TCS HyperVault plans 1GW AI data center campus in Hyderabad

HyperVault’s 1GW campus plan is another reminder that the race to AGI is increasingly bottlenecked by energy, land and interconnect rather than algorithms alone. A gigawatt-class site in Hyderabad, explicitly marketed at frontier AI companies and hyperscalers, effectively adds a new sovereign-scale node to the global compute map alongside clusters in the US, Europe and the Gulf. TCS is not just selling services here, it is positioning itself as a landlord for the largest training and inference jobs of the late 2020s. For the ecosystem, this matters in two ways. First, it deepens India’s role as an AI infrastructure hub, complementing its talent pool and offshoring base with hard assets for model training. That both diversifies geopolitical risk for Western labs and gives Indian regulators new leverage over how those clusters are used. Second, it underlines how capital-intensive the next phase of AI will be: 700 billion rupees in projected spend for a single campus aimed squarely at GPU racks. If projects like this stay on track, they will shorten the waiting time between model generations by making massive training runs logistically feasible rather than power-constrained.