tech · 2026-06-21
India's Data Centre Pipeline Hits 8.33 GW
6fbc1c2f-0755-4fc0-a334-5c04f3376947
India's data centre capacity pipeline has reached 8.33 GW, driven by surging AI and cloud computing demandThe boom reflects growing hyperscaler interest from companies like AWS, Google, and Microsoft expanding in IndiaPower availability and land acquisition remain key bottlenecks for converting pipeline capacity into operational centres
Why is AI driving India's data centre boom?
AI models require massive computing clusters that run 24/7. Training a single large language model like GPT-4 can consume ~50 GWh. Companies like Microsoft and Google are pre-leasing Indian data centre capacity to serve Asia-Pacific customers, pushing India's pipeline from ~2 GW in 2022 to 8.33 GW now.
How do AI workloads differ from cloud?
AI training workloads consume ~5x more power per rack than traditional cloud storage. A single Nvidia H100 GPU cluster draws ~700 kW continuously. Cloud workloads like email or SaaS are lighter, averaging ~10-15 kW per rack. This power intensity is why AI is reshaping data centre design standards.
Why are hyperscalers choosing India?
India offers cheaper land, lower construction costs, and a large domestic market. AWS announced a $12.7Bn India investment through 2030. Proximity to Southeast Asian users cuts latency. India's digital economy, projected at $1Tn by 2028, guarantees local demand that justifies the infrastructure buildout.
What role does GPU supply play here?
Nvidia GPU supply remains constrained globally. Indian data centre operators like Yotta and CtrlS compete with US and Middle Eastern buyers for H100 and B200 chips. A single AI-ready rack needs 8-16 GPUs costing ~$30K each. Chip allocation delays can push project timelines by 6-12 months.
What does 8.33 GW mean for India's grid?
8.33 GW equals roughly the output of 8 large nuclear reactors. If fully built, these centres would consume ~3-4% of India's total installed power capacity. States like Maharashtra and Tamil Nadu face grid strain. Adani Group's 1 GW data centre park in Noida alone requires dedicated power infrastructure.
Which Indian states bear the power burden?
Maharashtra hosts ~45% of India's operational data centre capacity, concentrated in Mumbai and Navi Mumbai. Tamil Nadu and Telangana follow. Uttar Pradesh is emerging, with Adani's 1 GW Noida campus. States offering power subsidies and single-window clearances attract the most investment.
Could this raise electricity costs for homes?
Yes, potentially. In northern Virginia, data centres now consume ~25% of regional power, pushing residential rates up ~8%. India's grid is tighter. If Maharashtra's data centres scale to 3-4 GW, competing demand with industrial users could pressure state electricity regulatory commissions to revise tariffs.
What backup power systems are being built?
Most Indian data centres maintain 48-72 hours of diesel backup. Newer facilities are shifting to lithium-ion battery banks and on-site solar. Reliance's Jamnagar data centre plans include dedicated renewable power. Industry-wide, ~60% of backup still relies on diesel generators, creating emissions concerns.
How does India's capacity compare globally?
India's 8.33 GW pipeline is ~5% of the global data centre capacity pipeline. The US leads with ~40 GW under development. India ranks behind the US, China, and parts of Europe but is the fastest-growing market in Asia, with capacity doubling in roughly 2 years according to JLL's 2026 report.
How far behind is India vs the US?
The US has ~40 GW in pipeline and ~15 GW operational. India has ~1.6 GW operational with 8.33 GW planned. The gap is roughly 10x in operational capacity. India's advantage is growth rate, doubling every ~2 years versus ~3 years for the US market.
Which Indian companies lead this race?
Adani Group leads with plans exceeding 1.5 GW across multiple campuses. Reliance Industries is building capacity in Jamnagar and Mumbai. Among pure-play operators, Yotta (Hiranandani) operates India's largest single campus at 250 MW in Navi Mumbai. NTT, STT, and CtrlS also hold significant market share.
What's India's realistic capacity by 2030?
JLL and CBRE project India reaching ~3-4 GW of operational capacity by 2030, with the full 8.33 GW pipeline taking until 2032-2035. Key constraints are power grid upgrades, water availability for cooling, and environmental clearances. Only ~1 in 3 announced projects historically reach completion on schedule.
Source: thehindubusinessline.com