Aleria and Hyperfusion to Build Modular AI Data Center Designed with NVIDIA Vera Rubin in Abu Dhabi

10MW development will pair the NVIDIA Vera Rubin platform with high-density, liquid-cooled modular infrastructure, targeting activation in under six months

UNITED ARAB EMIRATES, October 21, 2026 /EINPresswire.com/ -- News Summary
● Aleria and Hyperfusion are developing a 10MW modular AI data center in Abu Dhabi designed for NVIDIA Vera Rubin.
● The development is currently under construction, with service entry targeted for Q1 2027.
● Hyperfusion’s modular infrastructure is engineered for rack densities exceeding 200kW, with direct-to-chip liquid cooling and full redundancy across power, cooling and networking.
● The design targets a deployment cycle of less than six months from ground-break to activation.
● Aleria’s intelligence layer will connect the NVIDIA accelerated computing platform with workload orchestration, data management, agentic AI and enterprise applications.

BERLIN, Germany, October 20, 2026 (GTC) – Aleria, an AI infrastructure and platform company, and Hyperfusion, an Abu Dhabi-based AI infrastructure company, today announced the development of a 10MW modular AI data center in Abu Dhabi designed for NVIDIA Vera Rubin.

Currently under construction, the development is targeting service entry in Q1 2027, less than six months after ground-break. The modular infrastructure is being engineered around the power, cooling and density requirements of NVIDIA Vera Rubin, including rack densities exceeding 200kW MaxP and direct-to-chip liquid cooling.

The collaboration combines Hyperfusion’s land, power and modular physical infrastructure with Aleria’s and intelligence layer. Together, the companies are bringing the physical infrastructure, compute and software required to operate next-generation AI into a single environment.

“The next generation of AI infrastructure cannot take years to come online,” said Eric Leandri, CEO of Aleria. “The computing architecture is moving too quickly for that. In Abu Dhabi, we are bringing the infrastructure and compute onto the same cycle, so the next generation of NVIDIA AI infrastructure can move from bare-metal compute to production-ready capacity much faster.”

Infrastructure Designed Around NVIDIA Vera Rubin

NVIDIA Vera Rubin introduces new power and cooling requirements at the rack level, changing how the physical infrastructure around accelerated computing needs to be designed. Hyperfusion’s modular architecture is engineered for rack densities exceeding 200kW MaxP, with direct-to-chip liquid cooling, redundant power, cooling and networking, and structural capacity designed around high-density liquid-cooled systems.

A 45°C warm-water cooling loop is designed to reject heat efficiently in Gulf operating conditions while reducing water and energy consumption.
Rather than adapting a conventional data center to accommodate the computing environment after construction, the physical infrastructure is designed around the requirements of NVIDIA Vera Rubin from the outset.

From Ground-Break to Compute in Under Six Months

The Abu Dhabi development is targeting less than six months from ground-break to activation, with service entry planned for Q1 2027.
Hyperfusion’s modular approach moves much of the physical infrastructure into a repeatable manufacturing process, allowing power, cooling and networking systems to be configured around the computing architecture before deployment on site.

The approach addresses a growing challenge in AI infrastructure. Computing architectures are advancing on increasingly short cycles, while conventional data center development can take significantly longer. Modular infrastructure allows the physical environment to move more closely with the technology it is designed to operate.

“The speed of AI is changing what we expect from data center infrastructure,” said Quentin Reyes, Founder and CEO of Hyperfusion. “When the computing architecture changes every year, the infrastructure around it cannot take three years to deliver. We designed HyperPod to close that gap, bringing high-density power, liquid cooling and redundancy together in a modular environment that can be deployed on the same timeline as the technology it is built for.”

From Accelerated Computing to Production AI

Aleria’s intelligence layer connects the NVIDIA accelerated computing platform with workload orchestration, data management, agentic AI, enterprise applications and video AI. For critical environments, the architecture can extend into private and air-gapped deployments where infrastructure, data and AI workloads need to remain under greater organisational control.

The Abu Dhabi development adds another deployment model to Aleria’s sovereign AI operations in the UAE, combining next-generation accelerated computing with physical infrastructure designed around its power, cooling and density requirements.

About Aleria

Aleria is a full-stack AI company enabling organisations 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.

About Hyperfusion

Hyperfusion AI is an AI infrastructure company headquartered in Abu Dhabi, United Arab Emirates. The company designs and manufactures HyperPod, a modular AI data center platform and operates AI colocation and managed cloud services across the MENA region.

Faisal
Aleria
+971 52 200 1169
faisal@aleria.cloud
Visit us on social media:
LinkedIn
Instagram