AM Intelligence orders 20,000 NVIDIA Rubin GPUs, plans $20 billion AI infrastructure expansion

AM Intelligence Secures Major GPU Order for Asia Expansion

AM Intelligence announced two firm orders for 20,000 NVIDIA Rubin GPUs1

1

, positioning the AI infrastructure firm among the largest committed buyers of NVIDIA Vera Rubin platform hardware outside North America. The GPUs will be deployed as NVL72 rack-scale systems across facilities in India and Malaysia, with delivery scheduled for Q2 20272

2

. This expansion follows the company's August announcement of 9,000 NVIDIA Rubin GPUs for its Hyderabad AI factory3

3

, bringing total committed capacity to approximately 29,000 GPUs and close to 100 MW.

The new orders add roughly 70 MW of AI data center capacity to AM Intelligence's portfolio4

4

. Each facility will integrate the NVIDIA Vera Rubin platform with high-throughput RDMA over Converged Ethernet networking for efficient data movement, alongside advanced storage systems and liquid cooling infrastructure. This architecture allows the sites to accommodate successive generations of AI silicon within the same physical footprint, according to AM Intelligence Chairman Anil Chalamalasetty1

1

.

$20 Billion Capital Commitment Signals Aggressive Growth

AM Intelligence plans to invest more than $20 billion over the next 15 months to bring an additional 300 MW of AI compute capacity to market1

1

. This capital expenditure comes on top of the $6 billion already committed for the initial 100 MW deployment. The scale of investment reflects the company's bet that demand for Compute-as-a-Service will accelerate as enterprises shift from experimenting with AI tools to deploying AI agents and reasoning models at scale2

2

.

The firm, backed by Greenko promoters, has developed a pipeline of approximately 400 MW of Compute-as-a-Service capacity across India, the United States, Europe, and Malaysia3

3

. The three GPU orders placed to date represent the first committed tranche within that pipeline, with the balance under development for delivery in subsequent phases throughout 2027 and early 20284

4

.

India Emerges as AI Infrastructure Hub

The expansion positions Hyderabad at the center of India's AI infrastructure buildout. The city already serves as a technology and data center hub, and AM Intelligence's AI factory gives it a critical role in supporting the next wave of AI workloads2

2

. This development marks a shift in India's AI strategy from software and talent to physical computing infrastructure capable of supporting AI training and inference at scale.

Vishal Dhupar, managing director for Asia South at NVIDIA, noted 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 infrastructure3

3

. The deployment addresses a critical bottleneck for Indian AI companies: access to affordable, powerful computing capacity for training, fine-tuning, and deploying models.

Long-Term Vision Targets 1 Gigawatt Capacity

AM Intelligence's stated goal extends beyond the current orders. The company aims to deliver 1 gigawatt of Compute-as-a-Service capacity to global AI workloads, within a broader program of 5 GW of powered AI data centers across India, the United States, and Europe1

1

. This ambitious roadmap positions the firm to serve hyperscalers, neo cloud providers, sovereign AI initiatives, frontier labs, AI natives, and local developer communities4

4

.

The company plans to leverage what it calls electron-to-token economics by integrating energy infrastructure with computing capacity required to train and run AI models3

3

. This approach reflects the firm's background in power solutions through its Greenko connection, potentially offering cost advantages in an industry where energy consumption represents a significant operational expense. As AI development increasingly becomes a race for computing power rather than just better models, AM Intelligence's infrastructure-first strategy could prove decisive in determining which regions can support the next generation of AI applications.