Archive · 26 Aug 2026
AM Intelligence Orders 9,000 NVIDIA Rubin GPUs for Hyderabad
AM Intelligence wants to build a very large AI computer center in Hyderabad, India.
It ordered about 9,000 new NVIDIA Rubin GPUs for the project.
The center is planned to have 30 megawatts of computing capacity and is scheduled for deployment in early 2027.
Its computers are designed to run extremely large AI models, including models with a trillion parameters.
AMI says the facility could provide about 450 exaFLOPS of inference computing power.
A US customer has already purchased the first capacity, but the customer’s name was not disclosed.
AMI says low-cost renewable energy from Greenko will help keep operating costs competitive.
The Hyderabad site is part of a larger plan for 1 gigawatt of compute-as-a-service capacity across several countries.
AMI is also developing up to 5 gigawatts of AI data-center capacity and plans to fund its expansion through debt and equity.
AM Intelligence placed a binding order for about 9,000 NVIDIA Rubin GPUs for a 30-megawatt Hyderabad AI factory.
The facility is scheduled for deployment in the first quarter of 2027 and would be among Asia’s first large-scale Vera Rubin deployments.
AMI expects the cluster to deliver about 450 exaFLOPS of NVFP4 inference compute for trillion-parameter models and agentic AI applications.
The Hyderabad project is the first phase of a planned 1-gigawatt compute-as-a-service platform across India, the United States, Finland and Malaysia.
AMI plans to bring 200 megawatts online initially, requiring more than $8 billion in capital expenditure and targeting customers including hyperscalers and AI labs.
- Who
- AM Intelligence, promoted by Greenko Group and chaired by Anil Chalamalasetty, is developing the project.
- What
- AMI placed a binding order for about 9,000 NVIDIA Rubin GPUs for a planned 30-megawatt AI factory.
- When
- Deployment is scheduled for the first quarter of 2027; AMI plans to bring 200 megawatts of initial capacity to market in the near term.
- Where
- The first facility will be in Hyderabad, India, within a broader platform spanning India, the United States, Finland and Malaysia.
- Why
- AMI aims to provide large-scale AI computing by combining renewable-energy infrastructure, high-density data centers and advanced NVIDIA hardware.
- GPU order
- About 9,000 NVIDIA Rubin GPUs
- Hyderabad facility
- 30 megawatts
- Scheduled deployment
- First quarter of 2027
- Expected compute output
- About 450 exaFLOPS of NVFP4 inference compute
- Planned compute platform
- 1 gigawatt of compute-as-a-service capacity
- Initial capacity target
- 200 megawatts
- Planned capital expenditure
- More than $8 billion
