On September 17, 2026, Amazon Web Services announced the general availability of the Amazon EC2 T8i instance family. Powered by custom sixth-generation Intel Xeon Scalable Processors codenamed Granite Rapids, these low-cost burstable instances deliver up to 30 percent better price performance than previous-generation T3 configurations to target lightweight microservices, testing environments, and intermittent workloads.
As enterprise engineering budgets pivot aggressively toward heavy generative AI model training and inference pipelines, platform teams are forced to do more with less for the unglamorous backend services keeping the lights on. Idle compute overhead is no longer acceptable under strict FinOps governance frameworks. AWS designed the T8i architecture specifically to capture this long-tail market of spiky, low-to-moderate utilization tasks without punishing teams with over-provisioned hourly billing.
Granite Rapids Silicon and Burstable Credit Mechanics
Under the hood, the T8i lineup relies on custom silicon co-designed between Intel and AWS. By deploying Granite Rapids processors exclusively within the AWS fleet, the cloud giant has carved out a distinct cost advantage at the entry tier of its compute catalog. Burstable instances operate on a strict CPU credit accounting system. Each instance accumulates credits at a baseline rate, allowing containerized workloads or microservices to absorb sudden demand spikes by spending those reserves.
According to official specifications published by AWS, the family provides up to 70 percent higher compute performance, up to 1.25 times higher network bandwidth, and up to 2.4 times higher Amazon EBS bandwidth compared to legacy T3 instances. Unlimited credit mode ships enabled by default across the fleet. Engineering teams must monitor their credit burn rates carefully to prevent sudden throttling or unexpected on-demand charges during extended compute spikes.
Instance Specifications and Configuration Tiers
The T8i family introduces four distinct sizes, each packing two vCPUs configured as a single physical core. They maintain unique vCPU-to-memory ratios such as 1:0.25, 1:0.5, and 1:1 that remain absent in standard EC2 families.
| Instance Size | vCPUs | Memory (GiB) | Baseline Performance / vCPU (%) | CPU Credits Earned / Hour | Network Burst Bandwidth (Gbps) |
|---|---|---|---|---|---|
| t8i.nano | 2 | 0.5 | 5 | 3 | Up to 6.25 |
| t8i.micro | 2 | 1 | 10 | 6 | Up to 6.25 |
| t8i.small | 2 | 2 | 20 | 12 | Up to 6.25 |
| t8i.medium | 2 | 4 | 20 | 12 | Up to 6.25 |
For workloads requiring configurations larger than the medium tier, AWS recommends shifting to M8i Flex instances, which can scale seamlessly up to 16xlarge while maintaining competitive price-performance metrics against older hardware.
Availability and Migration Pathways
Existing T3 users can swap instance types to T8i without altering their underlying orchestration manifests or deployment scripts, retaining the exact same credit models and lightweight toolchains. The instances are available immediately via On-Demand and Spot purchasing options across a broad global footprint, including US East (N. Virginia, Ohio), US West (Oregon, N. California), multiple Asia Pacific hubs, Canada Central, and major European regions including Frankfurt, Ireland, London, and Paris. Furthermore, the micro and small configurations are included in the AWS Free Tier, lowering the barrier for developers spinning up new continuous integration and deployment pipelines or event-driven serverless architectures.
