Dell to Power NxtGen AI Factory with Nvidia Blackwell GPUs in India

New AI factory aims to deliver high-end GPU access for enterprises startups and public-sector users.

Topics

  • Dell Technologies will partner with Indian sovereign cloud provider NxtGen AI Pvt. Ltd to build what the companies call India’s first large-scale dedicated AI factory, aimed at expanding domestic computing capacity.

    The project is designed to support large-scale AI workloads, including generative, agentic and high-performance models, for enterprises, startups, research institutions and government programs.

    As part of the partnership, NxtGen will deploy Dell’s AI Factory with Nvidia solutions to build a high-performance AI cluster powered by more than 4,000 Nvidia Blackwell GPUs, supported by Nvidia BlueField-3 data processing units and Spectrum-X networking. 

    The deployment will include Dell’s liquid-cooled PowerEdge XE9685L servers, delivered through integrated rack systems, along with additional storage and compute hardware. 

    The cluster will be hosted entirely within India and operated under a sovereign cloud framework, keeping data and AI workloads within national borders.

    “We made a choice of NVIDIA B200SXM housed in Dell Technologies PowerEdge XE9685L servers to support large-scale AI training and inference within India,” NxtGen Cloud Technologies Managing Director and Chief Executive Officer Rajgopal A.S. said in a LinkedIn post.

    He said the AI cluster connects 4,096 GPUs using Nvidia BlueField-3 data processing units, creating what he described as a high-performance training environment.

    On storage, Rajgopal said the company is deploying 60 nodes of Dell PowerScale F710 along with high-performance NVMe disks dedicated to each GPU to support checkpoint storage and parallel file systems used during model training.

    “Training clusters require local storage for storing checkpoints and a parallel file system,” he said.

    He added that cluster stability has been addressed through the use of Nvidia transceivers and extensive optical fibre cabling supplied by CommScope.

    “Key to the stability of the cluster comes from the choice of NVIDIA transceivers and nearly 276 km of optic cables,” Rajgopal said.

    Cooling and power efficiency were also a focus, he said, with the infrastructure using direct-to-chip liquid cooling from Vertiv India and operating without chillers.

    “The cluster is kept cool with water running at room temperature, we do not use any chillers, saving power,” he said, adding that Vertiv is also providing power equipment and lithium-ion battery-backed UPS systems.

    Rajgopal said the deployment is completed with the full Nvidia AI Enterprise software stack and exceeds Nvidia’s reference architecture specifications.

    The announcement comes as India increases its focus on building domestic AI infrastructure, alongside investments in digital public infrastructure and advanced computing. Policymakers have increasingly emphasised data sovereignty and local compute capacity, particularly for sensitive government and enterprise workloads.

    Dell India President and Managing Director Manish Gupta said the collaboration reflects the need for scalable and secure infrastructure as AI adoption accelerates across sectors. 

    Topics

    More Like This

    You must to post a comment.

    First time here? : Comment on articles and get access to many more articles.