AM Intelligence expands NVIDIA Vera Rubin AI infrastructure across India and Malaysia

AM Intelligence (AMI) has announced two further firm and binding orders for 20,000 NVIDIA Rubin GPUs, expanding its planned AI infrastructure footprint across India and Malaysia.

The GPUs will be deployed as NVIDIA Vera Rubin NVL72 rack-scale systems, with the additional capacity expected to total approximately 70 megawatts and be delivered in the second quarter of 2027.

The orders build on AMI's initial commitment for 9,000 NVIDIA Rubin GPUs at its first AI factory in Hyderabad announced in August. Together, the three orders bring AMI's committed capacity to approximately 29,000 Rubin GPUs and close to 100 MW.

Expanding AI compute across India and Malaysia

The latest orders are part of AMI's strategy to develop large-scale Compute-as-a-Service infrastructure for cloud providers, enterprises, AI developers and sovereign AI initiatives.

The new facilities in India and Malaysia are planned to follow the same architecture standard as AMI's Hyderabad AI factory, combining NVIDIA Vera Rubin systems with high-throughput networking, advanced storage and liquid cooling.

The company says the infrastructure is being designed to support successive generations of AI silicon without requiring the physical footprint to be repeatedly rebuilt.

Building around NVIDIA Vera Rubin

AMI's infrastructure will use NVIDIA Vera Rubin NVL72 rack-scale systems, designed for demanding AI training and inference workloads.

The architecture will incorporate high-throughput RDMA over Converged Ethernet networking to support fast data movement, alongside advanced storage systems.

For AMI, the objective is to create infrastructure capable of supporting the increasing compute requirements associated with frontier AI models while maintaining flexibility as future generations of processors become available.

NVIDIA's Asia South managing director Vishal Dhupar said AMI's approach combines accelerated computing with high-performance networking and liquid-cooled infrastructure to support larger AI workloads.

From Hyderabad to a wider regional footprint

The new orders extend AMI's initial Hyderabad deployment, which is expected to be among the early large-scale NVIDIA Vera Rubin deployments in Asia.

The Hyderabad facility is the first committed tranche of AMI's broader Compute-as-a-Service strategy. The company previously announced a binding order for 9,000 Rubin GPUs for the site, with delivery scheduled for the first quarter of 2027.

With the latest commitments, AMI is expanding the model beyond a single AI factory and into a broader regional infrastructure network.

Aiming for 400 MW of global compute capacity

AMI says it has developed a global pipeline of approximately 400 MW of Compute-as-a-Service capacity across India, the United States, Europe and Malaysia.

The three GPU orders announced so far represent the first committed tranche of that pipeline, with additional capacity under development for subsequent phases through 2027 and early 2028.

The company plans to bring another 300 MW of capacity to market over the next 15 months, alongside more than $20 billion in additional capital expenditure, on top of the $6 billion already committed for its initial 100 MW programme.

Connecting energy and AI compute

AMI's infrastructure strategy is closely linked to the wider energy capabilities of the Greenko founders-backed group.

The company is positioning its platform around the integration of energy infrastructure, data-centre capacity and AI compute, with the aim of improving the economics of running high-density AI workloads.

Its planned customer base includes hyperscalers, neocloud providers, sovereign AI initiatives, frontier AI labs, AI-native companies and local developer communities.

The model reflects a growing focus on the infrastructure required to support AI training and inference at scale, particularly as demand for GPU capacity expands.

Toward gigawatt-scale AI infrastructure

Beyond the approximately 400 MW near-term pipeline, AMI says it is targeting 1 GW of Compute-as-a-Service capacity as part of a broader programme to develop 5 GW of powered AI data centres across India, the United States and Europe.

The company is building its facilities around high-density, liquid-cooled infrastructure designed to accommodate successive generations of AI hardware.

For India and the wider region, the expansion adds another significant private-sector investment in AI compute capacity as enterprises, developers and governments seek greater access to advanced infrastructure.

About AM Intelligence

AM Intelligence is an AI infrastructure platform established by the promoters of Greenko. The company is developing infrastructure spanning energy, data centres, AI hardware and compute services, with plans for gigawatt-scale AI capacity across India, the United States and Europe. Its strategy includes Compute-as-a-Service infrastructure for hyperscalers, enterprises, frontier AI labs, sovereign AI initiatives and developers.

Source and Credits

AM Intelligence official announcement