AI is getting cheaper to run, but powering its growth is getting pricier
business-standard.com Aug 17, 2026

AI is getting cheaper to run, but powering its growth is getting pricier

AI-summarised brief · reviewed before publication

Artificial intelligence deployment costs are declining, yet the infrastructure required to support its expansion is becoming increasingly expensive. Rising demand for computing capacity, data centres, and electricity is driving up global infrastructure expenditures. Gartner projects worldwide spending on AI-optimised infrastructure as a service will reach $42.3 billion in 2026, a 96.4 per cent increase from 2025. This surge is partly due to cheaper AI tasks encouraging broader corporate adoption across various workflows. India faces significant challenges in managing this growth, with the Ministry of Power projecting an additional 26.3 GW of electricity demand from data centres by 2031-32. Approximately 17 GW of applications have already been submitted to state transmission utilities, while another 9.3 GW awaits formal grid-connection processes. Consequently, while individual AI operations become more affordable, the aggregate strain on national power grids and data centre capacities continues to intensify rapidly.

💡 Why It Matters

  • · The paradox of cheaper AI driving higher infrastructure costs reveals a critical bottleneck in sustainable tech growth.
  • · Nations must urgently upgrade grid infrastructure to prevent power shortages from stifling the very innovation they seek to promote.