A fast CPU combined with a large amount of memory results in increased performance for applications such as electronic design automation, and certain relational databases. This can help customers speed time to market for product development.

Reducing the number of CPUs required for certain workloads reduces the licensing fees for applications that use a per-CPU pricing model. Software licensing costs for EDA applications amount to a large share of total costs and reducing the number of CPUs required can result in significant cost savings.

z1d instances support up to 25 Gbps of network bandwidth when launched within an EC2 placement group. For EBS volumes, z1d instances support a Non-Volatile Memory Express (NVMe) interface resulting in up to 14 Gbps of dedicated EBS bandwidth to both encrypted and unencrypted EBS volumes.

z1d high frequency instances deliver a sustained all core frequency of up to 4.0 GHz. The Intel Turbo capability enables the processor to achieve a higher clock rate based on available power and thermal headroom. z1d instances come in 7 different instance sizes, including bare metal, and provide up to 48 vCPUs per instance.

z1d instances deliver a 1:8 vCPU to memory ratio and up to 384 GiB of RAM.  The higher ratio of memory to vCPU is ideal for memory intensive applications such as relational databases that don't require many vCPUs.  This can help reduce software licensing costs that are based on number of vCPUs.  In addition, the Optimize CPUs feature allows customers to reduce the vCPU count (as the compute requirement can be handled by lesser number of cores) and still maintain the same amount of memory.

z1d instances are built on the AWS Nitro System, a rich collection of building blocks that offloads many of the traditional virtualization functions to dedicated hardware. By doing so, the AWS Nitro System enables high performance, high availability, and high security while also reducing virtualization overhead.