India is looking to establish itself as a major global hub for AI infrastructure, and while companies such as Sarvam are contributing to that goal, AM Intelligence is making a much larger investment in the computing infrastructure required for next-generation AI.
According to a Bloomberg report, AM Intelligence has placed a binding order for around 9,000 Nvidia Vera Rubin systems, making it one of the earliest companies in Asia to adopt Nvidia's latest AI computing platform.
AM Intelligence is backed by the founders of renewable energy company Greenko Group and plans to install the Nvidia systems at its first AI data centre in Hyderabad. The servers are expected to become operational next year and will form the initial stage of the company's planned 1GW compute-as-a-service platform, which will span India, the US, Finland and Malaysia.
Hyderabad data centre to support advanced AI
The Hyderabad facility is being designed to handle highly demanding AI workloads, including trillion-parameter models and advanced agentic AI applications.
The company plans to initially offer 200MW of AI computing capacity as part of a larger project that is expected to involve more than $8 billion in capital expenditure.
The facility is expected to provide computing infrastructure to major cloud-service providers, AI research labs and organisations developing Indian AI models. According to the report, the initial capacity has already been purchased by an unnamed US customer.
AM Intelligence intends to make its computing resources available to customers across India, the US, Finland and Malaysia. The company says India's undersea cable infrastructure could allow US customers to access AI computing resources hosted in India with latency of roughly 300 milliseconds.
Renewable energy to help control AI costs
Energy is a central part of the company's strategy. AM Intelligence has access to Greenko's renewable-energy infrastructure, which could help it keep the cost of AI computing competitive.
Greenko founder and president Mahesh Kolli said energy prices have a major impact on the economics of AI tokens and computing. AM Intelligence also claims that its infrastructure could rank among the lowest-cost AI computing platforms globally.
The company plans to expand well beyond Hyderabad. Its long-term target is to reach 5GW of AI data-centre capacity by 2030, with facilities spread across India, Europe and other international markets.
The expansion is expected to be financed through a combination of debt and equity, positioning AM Intelligence as a potentially significant player in the global race to build AI computing infrastructure.
A useful next step would be to compare AM Intelligence's planned 1GW platform with India's other major AI infrastructure projects.
