India's data centre sector experienced a breakout year in 2025, marking a pivotal shift in the country's digital infrastructure landscape. Capacity has surged more than threefold, growing from approximately 0.5 gigawatts (GW) in 2020 to an estimated 1.6 GW in 2025. This rapid expansion has been fuelled by a dramatic 30-fold increase in data traffic since 2017, the addition of 900 million internet users, and the accelerating adoption of artificial intelligence (AI). Notably, India has emerged as OpenAI's second-largest market, underscoring the nation's rising prominence in global AI consumption.

The year saw a wave of mega-investments totalling around $63 billion in announcements, with Reliance Industries leading the charge by unveiling the world's largest data center in Jamnagar, including a 1 GW AI facility developed in partnership with Nvidia. Other significant commitments included TCS launching its HyperVault initiative targeting 1–1.2 GW of AI-grade capacity over five years, and OpenAI pledging 1 GW amid data localization mandates under the DPDP Act. Hyperscalers leased nearly 98 MW in the first half of 2025 alone—a 48% year-on-year increase—with Mumbai dominating the market, holding about 50% share and near-full occupancy rates.

Projections indicate India's data center capacity could increase rapidly—potentially reaching 5–9 GW or more by 2030, depending on analyst estimates—with additional capex requirements of around $30 billion for incremental growth. Non-IT costs (such as cooling, electricals, and construction) account for roughly 45% of total expenses, creating opportunities for domestic players in EPC (L&T), backup power (Cummins), cooling systems (Voltas), and wiring (Polycab). Real estate developers, electrical and power system providers, telecom operators, and IT firms are all poised to benefit from this boom.

Chairman and Co-Founder of a data centre firm described 2025 as a "wake-up call" for India, which is catching up on both conventional cloud capacity (currently around 1.5 GW) and the AI surge. He estimated that scaling conventional cloud alone to 3 GW by 2030 could require $40 billion in total capex (including servers and IT equipment), with AI workloads pushing costs significantly higher—potentially $70–80 billion overall to reach 3–5 GW. He cautioned that many large announcements represent commitments rather than immediate build-outs, and emphasized India's lower risk of overbuild compared to saturated markets like the US.

Another entrepreneur in this area highlighted India's low capex and opex advantages globally, driven by data localization policies initiated by the Prime Minister. He argued that even if AI compute becomes cheaper, demand for data storage will only increase as more users adopt the technology. He noted that the next 3–4 years offer strong growth with limited downside risk, given India's low starting base and locked-in demand. Colocation typically yields 6–7year paybacks (13–14% ROI), while cloud services offer faster returns (3–4 years) but require customer visibility. He suggested investors focus on sector-specific data centres, Indian cloud providers, or specialized MEP/electrical suppliers. Both guests expressed confidence in sustained momentum, with Amit noting Anant Raj is well-prepared for future competition.

As 2025 draws to a close, India's data centre boom stands as one of the most compelling multi-year investment themes in the digital economy—balancing explosive growth potential with the need for sustainable power and execution. The sector's trajectory will remain a key focus for investors and policymakers in the years ahead.