AM Intelligence (AMI), an AI infrastructure platform established by Greenko’s founders, has made two additional orders totaling 20,000 Nvidia Rubin graphics processing units (GPUs) to be utilized in India and Malaysia.

The company plans to implement the GPUs in NVIDIA Vera Rubin NVL72 rack-scale systems throughout India and Malaysia. AMI will enhance the capacity of its initial AI factory in Hyderabad.

The growth will offer substantial computing resources nearer to AI developers and businesses.

The latest orders come after the company's first promise of 9,000 NVIDIA Rubin GPUs revealed on 25 August 2026. The extra capacity will amount to approximately 70 megawatts (MW), with shipments planned for the second quarter of 2027. Collectively, the three orders elevate AMI’s pledged capacity to approximately 29,000 Rubin GPUs and close to 100 MW.

AMI intends to introduce an extra 300 MW of capacity to the market in the upcoming 15 months. The growth will necessitate over $20 billion in capital investment, in addition to the $6 billion pledged for the initial 100 MW.

“Both facilities will be built to the same architecture standard as the Hyderabad AI Factory, integrating the NVIDIA Vera Rubin platform with high-throughput RDMA over Converged Ethernet (RoCE) networking for fast, efficient data movement, alongside advanced storage systems, ” says Anil Chalamalasetty, Chairman, AM Intelligence.

Also Read: DeepSeek Partners with Huawei to Build Tools for Ascend AI Chips

“Each site is engineered for high rack power density and liquid cooled throughout, allowing the infrastructure to accommodate successive generations of AI silicon within the same physical footprint.”

Vishal Dhupar, managing director, Asia South, NVIDIA, said, “AM Intelligence is bringing the next generation of AI factories to India and Malaysia with NVIDIA Vera Rubin”.

Also Read: SiMa.ai Hits $1.45B Valuation after Raising $500M to Scale Physical AI

“By combining accelerated computing with high-performance networking and liquid-cooled infrastructure, AMI is creating the foundation for developers and enterprises to build more ambitious AI”.

The firm has established a worldwide pipeline of approximately 400 MW of Compute-as-a-Service capability. The pipeline extends to areas throughout India, the US, Europe, and Malaysia. The three disclosed orders signify the initial committed portion within this pipeline.

The firm intends to carry out additional shipments until 2027 and the beginning of 2028. The organization intends to deliver 1 GW of Compute-as-a-Service capacity for worldwide AI tasks. This is a component of a larger initiative aimed at 5 GW of AI-powered data centers throughout India, the US, and Europe.

Also Read: HCLSoftware to Buy Croatia RPA Firm Robotiq.ai

AMI aims to combine energy infrastructure with computing power to facilitate large-scale AI training and inference. The service model will focus on hyperscalers, neo-cloud providers, sovereign AI projects, frontier labs, AI-native firms, and developer communities.