OCI flexible shapes, such as the E5 flex and Ampere A1 flex shapes, let you choose the number of OCPUs and the amount of memory per instance independently, in fine increments, rather than picking from a fixed menu of instance types. That independence is the quiet advantage: on the hyperscalers a workload that needs more memory usually forces you up to the next instance size and pays for cores you will never use, while on OCI you add only the memory. Ampere A1 arm cores often carry a materially lower price per core than x86, and a portion of A1 capacity has been available at no cost under the always free tier. Sized to real utilization, flexible shapes routinely remove a quarter or more of compute spend that fixed sizing leaves stranded.
Here is the mechanism, the price math, and how to right size flexible shapes safely.
What is a flexible shape on OCI?
A flexible shape lets you specify OCPUs and memory as separate dials when you create or resize a compute instance or an OKE node, within a supported range, instead of accepting the fixed core to memory ratio of a named instance type. The E series flexible shapes cover general x86 workloads, while the Ampere A1 flexible shape runs on arm cores. One OCPU on OCI maps to a full physical core, which is a different unit than a hyperscaler vCPU, so when you benchmark price, compare on cores rather than vCPUs or the comparison misleads you.
Where do the savings actually come from?
Three sources. First, no rounding waste: you provision the exact memory a workload needs without buying extra cores, and the exact cores without buying extra memory. Second, Ampere A1 arm pricing: A1 cores commonly price well below comparable x86 cores, so memory bound and scale out services that run cleanly on arm move at a lower unit rate. Third, precise scale to demand: because the increment is one OCPU and one GB of memory, autoscaling and scheduled resizing track demand closely rather than in coarse instance class jumps.
| Dimension | Fixed instance sizing | OCI flexible shape |
|---|---|---|
| Core to memory | Fixed ratio per type | Chosen independently |
| Memory bound app | Forces a larger instance, pays for idle cores | Add memory only |
| Arm option | Limited | Ampere A1 at a lower price per core |
| Resize granularity | Whole instance class jumps | One OCPU and one GB steps |
Rates are indicative and depend on region, shape, and term; verify against current OCI pricing.
How do you size a flexible shape without starving production?
Look at sustained CPU and memory utilization over a representative two to four week window, including peaks and batch windows. Size OCPUs to the busy hour CPU plus headroom, and memory to the working set plus cache and burst, not to the long run average. For arm, validate the workload and its dependencies run cleanly on Ampere before you move it. Set autoscaling on the OCPU dial where the platform supports it, and review the shape quarterly so it tracks the workload as it changes. The goal is high utilization with enough headroom to absorb a normal peak, not the smallest possible shape.
A worked example
A European SaaS company ran a memory heavy analytics tier on a fixed x86 instance type, paying for sixteen cores to get the memory it needed while CPU sat near 30 percent. Moving to an OCI E5 flexible shape sized to eight OCPUs and the same memory removed the idle cores, and shifting two stateless services to Ampere A1 flexible shapes cut their compute rate further. Compute on the tier fell by roughly a quarter with no change to latency. Figures are verified against billing data and anonymised.
Frequently asked questions
What are OCI flexible shapes?
Are Ampere A1 shapes cheaper than x86 on OCI?
How do I size a flexible shape?
Size your OCI compute to real demand
We help enterprises move to right sized OCI flexible shapes and Ampere A1 where it fits, as an independent advisory that takes zero provider commissions and answers only to you. Our guarantee: we reduce your cloud spend or we reimburse our service fee, on a Fixed Fee or a no risk Gainshare basis. Download the OCI Universal Credits guide, read the deeper OCI cost optimization guide, and see the OCI architecture choices that cut cost.
Put a defensible number on your cloud spend.
No provider in the room, no published price list. Tell us your footprint and we will scope the savings against your billing data — we reduce your cloud spend or we reimburse our service fee.
The Cloud Spend Navigator: what changed in cloud pricing, commitments, and FinOps — no vendor spin.