Over 50 % of Indian Firms to Build New Data-Centre Capacity in a Year as AI Workloads Grow

 

Over 50 % of Indian Firms to Build New Data-Centre Capacity in a Year as AI Workloads Grow

Mumbai, 12 November 2025 – According to a new report by Cisco, more than half of Indian organisations – 51 % – plan to build new data-centre capacity within the next 12 months to cope with the rapid growth of AI workloads. 

Key findings

• Over 50 % of Indian companies expect their AI workloads to increase by more than 50 % in the next 3–5 years. 

• 91 % of Indian organisations say they are deploying autonomous AI agents; yet only 37 % believe they can properly secure them. 

• Globally, 13 % of organisations that have deployed AI are outperforming peers because they made foundational infrastructure decisions early. Cisco calls them “pacesetters”. 

• These “pacesetters” generally:

  • build “network-first” foundations,

  • prioritise power infrastructure,

  • continuously optimise,

  • embed security from day one. 

• Meanwhile, around 45 % of Indian firms that don’t adopt this architectural foresight risk accumulating what Cisco terms “AI infrastructure debt” – technical debt that could lead to operational, security and competitive risks. 

Why this surge in data-centre capacity?

The backdrop to this push is two-fold:

Firstly, AI workloads are becoming far more demanding. According to industry commentary, traditional data-centres designed for moderate densities are now being challenged by racks with much higher power/heat requirements due to AI-compute workloads. 

Secondly, India’s data-centre market is rapidly expanding. For example, studies suggest the country’s data-centre capacity doubled from ~540 MW in 2019 to ~1,011 MW in 2023, with forecasts indicating an additional ~500 MW of capacity to be added over the next few years, largely driven by AI and cloud demand. 

Thus, the combination of rising AI compute, regulatory push (such as data-localisation), and enterprise digital transformation is accelerating investment in data-centre infrastructure.

Implications for Indian firms

• Opportunity: Firms that act early and build scalable, optimised infrastructure will be better positioned to extract value from AI, gain competitive advantage, and avoid bottlenecks.

• Risk: Those that delay may end up with “infrastructure debt” — needing to retrofit power, cooling, networking, security — which may hamper ROI from AI, increase cost and risk.

• Strategic focus: The built-capacity is not only about compute; Cisco highlights that network, power/energy planning, and security must be conceived from the start, not after deployment.

• Ecosystem effect: Given so many firms are planning new capacity, the data-centre ecosystem (land, power, cooling, networking, skilled operators) must scale accordingly — this offers investment and job opportunities across the value chain.

Challenges ahead

• Power & cooling: AI workloads are power-intensive, and cooling high-density racks is non-trivial. Scaling sustainably will be key.

• Security & compliance: The survey shows only ~37 % of Indian firms believe they can fully secure autonomous AI agents — a gap that becomes more acute with large-scale infrastructure.

• Talent & operational readiness: Building the capacity is one thing, operating it for AI workloads is another. Organisations will need new skills, processes and operational maturity.

• Capex & ROI: While the investment is significant, firms must ensure that infrastructure is aligned with workloads, not over-built or mis-matched — hence the emphasis on architectural foresight.

What this means for the Indian technology landscape

This trend signals that India is entering a new phase of digital infrastructure build-out. AI is not just a use-case but a driver of physical infrastructure expansion. As firms across sectors move AI from pilot to scale, the supporting infrastructure must keep pace.

For India, this could help in several ways:

• Strengthening domestic data-centre supply and reducing dependency on external jurisdictions.

• Enabling local AI compute for Indian workloads (including language, regional content, etc), improving latency and regulatory compliance.

• Attracting global cloud/hyperscale players who may prefer locations with robust local infrastructure.