AM Intelligence Orders 20,000 More NVIDIA Rubin GPUs for India AI Buildout
AM Intelligence has placed an order for roughly 20,000 additional NVIDIA Vera Rubin GPUs to support AI infrastructure in India and Malaysia. The company's total committed Rubin capacity is now about 29,000 GPUs, alongside a 400 MW development pipeline. Hyderabad remains a key location in its Indian AI expansion plan.
AM Intelligence, with backing from individuals associated with Greenko, has added roughly 20,000 NVIDIA Vera Rubin GPUs to its plans for India and Malaysia. Delivery is expected in a 2027 quarter. Combined with an earlier 9,000-GPU Hyderabad commitment from August 2026, its Rubin total reaches about 29,000.
The new systems will use NVIDIA Vera Rubin NVL72 rack-scale designs with networking, storage, and liquid cooling. They are expected to add about 70 MW across both countries, lifting committed capacity to nearly 100 MW within a 400 MW development pipeline. Hyderabad remains a focal point.
This buildout could affect AI developers, enterprises, researchers, and public programs in India and Malaysia by making large-scale computing more available through service models. Hyderabad's data-centre ecosystem may see added demand for power, cooling, land, and skilled workers. Smaller organizations might gain access to resources that are otherwise costly, though actual benefits may depend on pricing, energy capacity, and how widely these services are offered.