Core42 Enhances AI Infrastructure for Secure UAE Government Services Deployment

Core42 Enhances AI Infrastructure for Secure UAE Government Services Deployment

Core42 is advancing the deployment of secure and scalable AI infrastructure for UAE government services, a critical step as the nation aims to transition 50% of its government operations to Agentic AI. This initiative is designed to ensure that intelligence is delivered securely, economically, and at production scale, addressing the increasing complexity of autonomous agents that generate multiple inference events to complete tasks.

In an interview with Tahawultech.com, Raghu Chakravarthi, Chief Product and Technology Officer at Core42, emphasized the importance of integrated compute, cloud, data, and governance capabilities. These elements are essential for moving Agentic AI beyond isolated pilot projects and into comprehensive government-wide deployment.

Building a Sovereign AI Infrastructure

Chakravarthi highlighted that a sovereign, scalable AI infrastructure is foundational for achieving the UAE’s ambitious goals. The shift to Agentic AI changes the economic dynamics of AI, where a single user request can trigger numerous inference events. This shift necessitates careful management of costs, as consumption scales with autonomy rather than headcount.

What sets the Gulf region apart is its simultaneous pursuit of sovereignty and scale in AI infrastructure. This approach ensures that sovereign control is a primary requirement, supported by significant national investments in computing resources. The result is an infrastructure that is locally governed, globally competitive, and designed for large-scale AI adoption from the outset.

Key Capabilities for Effective AI Deployment

To facilitate the transition from pilot projects to secure government-wide deployment, Core42 emphasizes the need for a cohesive system that integrates compute, cloud, data, and governance. AI workloads demand diverse infrastructure capabilities, including various accelerator architectures that cater to different performance metrics such as latency, throughput, and cost.

Chakravarthi noted that the deployment architecture must allow for flexibility across real-time and batch processing, as well as cloud and on-premises environments. This flexibility is achieved through a workload classifier that automatically routes requests to the most suitable silicon and model, ensuring optimal performance and cost-efficiency.

Moreover, embedding governance directly into the inference layer is crucial. This includes in-country data residency, encryption, access controls, and strict data handling policies, all supported by over 170 security policies and SOC 2 Type II certification.

Controlling Costs and Enhancing Performance

As the deployment of Agentic AI increases, managing inference costs becomes paramount. Government entities must focus on the economics of the entire agentic workflow, understanding how many inference events each task generates and which stages require low latency. Cost controls should be implemented proactively, with budgets and alerts established before consumption scales.

Workload-aware orchestration is key to improving performance, sovereignty, and economic efficiency. By treating infrastructure as a dynamic workload-placement decision, Core42 aims to match tasks with the appropriate models and accelerators, optimizing for both quality and cost. This approach not only enhances the responsiveness of citizen-facing applications but also ensures better utilization of resources while maintaining strong control over sensitive data.