StackPath Introduces NVIDIA GPU Accelerated Edge Compute Instances

We are delighted to announce GPU-Accelerated StackPath Edge Compute Virtual Machine (VM) and Container instances powered by the NVIDIA A2 Tensor Core and NVIDIA A16 GPUs. We joined the NVIDIA Partner Network to bring GPU acceleration to the edge of the internet to provide data scientists and business analysts the capabilities they need to transform information into actions.

StackPath eliminates a major hurdle for AI inference and computer vision workloads by providing GPU-Accelerated computing that’s not only fast but also consistent. Our infrastructure’s geographic proximity to major metro locations minimizes network hops to improve performance and minimize latency

StackPath’s high-capacity VM and container configurations are now GPU-Accelerated on a platform with familiar controls, robust APIs, image capture and storage, autoscaling, and more. Use powerful computing with ease near users to provide inherent performance and cost efficiencies on an edge-native IaaS platform where data reaches instances up to 300% faster than on hyperscale cloud instances. Customers will see the impact of StackPath’s low latency for AI inference, computer vision, and natural language processing workloads while experiencing the ease of use, consistency, and efficiency our platform already provides.

GPU-Accelerated Edge Compute instances are currently available in Dallas, San Jose, and Frankfurt; locations will be added throughout 2024. If you have a specific location where you would like to utilize GPU acceleration at the edge, please contact us,and we will work to prioritize your next GPU-Accelerated AI workload.

GPU-accelerated instance pricing

StackPath Edge Compute instances are provisioned on demand through the StackPath Customer Portal or API. Instances are billed by the hour and by the volume of data transferred. Additional options include forming virtual private clouds and leveraging built-in L3-L4 DDoS protection, persistent storage, image capture and deployment, private IP addresses, and more.