AI is changing the way data centres are being built, with high-performance GPUs (Graphics Processing Units) needed to train and run large AI models, process huge volumes of data and support AI applications. GPUs are particularly critical for AI training, inference and other workloads that require massive computing power. As these workloads become more demanding, data centres are moving towards high-density computing, advanced networking, better cooling systems and much higher power capacity.
This shift is now taking shape in Hyderabad, with AM Intelligence, the AI infrastructure company set up by the promoters of Greenko, placing a firm and binding order for around 9,000 NVIDIA Rubin GPUs. The chips will power its first 30 MW AI Factory in the city, with the facility set to deploy them as Vera Rubin NVL72 rack-scale systems and emerge as one of Asia’s first frontier AI compute clusters based on the new architecture.
The facility is being designed to handle some of the world’s largest AI workloads, including trillion-parameter models and next-generation agentic AI applications. It will combine NVIDIA Vera Rubin architecture with high-throughput RDMA (Remote Direct Memory Access) over Converged Ethernet (RoCE) networking and advanced storage systems to enable faster data movement and processing.
The Vera Rubin NVL72 systems use NVFP4 (NVIDIA’s 4-bit floating-point computing format), a low-precision format designed to run AI workloads more efficiently, along with next-generation HBM4 (High Bandwidth Memory 4). According to AM Intelligence, the architecture can reduce AI inference token-serving costs by up to 10 times compared with the previous NVIDIA Grace Blackwell generation. The Hyderabad AI Factory is being engineered to deliver around 450 exaFLOPS of NVFP4 inference compute.
Beyond Hyderabad, AM Intelligence is developing 5 GW of powered AI data centres across India, the US and Europe. These facilities will be built for high-density AI computing and use liquid cooling to manage the heat generated by advanced AI chips. The infrastructure will also be designed to support newer generations of AI processors as they become available.
The company plans to combine its energy and computing infrastructure to deliver competitive “electron-to-token” economics, serving hyperscalers, neocloud providers, sovereign AI initiatives, frontier AI labs, AI-native companies and developers. AM Intelligence is part of the wider AM Group, which includes Greenko, AM Green Ammonia and AM AI businesses.
“For over two decades, we have focused on transforming electrons into value and addressing critical and emerging societal needs. Today, the electron-to-token opportunity allows us to take this capability further by converting power infrastructure into frontier AI compute at scale,” said Anil Chalamalasetty, Chairman, AM Intelligence.
“Bringing the latest generation NVIDIA Vera Rubin to India marks the next step in this journey and establishes AMI at the forefront of the emerging AI electron-to-token economy,” he added.
Got questions or experiences to share? Comment below or connect with me on Instagram @samuelmjosh and LinkedIn: https://www.linkedin.com/in/samuel-joshua-maddela-ab210b1b9
