Scality launches on-premises AI inference stack

Scality has launched the AI Inference Factory, an integrated on-premises AI inference stack designed for enterprises, government agencies, and neo-cloud providers. The product combines four key components: validated open-weight models, a disaggregated inference-serving layer that separates prefill from decode to improve GPU efficiency, a control plane for authentication and scheduling, and Scality AI Data Infrastructure (ADI).

The ADI component utilizes policy-governed operations to manage model state and enterprise datasets. By storing and retrieving key-value cache from high-performance object storage rather than requiring GPUs to recompute context, the system extends inference capacity beyond the limits of GPU high-bandwidth memory. This architecture is intended to support agentic and context-heavy workloads.

Scality positions this solution as an alternative to cloud-based AI services, specifically targeting organizations concerned with unpredictable per-token cloud billing, the need for model-version stability, and strict data-sovereignty requirements. CEO Jérôme Lecat stated that the product provides the reliability and control necessary for mission-critical production environments.