AM Intelligence has placed a binding order for approximately 9,000 NVIDIA Vera Rubin GPUs for its AI factory in Hyderabad [1].
The investment signals a massive shift in regional compute capacity, positioning India as a primary hub for frontier AI development. By deploying the Vera Rubin platform, AM Intelligence aims to provide the raw processing power required for the next generation of large-scale artificial intelligence models.
The GPUs are scheduled for delivery in the first quarter of 2027 [4]. This hardware will form the backbone of a computing cluster that is intended to be one of the first of its kind in Asia [2].
Financial commitments for the project are substantial. The company plans a total capital expenditure of over $8 billion [1]. This funding supports the build-out of a facility designed for extreme scale, though reports on initial power capacity vary. Business Standard said the initial capacity is 200 MW [1], while the Times of India said the facility will start at 30 MW [2].
Regardless of the starting phase, the long-term roadmap is expansive. AM Intelligence targets a full build-out of 1 GW of compute power [2]. This expansion is expected to reach a target of 450 exaFLOPS [2]—a level of performance designed to handle the most demanding AI workloads globally.
The Hyderabad facility is expected to attract significant interest from developers and enterprises seeking high-performance computing without relying entirely on U.S. or Chinese data centers. The project represents a strategic bet on the growing demand for localized, high-capacity AI infrastructure.
“AM Intelligence has placed a binding order for approximately 9,000 NVIDIA Vera Rubin GPUs”
This move indicates a strategic effort to decentralize frontier AI compute away from traditional tech hubs. By securing a massive allocation of NVIDIA's latest hardware and committing billions in capital, AM Intelligence is attempting to bridge the 'compute gap' in Asia. The scale of the 1 GW target suggests the facility is intended not just for local use, but as a global destination for training massive foundation models.
