Amazon Web Services (AWS) has officially announced the general availability of its new low-cost burstable Amazon Elastic Compute Cloud (EC2) T8i instances. Powered by custom sixth-generation Intel Xeon Scalable Processors, known by their codename Granite Rapids and available exclusively on AWS, these new instances are designed to deliver up to 30 percent better price-performance over the previous generation T3 instances. Aimed at a broad spectrum of low-to-moderate CPU utilization workloads, the T8i lineup provides developers, startups, and enterprise engineers with an economical compute option for a wide array of lightweight tasks, ranging from microservices architectures and low-traffic websites to development and testing environments, data processing jobs, and freemium services.
For years, thousands of AWS customers have relied on T-family instances to run lightweight workloads that require small, cost-effective compute configurations. These environments often include small databases, short-duration compute tasks, and event-driven applications. A primary driver of the T-family’s enduring popularity is its burstable performance model. This system provides applications with a reliable baseline level of CPU performance while granting the flexibility to burst significantly above that baseline whenever unexpected spikes in traffic or processing demands occur, utilizing a system of accrued CPU credits.
As modern software development practices continue to evolve, enterprise infrastructure is undergoing rapid modernization. Organizations are increasingly migrating their legacy operations from on-premises data centers to the cloud, adopting complex event-driven and microservices architectures, and experimenting with emerging technologies such as artificial intelligence inference workloads. Amidst these architectural shifts, AWS customers have consistently expressed a demand for newer generation, cost-optimized small instances. Specifically, they have requested better price performance to continuously reduce their total cost of ownership, alongside a seamless migration path that allows them to leverage their existing technical knowledge, operational tooling, and automation scripts without requiring massive re-architecting.
The newly introduced T8i instances are engineered to address each of these customer requirements directly. By combining the raw compute capability of custom sixth-generation Intel Xeon Scalable Processors with the well-established burstable credit mechanism, AWS provides an environment where lightweight applications can scale their CPU usage transparently.
In terms of physical and logical specifications, the T8i instance family is structured to offer four distinct sizes designed to fit precise workload footprints. These sizes include the t8i.nano, t8i.micro, t8i.small, and t8i.medium configurations. Each of these four instance sizes is equipped with two vCPUs offered as a single physical core, ensuring predictable threading behavior for lightweight applications.
The individual specifications vary primarily by memory allocation, baseline performance percentage, and hourly credit generation. The t8i.nano instance provides 0.5 GiB of memory, a baseline performance of 5 percent per vCPU, and earns 3 credits per hour, coupled with a network burst bandwidth of up to 6.25 Gbps. Stepping up to the t8i.micro configuration yields 1 GiB of memory, a 10 percent baseline performance per vCPU, and 6 credits earned per hour. The t8i.small instance doubles the memory again to 2 GiB, offering a 20 percent baseline performance per vCPU and 12 credits earned per hour. Finally, the t8i.medium instance matches the small configuration’s 2 GiB baseline performance and 12 credits per hour, but expands the available memory to 4 GiB. Network burst bandwidth across all four variants remains robust, scaling up to 6.25 Gbps.
Much like their predecessors in the T3 family, the T8i instances offer unique vCPU-to-memory ratios—such as 1:0.25, 1:0.5, and 1:1—that are notably absent from standard general-purpose EC2 instance categories. This specialized allocation makes them uniquely suited for memory-conscious lightweight applications. Furthermore, T8i instances retain the familiar CPU credit system, supporting both the Standard and Unlimited credit configuration modes. In Standard mode, an instance accumulates CPU credits when running below its baseline and spends them when bursting. In Unlimited mode, which is configured as the default setting on all T8i instances, an instance can burst beyond its baseline even when its credit balance is fully depleted, ensuring that critical microservices or login gateways do not experience performance throttling during unexpected traffic surges.
For workloads that outgrow the small footprint of the T8i lineup—moving beyond the nano, micro, small, and medium sizes—AWS recommends transitioning to M8i Flex instances. These alternative offerings deliver comparable price-performance improvements over previous-generation equivalents while providing the architectural flexibility to scale up to massive 16xlarge configurations as application requirements grow.
The rollout of Amazon EC2 T8i instances begins immediately, with availability spanning multiple major AWS regions across the globe. Customers in the United States can access the new instances in the US East regions covering Northern Virginia and Ohio, as well as the US West regions in Oregon and Northern California. Across the Asia Pacific region, availability includes Hyderabad, Malaysia, Mumbai, Seoul, Singapore, Sydney, and Tokyo. In Canada, the instances are live in the Canada Central region, while European customers can deploy them in Frankfurt, Ireland, London, and Paris. Organizations seeking specific regional availability details or looking ahead to upcoming geographic expansions can check the CloudFormation resources tab within the AWS Capabilities by Region portal.
Purchasing options for T8i instances are flexible, with availability spanning On-Demand instances and Spot instances, while Savings Plan options are slated to arrive in the near future. These instances are built to support shared tenancy exclusively, meaning they do not support Dedicated tenancy or Dedicated Hosts. For developers and organizations looking to test the new infrastructure without immediate financial commitment, the t8i.micro and t8i.small instances are also included under the AWS Free Tier. Detailed pricing information can be found directly on the official Amazon EC2 Pricing page.
Developers and systems administrators can begin provisioning and testing T8i instances today through the Amazon EC2 console. AWS has encouraged users to share their feedback, deployment experiences, and performance observations through AWS re:Post for EC2 or by reaching out through their standard AWS Support communication channels.