Indian AI infrastructure platform AM Intelligence (AMI) has placed a binding purchase order for 9,000 Nvidia Vera Rubin computing systems. The procurement represents one of the earliest hyperscale commitments for Nvidia's next-generation Rubin architecture across Asia and anchors an $8 billion capital expenditure initiative to build 1 gigawatt (GW) of dedicated AI computing capacity.
The first phase of the deployment will take place at AMI's upcoming data center facility in Hyderabad, India. The initial site is planned for 200 megawatts (MW) of power capacity, with hardware deliveries scheduled to begin in the first quarter of 2027.
Energy Integration and Multi-Region Expansion
AM Intelligence was established by the founders and promoters of Greenko Group, one of India's largest renewable energy producers. The venture intends to co-locate compute clusters with utility-scale clean energy assets to reduce operational electricity expenses and lower the total cost of ownership for high-throughput inference and training workloads.
According to AM Intelligence leadership, the 1 GW expansion roadmap spans four target regions:
- India (Hyderabad): 200 MW initial facility serving domestic sovereign AI foundation models and regional cloud providers.
- United States: High-density facilities targeting North American AI labs and enterprise workloads.
- Finland and Malaysia: Secondary international hubs providing geographical diversity and access to regional power grids.
The company stated that international customers will be able to utilize Hyderabad-based compute capacity over undersea fiber connections, with round-trip latencies of approximately 300 milliseconds for asynchronous batch training and agentic workflows.
Workload Targeting and Hardware Architecture
Nvidia's Vera Rubin architecture incorporates next-generation Vera CPUs paired with Rubin GPUs and high-bandwidth memory, utilizing full liquid cooling across rack-scale configurations. AMI's deployment will target:
- Trillion-parameter model training: Providing the memory bandwidth and inter-node fabric necessary for large-scale mixture-of-experts (MoE) and dense frontier foundation models.
- Agentic AI systems: Supporting multi-step reasoning workloads and long-context inference pipelines that require sustained floating-point throughput.
- Compute-as-a-service: Offering bare-metal and managed compute instances for external cloud platforms and Indian government-backed sovereign AI projects.
AMI plans to scale the Hyderabad campus beyond its initial 200 MW footprint as subsequent Vera Rubin hardware allocations arrive through 2027 and 2028.
