AI inferencing to drive India data centre growth: Iron Mountain CEO - Business Standard

2026-04-01 · Source: artifical intelligence via Google News · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure · Depth: Fundamental Awareness, quick

Summary

Iron Mountain's Data Centers India CEO, Rajesh Tapadia, predicts that AI inferencing will be the primary driver of data center demand in India, surpassing the growth from large language model (LLM) training. While both global hyperscalers and Indian startups are utilizing India for LLM training, Tapadia emphasizes that the subsequent growth phase will be dominated by AI inference companies as AI applications scale. Iron Mountain is strategically positioning itself to meet this demand by developing 100-500 Megawatt (Mw) capacity data centers, rather than Gigawatt-scale campuses, which aligns with the anticipated higher growth in inference workloads within India. These facilities will also accommodate some domestic LLM training needs from Indian startups.

Key takeaway

For data center investors and operators evaluating market opportunities in India, your strategy should heavily prioritize infrastructure optimized for AI inferencing. The projected surge in inference workloads suggests that smaller, regionally focused data centers (100-500 Mw) will yield better returns than massive Gigawatt-scale campuses, while still accommodating some LLM training needs.

Key insights

AI inferencing, not LLM training, will drive India's data center growth.

Principles

• Inference workloads scale faster than training.
• Regional demand dictates data center capacity.

In practice

• Focus on 100-500 Mw data centers.
• Prioritize inference-optimized infrastructure.

Topics

• AI Inferencing
• India Data Centers
• Iron Mountain
• LLM Training
• Data Center Growth

Best for: VP of Engineering/Data, Director of AI/ML, Entrepreneur, CTO, Executive, Investor

Related on AIssential

• OpenAI taps Tata for 100MW AI data center capacity in India, eyes 1GW — OpenAI is significantly expanding its AI infrastructure and enterprise adoption in India through a major partnership with Tata Group.
• Your Next AI Query May Travel Where the Power Is — Distributing AI inference workloads across micro data centers near substations enhances grid flexibility and power access.
• The Professor of Outputmaxxing — Anjney Midha, AMP — Output maxing and full-stack alignment are crucial for scaling AI responsibly and efficiently, from compute infrastructure to patient empowerment.
• AI Network Infrastructure Growth Creates Up to $500 Million Opportunity for Test and Measurement Providers — Rapid AI network growth creates a substantial market for advanced test and measurement solutions, demanding new validation approaches.
• Canadian pension giant joins race to fund India’s AI-fueled data center boom — India is rapidly becoming a global AI and cloud data center hub, attracting major international and domestic investments despite challenges.
• What if building more grid capacity isn’t the answer? Solving ‘phantom compute’ could address data centre efficiency — "Phantom compute" from junk data bottlenecks AI development; efficient data governance can boost GPU utilization and cut compute demand.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by artifical intelligence via Google News.