Salad

Salad is a distributed computing platform that turns underutilized consumer and enterprise hardware into a shared resource for running AI, rendering, and other high-performance workloads. Its primary purpose is to provide developers and organizations with on-demand access to GPU and CPU capacity through a secure, scalable, and cost-efficient network. Users contribute their idle machines, while businesses tap into this pooled infrastructure via standardized interfaces and APIs.

Key features of Salad include a large, geographically distributed pool of GPUs suitable for training and inference of machine learning models, as well as support for containerized workloads to simplify deployment. The platform offers automated workload scheduling, resource allocation, and scaling, allowing users to match performance needs with available capacity. Built-in monitoring, logging, and usage analytics help teams understand cost, performance, and resource utilization over time. Salad also incorporates security and isolation mechanisms to protect workloads and host devices, including sandboxed execution and strict access controls.