Why the forecasts differ
Wood Mackenzie projects India's operational capacity growing from 2.2 GW in 2025 to 12 GW by 2030, with AI-dedicated capacity rising from 275 MW to over 6.5 GW. CareEdge Ratings expects about 4 GW by FY2030, and Redseer projects 5–6 GW of installed IT load.
The numbers differ mainly because some count IT load and others grid draw, and they start from different base years. All of them point up sharply.
The global picture
Gartner expects worldwide data-centre spending to exceed $650 billion in 2026, up 31.7% in a year, driven largely by AI infrastructure.
Power is the constraint everyone mentions. Wood Mackenzie calls reliable, cost-competitive power the defining factor in where data centres get built.
What to do with this
More Indian capacity means lower latency and easier data-residency compliance for Indian users. It does not automatically mean cheaper cloud bills; AI demand is competing for the same hardware.
Keep architectures portable, tag and review cloud spend monthly, and avoid paying for idle GPUs. Those habits will matter more, not less, as AI workloads grow.
Sources
Have a project like this in mind?
Talk To Our Team →
