NxtGen AI and Vertiv build sovereign AI factory in India

India’s sovereign Cloud ambitions have moved into execution mode. A sovereign AI factory in India is now being deployed with more than 4,000 NVIDIA Blackwell GPUs, integrated rack systems and liquid-cooled servers designed for high-density AI workloads.

The project brings together compute, power and thermal infrastructure in a single, tightly integrated stack. The aim is clear: support AI training and inference at population scale within India’s sovereign Cloud framework.

High-density compute at population scale

At the heart of the sovereign AI factory in India are over 4,000 NVIDIA Blackwell GPUs. These are being deployed through Dell Integrated Rack Scalable Systems IR5000, running on Dell PowerEdge XE9685L liquid-cooled servers.

The infrastructure is designed for high-performance AI training and inference. The emphasis is on density. More compute per rack. More performance per square foot. And sustained operation under heavy AI workloads.

This is not a pilot cluster. It is national in scale.

A. S. Rajgopal, MD and CEO, NxtGen AI, said building a national-scale AI factory requires accelerated computing along with a resilient and scalable AI infrastructure foundation. He added that the deployment supports the vision of strengthening India’s sovereign AI capabilities.

Power and thermal backbone

The physical datacentre layer is being delivered by Vertiv. The focus is on handling extreme power densities and cooling requirements typical of modern AI systems.

The power train includes:

• Distribution transformers

• Switchboards

• UPS systems

• Room power distribution units

• Racks and rack PDUs

On the thermal side, the deployment integrates advanced liquid cooling systems. These include:

• High-capacity Coolant Distribution Units

• Modular secondary piping infrastructure skids

• High-efficiency perimeter cooling systems

• Ambient-compatible heat rejection units

The design allows heat to be removed directly using water at ambient temperatures. This reduces cooling overhead and improves thermal stability for high-value GPUs.

In practical terms, this means lower energy consumption for cooling and more stable GPU performance during sustained AI training runs.

Liquid cooling and lithium-ion energy storage

The sovereign AI factory in India combines liquid cooling with lithium-ion battery-based energy storage systems.

Compared to traditional lead-acid batteries, lithium-ion systems offer:

• Longer service life

• Fewer replacement cycles

• Reduced material waste

• Lower operational emissions over time

Together, liquid cooling and lithium-ion power systems are positioned to lower the energy footprint of high-density AI infrastructure compared to conventional cooling and power models.

Subhasis Majumdar, Managing Director, Vertiv India, said AI factories are changing how datacentres are designed and operated. He noted that the infrastructure approach supports the latest generation Dell compute systems accelerated by NVIDIA Blackwell GPUs, with a focus on power efficiency and scalable operations.

Infrastructure built for sovereign Cloud demand

The deployment is supported by a domestic supply chain and regional engineering capabilities, including an Integrated Business Services hub in Pune. The stated goal is to meet rising demand for sovereign Cloud and AI infrastructure across enterprises, startups and public-sector organisations in India.

The sovereign AI factory in India reflects a broader shift. AI is no longer confined to isolated research clusters. It is moving into national infrastructure territory, where power density, cooling efficiency and operational resilience become as critical as compute performance.

For enterprises and public institutions evaluating AI at scale, this deployment offers a clear signal. Sovereign AI capability is no longer conceptual. It is being built, rack by rack, GPU by GPU, within India’s borders.

The next phase will be defined not just by hardware volumes, but by how effectively this infrastructure supports real-world AI workloads across industries.

Read More: 

Databricks APJ Financial Services GTM Lead Appointment Strengthens Regional Strategy

Redington and GENERA 3D Printed Eyewear Partnership India Announced