AM Intelligence — an Indian AI infrastructure player tied to Greenko Group — just placed an order for 9,000 $NVDA Vera Rubin NVL72 rack-scale systems. These will anchor their Hyderabad AI factory, aiming for early Asian deployment of Nvidia's next-gen platform.
Delivery is slated for Q1 2027. The first 30 MW phase will support trillion-parameter models and agentic AI workloads, with inference costs potentially dropping 10x versus the prior Blackwell generation.
This is the opening move in an $8B plan for 1GW of compute-as-a-service infrastructure spanning India, the US, Finland, and Malaysia — serving cloud providers, AI labs, and local LLM developers.
The scale and timing matter. If they execute, this becomes one of the first large-scale Vera Rubin deployments outside the US, and a signal that AI infrastructure spend is accelerating in emerging markets. Watch how this order influences $NVDA's forward guidance and whether other regional players follow suit.
Delivery is slated for Q1 2027. The first 30 MW phase will support trillion-parameter models and agentic AI workloads, with inference costs potentially dropping 10x versus the prior Blackwell generation.
This is the opening move in an $8B plan for 1GW of compute-as-a-service infrastructure spanning India, the US, Finland, and Malaysia — serving cloud providers, AI labs, and local LLM developers.
The scale and timing matter. If they execute, this becomes one of the first large-scale Vera Rubin deployments outside the US, and a signal that AI infrastructure spend is accelerating in emerging markets. Watch how this order influences $NVDA's forward guidance and whether other regional players follow suit.
