Aleria는 NVIDIA Rubin GPU 14,000기를 도입해 UAE 소버린 AI 역량을 확대하고 있다.

**News Summary**

- Aleria is expanding sovereign AI capacity in the UAE through a dedicated data center.

- Aleria will deploy approximately 14,000 NVIDIA Rubin GPUs at the 60MW data center.

- The 60MW deployment will also include 6,000 NVIDIA Groq 3 LPX accelerators, 8,000 NVIDIA Vera CPUs, and NVIDIA CMX integrated with NVIDIA Dynamo, supported by multiple petabytes of DDN's intelligent data layer delivering more than 17 TB/s of aggregate read throughput.

- Aleria's intelligence layer will connect the NVIDIA accelerated computing platform with workload orchestration, data management, agentic AI, and enterprise applications.

- The additional capacity is designed for large-scale AI training and inference in critical environments across the UAE.

Aleria, a full-stack AI company, today announced an expansion of its sovereign AI capacity in the UAE through a dedicated 60MW data center.

The facility is being developed around a dedicated computing profile. The data center will deploy approximately 14,000 NVIDIA Rubin GPUs alongside 6,000 NVIDIA Groq 3 LPX accelerators, 8,000 NVIDIA Vera CPUs, and NVIDIA CMX Context Memory Storage integrated with NVIDIA Dynamo.

The expansion builds on Aleria's existing sovereign AI operations in the UAE, adding computing capacity as demand for locally operated AI grows across critical environments.

"The UAE has made its AI ambition clear, and the speed at which it is building around that ambition is remarkable," said Eric Leandri, CEO of Aleria. "We are no longer explaining why it matters. The question now is how quickly we can add capacity to meet demand. We are doing exactly that, expanding a model that is already running real workloads and putting more sovereign AI capacity where it is needed."

**Adding Capacity to a Working Model**

Aleria's sovereign AI infrastructure model is already operating in the UAE, combining accelerated computing with the software and applications required to put AI into production. The expansion increases the computing capacity available to that environment as the scale and complexity of AI workloads grow.

Aleria's intelligence layer connects the NVIDIA accelerated computing platform with workload orchestration, data management, agentic AI, enterprise applications, and video AI. It links the underlying computing infrastructure with the workloads and applications organizations ultimately operate.

For critical environments, the same architecture can extend into private and air-gapped deployments, where infrastructure, data, and AI workloads need to remain under greater organizational control. The additional capacity will support large-scale training and inference, providing the computing foundation required to run increasingly demanding AI workloads.

**Sovereign AI Capacity Across the Deployment**

The data center will provide the AI computing environment, built around approximately 14,000 NVIDIA Rubin GPUs, 6,000 NVIDIA Groq 3 LPX accelerators, and 8,000 NVIDIA Vera CPUs. NVIDIA Groq 3 LPX adds specialized inference capacity for ultrafast token generation and latency-sensitive agentic AI workloads. The NVIDIA CMX Context Memory Storage Platform, integrated with NVIDIA Dynamo, will support the growing context and memory requirements of advanced inference workloads. The architecture will be supported by multiple petabytes of DDN's intelligent data layer, delivering more than 17 TB/s of aggregate read throughput.

NVIDIA Vera Rubin NVL72 systems will provide the accelerated computing foundation, while Aleria's intelligence layer connects that capacity with orchestration, data management, agents, and applications.

**From Infrastructure to Intelligence**

The deployment combines accelerated computing infrastructure with Aleria's intelligence layer. Aleria brings together the NVIDIA accelerated computing platform with its own intelligence layer, connecting the infrastructure to orchestration software and AI applications. The result is an integrated infrastructure model spanning physical capacity and power, accelerated computing, software, and production AI.

**About Aleria**

Aleria is an Abu Dhabi-based, full-stack AI company that enables organizations to deploy and scale AI in critical environments. As an NVIDIA Cloud Partner, Aleria combines NVIDIA accelerated computing with its own software to deliver integrated capabilities, from AI factories to enterprise applications. Its global offering incorporates data sovereignty and operational autonomy, with private and air-gapped deployment options. For more information, visit aleria.ai.