AMI orders 9,000 NVIDIA Rubin GPUs for Hyderabad AI factory

The Greenko-promoted AI infrastructure platform is investing more than $8 billion in an initial 200MW rollout, with the Hyderabad facility expected to deploy next-generation Rubin systems for large-scale AI workloads.

AM Intelligence (AMI), an AI infrastructure company promoted by the founders of Greenko, has placed a binding order for around 9,000 NVIDIA Rubin GPUs for its first AI factory in Hyderabad. The systems are scheduled for delivery in the first quarter of 2027, positioning the facility among the early large-scale deployments of NVIDIA’s next-generation Vera Rubin architecture in Asia.

The Hyderabad project will serve as the first phase of AMI’s broader plan to establish 1 gigawatt (GW) of compute-as-a-service capacity across India, the US, Finland and Malaysia. The company plans to bring around 200 megawatts (MW) of capacity to market initially, with the associated capital expenditure expected to exceed $8 billion.

AMI is building its infrastructure around an integrated model that combines energy resources, computing infrastructure and AI capabilities. The company aims to provide cost-efficient compute capacity for a wide range of customers as demand for AI infrastructure expands globally.

Hyderabad facility targets advanced AI workloads

The Hyderabad AI factory is being designed to support some of the most computationally demanding AI applications, including trillion-parameter models and emerging agentic AI systems. By adopting NVIDIA’s Vera Rubin architecture, AMI expects to provide high-speed data movement and advanced storage capabilities alongside accelerated AI computing.

The NVIDIA Vera Rubin NVL72 platform introduces NVFP4, a low-precision computing format intended to improve efficiency for AI workloads. The architecture also incorporates next-generation HBM4 memory, supporting the bandwidth requirements of increasingly complex AI models.

AMI said the Hyderabad facility is being engineered to deliver approximately 450 exaFLOPS of NVFP4 inference computing capacity. This is intended to provide the processing scale required to operate advanced AI models and applications.

Anil Chalamalasetty, chairman of AMI, said: “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. 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 expands AI infrastructure footprint

Beyond Hyderabad, AMI is developing around 5GW of AI data centre capacity across India, the US and Europe. The facilities are being designed for high-density AI computing and will use liquid-cooling technologies to manage the power consumption and heat generated by advanced AI processors.

The infrastructure is also being planned to accommodate successive generations of AI silicon, allowing AMI to upgrade computing environments as accelerator technologies evolve. The company intends to offer compute services to hyperscalers, neocloud providers, sovereign AI programmes, frontier AI laboratories, AI-native companies and developer communities.

AMI’s strategy extends beyond data centre capacity, with plans for a full-stack AI infrastructure ecosystem covering energy systems, data centres, hardware and customised AI models. The company said its approach is built around its experience in developing large-scale energy infrastructure and applying that capability to the growing demand for AI compute.

The company’s near-term 200MW rollout forms part of its larger 1GW compute-as-a-service ambition, while its planned 5GW data centre footprint is intended to establish a broader global platform for AI workloads.