Hyderabad-based AI firm AM Intelligence placed a massive order for 9,000 Nvidia Vera Rubin GPUs to power its new regional AI Factory.

Hyderabad-based AI infrastructure company AM Intelligence (AMI) has placed a firm binding order for 9,000 Nvidia Vera Rubin GPUs, deployed as Vera Rubin NVL72 rack-scale systems. This acquisition potentially positions AMI as one of the first adopters of the advanced computing platform in Asia. Scheduled for delivery in the first quarter of 2027, the hardware will power the initial 30-megawatt phase of AMI’s upcoming AI factory in Hyderabad, establishing it as one of the region’s first frontier AI compute clusters.

Established by the promoters of Greenko, AM Intelligence operates as a vertically integrated AI infrastructure platform focused on energy, compute, and innovation. The order represents the company’s first concrete step toward its vision of becoming a leading intelligence infrastructure provider for the AI economy. “For over two decades, we have focused on transforming electrons into value, addressing critical and emerging societal needs,” said Anil Chalamalasetty, Chairman of AM Intelligence. “Today, the electron-to-token opportunity allows us to take this capability further, converting power infrastructure into frontier AI compute at scale. 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.”

To support these ambitions, the Hyderabad AI Factory is engineered with high-throughput RDMA over Converged Ethernet (RoCE) networking and advanced storage systems to enable rapid, efficient data movement. This architecture is designed to run some of the world’s largest and most sophisticated AI models, including trillion-parameter models and next-generation agentic AI applications. The Nvidia Vera Rubin NVL72 introduces NVFP4, a new low-precision computing format optimized for AI workloads, alongside next-generation High Bandwidth Memory (HBM4). Designed to reduce AI inference token-serving costs by up to ten times compared to the preceding Nvidia Grace Blackwell generation, the system will deliver approximately 450 exaFLOPS of NVFP4 inference compute, providing massive capacity for scaling advanced AI models.

Beyond Hyderabad, AMI is developing five gigawatts of powered AI data centres across India, the United States, and Europe. These facilities are built for high-density AI computing and utilize liquid cooling to manage the substantial power and thermal demands of advanced AI chips, while maintaining the flexibility to accommodate successive generations of silicon. As part of this broader initiative, the Hyderabad facility serves as the first tranche of AMI’s plan to deploy one gigawatt of Compute-as-a-Service capacity across India, the United States, Finland, and Malaysia. Approximately 200 megawatts are scheduled to come online in the near term, with the entire buildout backed by capital expenditure exceeding USD 8 billion.