
New Delhi team Your working day- Shared working time
- ≈ 4.5 hours today
- New Delhi is
- 5.5 h ahead
- Live calls, your time
- 9 a.m. to 1:30 p.m. UK
Three ways to work together
- 01
Fixed-scope project
A written scope, one fixed price and a dated plan. Suits websites, apps and MVPs.
- 02
Dedicated team
Designers and engineers on your roadmap every month, billed monthly.
- 03
White label
For US and UK agencies: our work, your brand, unbranded deliverables.
Every engagement
- Fixed quote in USD, GBP or INR after a free 30-minute call
- NDA before discovery. You own all code, designs and content
- A working demo every sprint, not status slides
- Reply within one business day, calls inside your hours
Book a call
Calls on ET or UK hours
FIG 17 · Commitments
Savings Plans vs Reserved Instances vs Spot
Savings Plans, Reserved Instances and spot capacity solve different problems. The vendors describe the mechanics; we size them to your baseline.
| Savings Plans | Reserved Instances | Spot | |
|---|---|---|---|
| What you commit to | A spend per hour for 1 or 3 years | Specific instance capacity for 1 or 3 years | Nothing: spare capacity that can be reclaimed |
| Flexibility | High: Compute Savings Plans cover instance family, size, Region, Fargate and Lambda | Lower: tied to instance attributes | Highest, but workloads can be interrupted |
| Best for | Steady compute that changes shape over time | Very stable databases and fixed fleets | Batch jobs, CI runners and stateless workers |
| Main risk | Committing above your steady baseline | Paying for capacity you stop using | Interruptions without retries |
FIG 18 · Rules and dates
Cloud cost and FinOps facts that affect your plan (checked 24 Oct 2026)
Standards, survey figures and tool defaults that shape a FinOps plan, with what changed and when. Checked 24 Oct 2026.
| What changed | When | Source | What it means for you |
|---|---|---|---|
| FOCUS 1.4 ratified; adds Invoice Detail and Billing Period datasets | 4 Jun 2026 | FinOps Foundation, 10 Jun 2026 | Invoice-level reconciliation in one format across clouds |
| Wasted cloud spend estimated at 29%, the first rise in five years | 18 Mar 2026 | Flexera 2026 State of the Cloud | Waste is growing again, partly from AI workloads |
| 98% of FinOps practitioners manage AI spend, up from 31% two years earlier | 2026 survey | FinOps Foundation, State of FinOps 2026 | AI spend belongs in the same review as compute |
| Savings Plans: “save up to 72%” on compute, 1- or 3-year terms | In effect (checked 24 Oct 2026) | AWS Savings Plans user guide | Size commitments to your steady baseline only |
| Compute Optimizer: 14-day default lookback, up to 93 days with paid enhanced metrics | In effect (checked 24 Oct 2026) | AWS Compute Optimizer user guide | Seasonal workloads need the longer window |
| Azure Advisor: 7-day default lookback; estimates ignore existing RIs and savings plans | 18 Feb 2026 | Microsoft Learn, Advisor cost recommendations | Check reservations before acting on a resize |
| Google resource-based CUDs apply only to Compute Engine | Updated 21 Oct 2026 | Google Cloud docs, committed use discounts | Other services need spend-based commitments |
Sources: FinOps Foundation, Flexera, AWS, Microsoft Learn and Google Cloud documentation, all checked 24 Oct 2026. Next check: January 2027.
FIG 19 · Compare
Engineer-led FinOps vs a cost tool alone vs a one-off spreadsheet review
| Engineer-led FinOps | Cost tool alone | One-off spreadsheet review | |
|---|---|---|---|
| Finds waste | Yes, with context from your architecture | Yes, from billing and metrics | Partly, at one point in time |
| Makes the change | Yes: code changes, tested and shipped | No: someone still has to act | No |
| Keeps costs down | Budgets, alerts and policies in your code | Alerts, if someone watches them | Savings drift back within months |
| Best for | Teams that want savings that last | Teams with engineers free to act | A quick first look |
FAQ
Frequently asked questions
Straight answers about Cloud Cost Optimization with Elgrith.
Cloud cost optimization is matching what you pay AWS, Azure or Google Cloud to what your product needs: allocating spend to owners, removing idle and oversized resources, buying commitments for steady usage, and keeping it that way with budgets, alerts and policies.
Free 30-minute call
Send us the bill that worries you. We’ll show where it goes.
Share which clouds you use and what you spend on roughly. After a free call you get a written assessment scope and a fixed quote.
Pick a day for your call
OpenPick a day. You'll choose a time in your own time zone on the next step.