AM Intelligence Ordered 20,000 Nvidia Rubin GPUs
The $4 billion purchase adds 70MW of AI compute capacity to hubs across India and Malaysia.
Updated on Oct. 5, 2026 in Data Centers
Live Poll
Do you believe the rapid expansion of AI infrastructure in your country is a positive development?
AM Intelligence has finalized a $4 billion order for 20,000 Nvidia Rubin GPUs to bolster its AI infrastructure. This capacity expansion, which follows an earlier deployment of 9,000 GPUs in Hyderabad, aims to serve hyperscalers and sovereign AI initiatives.
Why it matters
The investment accelerates the scale of specialized AI compute in South Asia to meet rising demand from cloud providers. It leverages dedicated energy platforms to bridge the gap between current infrastructure and projected national capacity needs.
The purchase adds 70MW of compute capacity, supporting AM Intelligence's plan to bring 300MW online within 15 months. This grows the firm's footprint toward a 1GW target by 2030, set against India's current 1.2GW of total installed data center capacity.
The players
AM Intelligence
A compute provider utilizing a GPU-as-a-rental business model to supply AI capacity to hyperscalers and cloud providers.
Nvidia
The designer and manufacturer of high-performance graphics processing units essential for large-scale AI training and inference.
Greenko Energy Holdings
An energy platform operator managing 50GW of infrastructure, providing the utility-scale power required for dense data centers.
The details
AM Intelligence operates a GPU-as-a-rental model, providing compute power to third-party cloud providers and sovereign entities. The company sustains this infrastructure using energy sourced from a 50GW platform managed by AM Green and Greenko Energy Holdings. This integration ensures the consistent, high-wattage power required for high-density GPU clusters, which perform parallel processing tasks essential for training and running large-scale AI models.
Timeline
March 2024: IndiaAI Mission launched.
June 2025: AM Green acquired 17.5% stake in Greenko Energy.
August 2026: 9,000 GPUs deployed in Hyderabad.
Q2 2027: Projected deadline for AM Intelligence to hold largest compute capacity in India.
2030: Target date for completion of 1GW compute hub.
The Tech Race
The expansion follows the trajectory set by the IndiaAI Mission, which provides the national policy framework for scaling AI infrastructure. The move positions AM Intelligence to command significant market share as India looks to scale its total data center capacity to 10GW by 2030.
This development primarily affects hyperscalers and sovereign AI initiatives requiring high-density, rented compute capacity in the Asian market. Customers should monitor the Q2 2027 milestone to see if the firm meets its target for the largest compute capacity in India.
The takeaway
The move signals a shift toward massive, energy-integrated compute clusters as the primary competitive advantage for data center operators. Watch for the completion of the 300MW capacity milestone over the next 15 months to confirm the deployment speed of this new hardware.
Further reading
For more on the rapid expansion of physical AI infrastructure, read our coverage on Data Centers.
Live Poll
Do you believe the rapid expansion of AI infrastructure in your country is a positive development?
