AI Workloads and Hyper‑Scale Cloud Drive Massive Growth in India’s Data Centres

Artificial‑intelligence workloads and the rapid deployment of hyper‑scale cloud platforms are powering a dramatic expansion of data‑centre capacity across India. According to a report published on 2 September, the nation’s total capacity, measured at 1.6 GW in mid‑2026, is forecast to climb to approximately 6 GW by 2029, contingent on an influx of $110 billion in investment.

The first half of 2026 saw demand for 101 MW of new power, a rise of more than 20 % compared with the three‑year average for the same timeframe. Supply kept pace, with 85 MW of fresh capacity added, largely centred in Mumbai and Chennai. Speedy commissioning of pre‑sold space pushed overall vacancy to a historic low of 2.8 %.

Global hyper‑scale operators are reshaping the market dynamics by committing to build about 30 % of the forthcoming capacity through self‑constructed facilities. Under this model, 1.4 GW of self‑built capacity is projected to be operational by 2029, turning Hyderabad and Visakhapatnam into emerging gigawatt‑scale data‑hub cities while Mumbai and Chennai continue to expand their established footprints.

The central government’s recent budget introduced a 20‑year tax holiday for foreign cloud service providers, extended to 31 March 2047. This fiscal incentive has already attracted more than $50 billion of foreign direct investment, reinforcing India’s goal of becoming a global hub for AI and cloud services and mitigating long‑term investment risk.

Capital allocation for the expansion breaks down into $4 billion for real‑estate development, $18 billion for mechanical, electrical and plumbing (MEP) infrastructure, and $88 billion for servers, racks, and related IT equipment. Analysts also note that the rollout of small modular reactors (SMRs) by 2033 could provide off‑grid electricity for large data‑centre campuses, offering a potential solution to future power‑supply challenges.