Indian data centre company AM Intelligence recently placed an order for roughly 9,000 Nvidia Rubin GPUs, aiming to deploy one of Asia’s first Vera Rubin clusters at its new AI factory located in southern India.
The company said the GPUs will be deployed as Nvidia Vera Rubin NVL72 rack-scale systems at a 30MW AI factory in the city of Hyderabad. The facility is slated to come online in Q1 2027.
AM Intelligence said the site is being designed to run trillion-parameter models and agentic AI applications and will target customers including cloud providers, AI labs and organisations developing sovereign AI models. It further explained Vera Rubin’s platform architecture is designed to cut AI inference token-serving costs by up to 10 times compared with Nvidia’s previous Grace Blackwell generation.
“The electron-to-token opportunity allows us to take this capability further, converting power infrastructure into frontier AI compute at scale”, noted AM Group Chairman Anil Chalamalasetty.
Based in Hyderabad, AM Intelligence is the AI infrastructure arm of Indian green energy and technology conglomerate AM Group. The firm builds large-scale, renewable-powered AI computing infrastructure and data centres to support AI models, cloud-service providers and AI labs.
The Hyderabad site marks the first phase of AM Intelligence’s planned buildout of 1GW of Compute-as-a-Service capacity across India, the U.S., Finland and Malaysia. The company plans to bring an initial 200MW online, with the wider buildout involving more than $8 billion in capital expenditure.
By 2030, AM Intelligence aims to hit 5GW of data centre capacity across India, US and Europe, with facilities designed for high-density AI compute and liquid cooling.
Source: Mobile World Live
Image Credit: ShutterStock
Source: Tahawul Tech