QIT

Active / Live

QIT

GPU containers. Built for AI.

QIT builds modular AI infrastructure around available power. The company deploys liquid-cooled GPU containers, shipping-container-style modular data centers for AI inference, training and other GPU workloads, at sites that already have usable electrical capacity. Projects are planned around a six-month deployment target from an agreed project scope.

QIT also runs its own compute platform: a Model API with 407 published models across image, video, audio, text, 3D and training workloads (pay-as-you-go, only successful requests billed); serverless GPUs (QIT Pods) that scale from zero and are billed per second; and reserved capacity and bare metal at a fixed fee per GPU-hour.

Published GPU profiles include L40S (from $1.25/hr), RTX PRO 6000 ($1.55/hr), H100 ($3.08/hr), H200 ($3.51/hr), B200 ($4.68/hr) and B300 ($5.46/hr). New business accounts receive $1.56 in free credits. The Model API and serverless control plane carry a 99.99% monthly uptime SLA, EU and US data residency is available, and customer content is not used to train models.

Contact: ismail@qit.energy