Research Note: Inference Hardware on the Edge

Type: Research Report

This research note examines the rapidly evolving ecosystem of inference hardware for edge deployment in capital markets.

We explore available technologies, highlighting trends and useful categorization of products. We also discuss the different architectures by their ability to meet the demands of Latency (inference speed) and Throughput (inference volume) as well as the constraints imposed by the edge.

The report seeks to aid decision-makers in navigating the complexities of hardware on the edge.