AMI places an $8 billion order for 9,000 NVIDIA Rubin GPUs to power a new AI factory in Hyderabad.
AM Intelligence (AMI), the AI infrastructure platform established by the promoters of Greenko, has placed a binding order for approximately 9,000 NVIDIA Rubin GPUs for deployment at its inaugural AI factory in Hyderabad. This marks one of the earliest large-scale deployments of NVIDIA’s next-generation Vera Rubin architecture in Asia. The cluster is scheduled for delivery in the first quarter of 2027 and will utilize NVIDIA Vera Rubin NVL72 rack-scale systems, occupying 30 MW of capacity at the Hyderabad facility. According to AMI, the deployment will establish the site as one of Asia’s first frontier-scale AI compute clusters and serve as the initial tranche of its planned 1 GW Compute-as-a-Service platform across India, the US, Finland and Malaysia.
The company intends to initially bring 200 MW of AI compute capacity to market, requiring capital expenditure exceeding $8 billion. This investment comes as infrastructure firms and data centre operators compete to build the necessary power, cooling, and compute resources for increasingly large AI models.
“For over two decades, we have focused on transforming electrons into value — addressing critical and emerging societal needs. Today, the electron-to-token opportunity allows us to take this capability further, converting power infrastructure into frontier AI compute at scale,” said Anil Chalamalasetty, chairman of AM Intelligence. “Bringing 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.”
AMI is leveraging its energy assets, including access to renewable generation and storage through Greenko, as a core competitive advantage for building energy-intensive AI infrastructure. The Hyderabad facility will integrate Vera Rubin systems with high-throughput RDMA over Converged Ethernet (RoCE) networking and advanced storage infrastructure, enabling the execution of trillion-parameter models and next-generation agentic AI applications. The facility is engineered to deliver approximately 450 exaFLOPS of NVFP4 inference compute. NVIDIA’s Vera Rubin architecture introduces the NVFP4 low-precision computing format alongside next-generation High Bandwidth Memory (HBM4), both designed to enhance the efficiency and economics of serving AI workloads.
The Hyderabad deployment aligns with India’s broader efforts to expand domestic AI computing infrastructure and reduce reliance on overseas capacity for training and running advanced models. AMI stated that its platform will serve hyperscalers, neo-cloud providers, sovereign AI programmes, frontier AI labs, AI-native companies and local developer communities. Beyond its initial 1 GW Compute-as-a-Service roadmap, AMI is developing 5 GW of powered AI data centres across India, the US and Europe. These upcoming facilities are being designed for high-density computing and liquid cooling, critical requirements as successive generations of AI chips demand significantly higher power levels and generate greater thermal output.