$10k off
a multi-tenant
hosting platform.
An IT services company hosted dozens of customer applications across an equally large number of AWS accounts — one per client, all under a single payer organization. Optimization meant working across the entire fleet, not just one account.
Outcome
Before
$40,000
monthly AWS spend
After Koali
$30,000
monthly AWS spend
Annualized savings
Realized, ongoing
$120,000
$40k
Starting monthly spend
$30k
Ending monthly spend
15 days
End-to-end engagement
$120k
Annualized savings
the challenge
Optimization across a fleet, not a single account.
The client ran an IT services business hosting a wide range of customer applications on AWS. To keep tenant workloads isolated and billing clean, every customer lived in their own dedicated AWS account inside the parent organization. The result was an environment with dozens of accounts to analyze, each with its own EC2 fleet, storage footprint, and usage profile.
Their existing approach to commitment coverage hadn't kept up with the growth. Savings Plans and Reserved Instances were either missing, mis-sized, or scoped to individual accounts in ways that left compute on-demand across the broader fleet. EC2 instance sizing was inconsistent — some tenants ran on conservative-but-oversized x86 instance types that had been chosen years ago and never revisited.
The opportunity was clear: aggregate the optimization across the whole organization, lean into Compute Savings Plans at the payer level, right-size where customer workloads allowed it, and migrate compatible services to ARM-based Graviton instances for better price-performance.
the audit
Two days across dozens of accounts.
Read-only access at the organization level let us pull a consolidated view across every tenant account, identify the biggest aggregate opportunities, and prioritize the workloads where the math worked best.
Suboptimal commitment coverage
Existing Savings Plans and RIs were either undersized or scoped to individual accounts. A meaningful portion of compute hours were still being billed on-demand despite stable, predictable usage.
Conservative EC2 sizing
Many tenant workloads were running on instance types two sizes larger than needed — leftover decisions from when the workloads were first migrated and never revisited.
x86 workloads with Graviton-ready code
A subset of tenant applications were running on x86 instance families but had no real dependency on the architecture. They were prime candidates for Graviton migration with no application changes needed.
Account-level fragmentation
Cost optimization had historically been done per-account, missing the aggregation benefits available at the payer level. Treating the organization as a single fleet unlocked significantly more leverage.
the execution
One week to ship every change.
The plan was approved on day three. Execution rolled across tenant accounts in sequence over the following week, with each workload validated before moving to the next.
Days 1–2
Audit
Pulled a consolidated view across the organization, mapped commitment coverage gaps, and identified candidate workloads for right-sizing and Graviton migration.
Day 3
Plan & approval
Walked through a per-tenant change plan with the client. Each workload had its own risk profile, expected savings, and rollback path. Approval came back same day.
Days 4–10
Right-size, migrate, commit
Right-sized EC2 instances across approved tenant workloads. Migrated compatible applications to Graviton instance types. Restructured Compute Savings Plan coverage at the payer level to capture the aggregated baseline of the entire fleet.
Days 11–15
Validate & invoice
Confirmed realized savings against the baseline. Documented every change per tenant for the client's records. Final invoice followed only after the savings showed up on their bill.
the result
$10k a month, recurring.
The combination of right-sized fleets, Graviton migration where compatible, and properly-structured commitment coverage cut the client's monthly AWS spend by 25% — savings that compound month after month, with no impact on tenant workloads.
Outcome
Before
$40,000
monthly AWS spend
After Koali
$30,000
monthly AWS spend
Annualized savings
Realized, ongoing
$120,000
ready when you are
Got a fleet to optimize?
Multi-account organizations are where the math gets the most interesting. Free audit, read-only access, dollar-backed findings.
$10k+ monthly AWS spend · No risk