Hyderabad:AM Intelligence (AMI), an AI infrastructure platform established by the promoters of Greenko, has placed firm orders for 20,000 NVIDIA Rubin GPUs. These GPUs will be deployed as NVIDIA Vera Rubin NVL72 rack-scale systems in India and Malaysia.

According to BERNAMA News Agency, This move builds upon the initial AI factory set up in Hyderabad, India, and aims to provide enhanced computing capacity for cloud providers, enterprises, and AI developers. The new orders follow AMI's earlier commitment announced on August 25, which involved 9,000 NVIDIA Rubin GPUs at the Hyderabad facility.

The expansion will add approximately 70 megawatts of capacity and is scheduled for delivery in the second quarter of 2027. Both the Indian and Malaysian facilities will mirror the architecture of the Hyderabad AI Factory, featuring high-throughput RDMA over Converged Ethernet networking for efficient data movement and advanced storage systems. The infrastructure is designed for high rack power density and uses liquid cooling to support future generations of AI silicon.

With the latest orders, AMI's total commitment now stands at approximately 29,000 Rubin GPUs and nearly 100 megawatts of capacity. The company also plans to introduce an additional 300 megawatts of capacity over the next 15 months, with an investment exceeding US$20 billion.

AMI has outlined a global pipeline of about 400 megawatts of Compute-as-a-Service capacity across locations including India, the United States, Europe, and Malaysia. These orders are the first committed segment within this pipeline, with the remainder planned for delivery in phases through 2027 and early 2028. This initiative supports AMI's goal of providing one gigawatt of Compute-as-a-Service capacity for global AI workloads within a larger program aiming for five gigawatts of powered AI data centers.