The data-centre efficiency crisis no one is talking about

For all the excitement surrounding liquid cooling, AI-driven optimisation, and next-generation chillers, the fastest path to improved power usage effectiveness (PUE) rarely begins with a large capital project.

In my experience across data centres in India and the wider Asia-Pacific region, efficiency almost always breaks down much earlier, inside the rack itself. This is the industry’s blind spot. As India transitions into an era of high-density computing, it is also becoming one of its biggest risks.

According to CBRE’s India Data Centre Market in a New Era – A 2025 YTD Market Update report, India’s operational data centre stock reached nearly 1,530 MW (translating to 23 million sq. ft.), as of 9M (January-September) 2025, with 260 MW of new supply added during the year.

Operators spend months evaluating new cooling technologies or planning infrastructure expansions. However, many facilities still lose cold air through gaps in the rack base, around rails, or through unsealed cable openings. I’ve walked into halls where the cooling plant was running at peak load, not because the design was wrong, but because airflow discipline inside the rack had never been prioritised.

In a market as fast-growing as India, this is understandable. Facilities expand in phases, new racks coexist with legacy ones, and density demands shift with every new AI or cloud deployment. However, this is exactly why airflow becomes strategic: small leakages compound at scale, and the energy penalty follows for years.

What high-performing data centres do differently

There is a consistent pattern in facilities that operate with strong efficiency and predictable thermal behaviour. They treat airflow management as a discipline, not a maintenance task.

1. They seal the rack like it actually matters: Blanking panels are not optional. Gaps around 19-inch rails are treated as energy leaks, not cosmetic imperfections.
2. They don’t ignore the base: The base-to-floor interface is one of the biggest sources of cold air loss, yet often the least addressed.
3. They replace brush strips with airtight cable seals: Brush strips are convenient. Airtight seals are efficient. The difference in leakage can be dramatic.
4. They address the “silent gaps” between adjacent cabinets: Those narrow spaces become channels for hot air to loop back into cold aisles if not sealed properly.
5. They use partial containment intelligently: You don’t always need full aisle containment. Even lightweight interventions create stability and allow higher set points.

Efficiency is not achieved through any single measure; it is achieved through a mindset that every cubic foot of cold air must be controlled, directed, and utilised.

Why this matters more now than ever

India is entering a new phase of compute intensity. According to a report by Modor Intelligence, hyperscalers led with 42.6% of India’s data centre cooling market share in 2024 and are expanding at a 26.5% CAGR by 2030. Tier III facilities commanded a 63.4% share of the India data centre cooling market size in 2024, while Tier IV facilities are projected to grow at a 27.3% CAGR through 2030.

Air-based solutions held 67.2% revenue share in 2024; liquid solutions represented the fastest growth at 27.7% CAGR, equipment accounted for an 81.2% share of the India data centre cooling market size in 2024, and services are advancing at a 26.8% CAGR.

AI, analytics, and high-density cloud workloads are transforming thermal profiles inside the rack. Operators who once planned for 6–8 kW per rack are now seeing sustained loads double or triple that.

If airflow is not managed, operators compensate the only way they can: lowering temperature set points, increasing fan speeds, or overprovisioning cooling.

Each of these choices worsens PUE and inflates operational cost even before considering sustainability targets that are becoming non-negotiable across the region.

In contrast, facilities that improve airflow control unlock three strategic benefits:

• Higher safe temperature set points, without compromising reliability
• More consistent intake temperatures, reducing thermal stress
• Ability to support higher rack densities without incremental cooling capex

This is especially valuable in India, where real estate and power availability are becoming constraints in high-demand markets.

Measurable impact on PUE and capacity planning

Operators often view airflow optimisation as a comfort improvement rather than a strategic lever. However, field deployments show strong economic outcomes:

• Energy savings: Often several lakh rupees annually per data hall, simply by reducing cooling load.
• ROI: Typically under 18 months, since interventions are inexpensive and non-disruptive.
• Thermal stability: More consistent intake temperatures reduce the risk of thermal stress.
• Raising set points: Better airflow allows operators to safely raise temperatures, improving PUE without compromising reliability.
• Reclaimed capacity: Facilities reporting “cooling bottlenecks” often find they can support higher rack densities after airflow optimisation, without new cooling infrastructure.

In a market as cost-sensitive and growth-driven as India’s, the ability to increase density without expanding footprint or power draw is a competitive advantage.

Sustainability begins with the fundamentals

Sustainability is often framed as a technology challenge. But it is equally a discipline challenge. Airflow optimisation does not require new cooling plants, complex retrofits, or operational downtime. It requires intent and a recognition that efficiency starts far closer to the server than most people assume.

If India is to build the next generation of sustainable, high-density digital infrastructure, we must stop treating airflow as hygiene and start treating it as strategy. Because the takeaway is simple: The future of PUE doesn’t start with the cooling plant. It starts inside the rack.

Sanjay Motwani is the Vice President – Asia Pacific at Legrand Data Center Solutions