India is not short of money for AI. What it is short of is compute. According to the Economic Survey 2025-26, India’s ability to build large AI data centres is not limited by demand or funds alone. It is more constrained by access to GPUs and supporting infrastructure. The survey said that AI should not be seen only as a new tech but more as a strategic priority that possesses the ability to shape India’s infrastructure, labour market, foreign policy and even culture. Yet, when it comes to execution, the country’s AI ambitions are running into a hard physical limit. The Survey flags that, to train advanced AI models, India needs a lot of compute power, mainly GPUs. But, GPUs are becoming more expensive as global demand is on the rise, leading to big players cornering GPU supply. Besides, the supply is also limited due to shortages in high-bandwidth memory and storage chips. This is pushing up costs and creating uncertainty around AI expansion plans. In simple terms, even when money is available, AI projects cannot move forward unless hardware is secured. To understand this better, the Economic Survey ran an agent-based simulation to study how AI compute capacity expands… Read MoreInc42 Media
