Saudi AI infrastructure firm OmniOps launches Bunyan Inference Hub for local AI hosting

OmniOps, a Saudi AI infrastructure company, has launched Bunyan Inference Hub, a platform that lets governments, enterprises, startups and developers access and deploy open-source AI models while keeping their data and computing workloads hosted within Saudi Arabia.

The platform brings AI models and computing resources into a single environment. Organisations can choose from an existing catalogue or integrate their own models without building and managing separate infrastructure, and customers can purchase predefined token bundles or access dedicated GPU infrastructure for workloads requiring additional capacity.

Developed by OmniOps’ research and development team, the hub uses a proprietary inference optimisation engine that adapts AI models to available hardware. The technology is designed to improve GPU utilisation, reduce processing costs per token and maintain consistent performance as demand increases.

The platform supports different GPUs and AI accelerators, automatically directing requests to suitable computing resources and adjusting capacity according to demand. Organisations can also monitor performance, track token consumption and establish usage limits across teams.

The launch builds on OmniOps’ existing Bunyan infrastructure, introduced in July 2025 to support sovereign AI deployment within Saudi Arabia. By hosting data and processing locally, the company aims to address security and data residency requirements that can complicate the adoption of externally hosted AI services.

According to research cited by OmniOps from ServiceNow, Saudi organisations increased AI spending by 124% year-on-year and plan to allocate 20% of their IT budgets to AI through 2027. Gartner forecasts global spending on AI inference workloads to reach $23.3 billion by the end of 2026.