AM Intelligence Orders 20,000 More NVIDIA GPUs in $20 Billion AI Infrastructure Push

AM Intelligence Commits $20 Billion to Scale AI Infrastructure

AM Intelligence, the AI infrastructure firm established by Greenko promoters, announced plans to invest over $20 billion in the next 15 months to dramatically expand its AI data center network1

1

. This massive capital expenditure comes on top of the $6 billion already committed for its initial 100 MW deployment, signaling one of the most aggressive AI infrastructure buildouts outside North America. The company placed two firm and binding orders for 20,000 NVIDIA GPUs, specifically the advanced Rubin chips configured as NVIDIA Vera Rubin NVL72 rack-scale systems2

2

. These orders follow the company's inaugural August 2024 commitment for 9,000 NVIDIA Rubin GPUs at its first AI factory in Hyderabad, bringing total committed capacity to approximately 29,000 GPUs across 100 MW.

Expansion Targets India and Malaysia Markets

The newly ordered 20,000 NVIDIA GPUs will power approximately 70 megawatts of AI compute capacity split between facilities in India and Malaysia, with delivery scheduled for Q2 20271

1

. This AI factory buildout establishes AM Intelligence among the earliest and largest committed buyers of Rubin GPUs outside North America. Both facilities will mirror the architectural standards of the Hyderabad AI Factory, integrating the NVIDIA Vera Rubin platform with high-throughput RDMA over Converged Ethernet (RoCE) networking for efficient data movement2

2

. Anil Chalamalasetty, Chairman of AM Intelligence, emphasized that each site is engineered for high rack power density and liquid-cooled infrastructure throughout, allowing the facilities to accommodate successive generations of AI silicon within the same physical footprint.

400 MW Pipeline Advances 1 Gigawatt Vision

AM Intelligence has developed a robust pipeline of approximately 400 MW of Compute-as-a-Service capacity across global locations including India, the United States, Europe, and Malaysia2

2

. The three orders placed to date represent only the first committed tranche within that pipeline, with the balance under development for delivery in subsequent phases throughout 2027 and early 2028. This pipeline advances the company's stated intention to bring 1 gigawatt of Compute-as-a-Service capacity to global AI workloads, positioned within a broader program targeting 5 GW of powered AI data centers across India, the United States, and Europe. The ambitious expansion plan aims to bring an additional 300 MWs of capacity to market within the next 15 months1

1

.

Strategic Positioning for AI Training and Inference at Scale

Vishal Dhupar, managing director of Asia South at NVIDIA, stated that AM Intelligence is creating the foundation for developers and enterprises to build more ambitious AI by combining accelerated computing with high-performance networking and liquid-cooled infrastructure2

2

. As a strategic builder of long-term assets, AM Intelligence plans to leverage its unique power solutions, capital expenditure advantage, and access to next-generation AI infrastructure to deliver optimized electron-to-token economics. The company aims to integrate energy infrastructure with computing capacity required for AI training and inference at scale, offering collaborative services spanning hyperscalers, neo cloud providers, sovereign AI initiatives, frontier labs, AI natives, and local developer communities. This approach positions the AI infrastructure firm to serve cloud providers, enterprises, and AI developers seeking large-scale computing capacity closer to where they build and deploy AI applications.