AM Intelligence, an Indian AI infrastructure company backed by Greenko Group, has placed a binding order for approximately 9,000 Nvidia Vera Rubin NVL72 rack-scale systems in a deal valued at roughly $8 billion in capital expenditure. The Hyderabad-based firm announced the order yesterday, targeting delivery in the first quarter of 2027 and positioning itself among the earliest adopters of Nvidia latest AI compute platform in Asia. The purchase ranks as one of the largest single orders for Nvidia Vera Rubin architecture announced to date.
The systems will be deployed at a new facility in southern India engineered to deliver approximately 450 exaFLOPS of NVFP4 inference compute, backed by Greenko renewable energy portfolio. The initial buildout targets 200 megawatts of capacity, with a stated long-term plan to scale toward one gigawatt of compute-as-a-service spanning India, the United States, Finland and Malaysia. The company said a US customer has already reserved initial capacity at the Hyderabad site.
Nvidia unveiled the Vera Rubin platform at the GPU Technology Conference in March 2026, naming it after astronomer Vera Rubin. The NVL72 configuration bundles 72 graphics processing units in a single rack-scale system designed for both training and inference workloads. Supplies of high-end Nvidia compute clusters remain tight as demand from cloud providers, AI labs and sovereign computing programs outpaces available production capacity. Securing 9,000 units represents a significant allocation in a supply-constrained market.
The order highlights the rapid growth of AI infrastructure investment across India, where companies and government bodies are racing to build domestic compute capacity to reduce dependence on overseas data centers. Indian firms including Reliance Jio and Tata Group have already announced large Nvidia-based deployments. The Indian government has signaled interest in homegrown AI model development, further fueling demand for frontier compute hardware. AM Intelligence said its customer base includes major cloud-service providers, AI laboratories and organizations building Indian-language AI models.
Greenko Group is one of India largest renewable energy producers, operating wind and solar assets across multiple states. AM Intelligence parent connection gives the company access to low-cost clean power, a critical cost driver in AI infrastructure where electricity represents a significant portion of operating expenses. Deepak Kolli, AM Intelligence founder, said the company energy advantage translates directly into lower inference costs. In global token economics, energy prices play a huge part, he said. The company plans to fund the buildout with a mix of debt and equity.
The deal comes as Nvidia has cemented its position as the dominant supplier of AI compute hardware, with demand surging across every region. The company data center revenue has grown rapidly as enterprises and governments invest in AI capabilities. Nvidia Vera Rubin succeeds the Blackwell architecture and is manufactured using TSMC advanced process nodes, delivering substantial performance gains over its predecessor. The tight supply environment means that large, committed orders such as AM Intelligence are increasingly rare and signal serious financial backing and customer demand.
India AI market is expanding quickly, driven by a large pool of engineering talent, growing startup ecosystems and government initiatives to build sovereign AI capabilities. The country relatively low energy costs, particularly from renewable sources, make it an attractive location for power-hungry AI workloads. Several global technology firms have announced plans to expand data center operations in India, and the government has introduced incentives for domestic AI compute infrastructure. Industry analysts say India could emerge as one of the top five markets for AI compute capacity within the next three years.
AM Intelligence order also underscores the global shift toward large-scale, purpose-built AI infrastructure. As AI models grow in size and complexity, the compute required to train and serve them continues to increase dramatically. Companies that can secure reliable access to cutting-edge hardware while keeping energy costs low are positioned to capture a growing share of the AI services market. The announcement signals that the AI infrastructure boom is no longer confined to a handful of US and Chinese hyperscalers, but is broadening to include well-capitalized players across South Asia and beyond.
discussion