GPUX.AI

GPUX is a serverless GPU platform for running machine learning inference. A developer calls a model through a simple HTTP request, for example a single curl command that generates an image from a prompt with Stable Diffusion XL, and the platform runs it on GPU hardware without the developer managing servers. The site lists models such as Stable Diffusion XL, ESRGAN upscaling, Whisper speech recognition and an Alpaca language model. The platform describes one second cold starts, read and write storage volumes and peer to peer capacity. Owners of private models can sell inference requests on those models to other organizations, and the site also invites people to earn by contributing GPU capacity from Windows or Linux machines. A blog covers releases, how to guides, case studies and performance work, including a post on running Stable Diffusion XL faster on RTX 4090 cards. The product is aimed at developers and teams who want to deploy a model behind an API quickly, or who want to run GPU workloads on demand. The company, GPUX Inc., is based in Toronto, and the team page lists staff in several countries. The most recent announcements and posts on the site date from 2023. The site does not publish prices. Visitors are directed to open the V2 app or to talk to a founder.