OmniOps launches Bunyan Inference Hub to keep Saudi AI workloads in-country

OmniOps, a Saudi sovereign AI infrastructure company, has launched the Bunyan Inference Hub, a platform that gives governments, enterprises, startups and developers access to open-source AI models and computing resources while keeping data and workloads inside the kingdom.

The problem it addresses sits between a working prototype and a running service. Moving AI applications into production multiplies infrastructure complexity. Each model can arrive with its own provider, its own computing environment and its own technical integrations. Organizations bound by strict data residency or security requirements often decide the only safe path is to build and operate that infrastructure themselves.

Bunyan Inference Hub pulls model access, computing resources and infrastructure management onto a single platform. Users can work from a catalogue of AI models or bring their own, and draw on compute without assembling separate infrastructure for every model or use case. Customers buy predefined token bundles across supported models and workloads. Those needing more capacity or isolation can take dedicated GPU infrastructure.

Underneath runs an inference optimization engine developed by OmniOps' R&D team. It adapts how AI models execute on the hardware available to match the requirements of a given workload, lifting GPU utilization, cutting cost per token and holding performance steady as demand grows. The platform operates across different GPUs and AI accelerators rather than depending on one hardware supplier. It routes requests to the appropriate model and compute, manages scaling, monitors performance and usage, and lets customers set usage limits and track token consumption across teams.

"Managing AI model consumption efficiently with correct security and cost guardrails has never been more critical," said Mohammed Altassan, founding CEO of OmniOps. He said the hub combines the company's proprietary Bunyan Inference technology with its sovereign AI deployment experience, and that public and private sector organizations can plug it into their operations to deploy AI solutions more efficiently.

Demand for inference capacity is climbing worldwide as more organizations push models into production. ServiceNow research shows Saudi-based entities increased AI spending by 124% year on year and plan to allocate 20% of IT budgets to AI solutions through 2027. Gartner forecasts global spending on AI inference workloads will reach $23.3 billion by the end of 2026.

For buyers weighing the platform, the practical question is narrower than the market forecasts: whether metered token bundles and shared accelerator pools can satisfy residency and security reviews that have so far pushed many of them toward running their own clusters.