India’s data centre industry is in the midst of an unprecedented expansion. Billions of dollars are flowing into AI-ready campuses, hyperscalers are racing to secure capacity, and operators are deploying increasingly dense GPU clusters to meet surging AI demand.
Yet, beneath the headlines about megawatts and investments lies a fundamentally important question: are operators investing enough in research and development (R&D) to address the country’s unique climatic conditions, rather than largely adapting technologies built for vastly different environments?
The answer, industry leaders argue, will determine whether India simply becomes a destination for AI infrastructure or evolves into a global innovator in data centre engineering.
According to the Council on Energy, Environment and Water (CEEW) and SYSTEMIQ, India’s installed data centre capacity has nearly tripled from around 520 MW in 2020 to almost 1.5 GW by mid-2025 and is projected to reach between 4.5 GW and 6.5 GW by 2030.
More than $95 billion has already been committed to the sector since 2019, with investments expected to exceed $100 billion by 2027.
However, the report cautions that scaling capacity alone will not be enough. As AI workloads drive higher power densities and resource consumption, the industry must invest in innovations around cooling, energy management and water efficiency that are tailored to India’s operating conditions.
Despite more than $95 billion committed to India’s data centre industry since 2019, dedicated R&D infrastructure remains scarce. Outside equipment manufacturers such as Daikin, Schneider Electric, Vertiv and Delta, most operators are still focused on deploying capacity rather than building research ecosystems for cooling, power systems and tropical-climate engineering.
Beyond Imported Technology
A recent development also highlights why India is becoming important for data centre R&D. Japanese air-conditioning major Daikin has selected India as the location for its first global R&D centre outside Japan, investing ₹1,000 crore in a facility in Haryana.
The centre will develop next-generation chiller and cooling technologies for data centres, alongside residential and commercial HVAC systems, with innovations intended for global markets.
For Amit Agrawal, President of Techno Digital, the industry’s biggest R&D gap is not centred on a single piece of equipment but on climate-specific systems engineering.
He says India needs a far better understanding of how cooling, power, humidity control, water availability, automation, and materials interact under local operating conditions, rather than relying on equipment specifications developed in temperate markets.
“The industry therefore needs more local test data, full-load validation and digital models based on Indian weather patterns, not just imported equipment specifications,” Agrawal tells AIM. “We need to know how systems perform across a Chennai summer, a coastal monsoon, a northern dust event and prolonged periods of high humidity.”
The next phase, he says, should move from assembling imported technologies to developing Indian engineering intelligence capable of optimising those systems for local conditions.
AI is Changing Infrastructure Design
AI is also forcing the industry to rethink how data centres themselves are designed.
Traditional facilities were built around rack densities of 5–15 kW. Today’s AI clusters are increasingly operating at 30–100 kW per rack, with next-generation deployments expected to go much higher.
Sanjib Seal, Head of Design and Engineering at Aurionpro Solutions, says one of the biggest assumptions becoming obsolete is that every workload requires the same infrastructure architecture.
“AI is fundamentally changing the way data centres are designed,” he tells AIM. “Going forward, infrastructure will be designed around the specific application, whether it is AI training, inference, cloud or enterprise workloads, rather than a one-size-fits-all model.”
He also believes conventional air cooling alone will no longer be sufficient for many AI deployments. Instead, hybrid architectures combining air and liquid cooling will become the norm depending on workload density.
Aurionpro is investing in modular power systems, hybrid liquid cooling and scalable electrical infrastructure that can evolve alongside rapidly changing GPU roadmaps, rather than treating power and cooling as secondary design considerations.
Seal adds that while retrofitting existing facilities remains necessary, legacy infrastructure has inherent limits. Buildings originally designed for low-density workloads often cannot economically support AI racks exceeding 100 kW because of constraints around structural loading, cooling capacity and electrical distribution.
“The industry therefore needs to move beyond isolated technology upgrades,” he says. “AI-ready infrastructure requires integrated engineering, where power, cooling, structural design and digital operations are planned together from the outset.”
The AI era is also changing how operators think about cooling. Agrawal believes the next decade of R&D will be defined by the co-design of power and cooling, orchestrated through intelligent automation rather than by optimising each independently.
Direct liquid cooling will play a central role, but only when integrated with efficient chillers, pumps, coolant distribution, water treatment and intelligent controls designed around local climatic conditions.
At Techno Digital’s Chennai campus, the company is developing a climate-adaptive, build-to-suit architecture capable of supporting air-cooled, hybrid and liquid-cooled deployments.
Sustainability Needs India-Specific R&D
As AI increases both electricity and water consumption, sustainability experts argue that certification alone will not solve the industry's challenges.
P GopalaKrishnan, Managing Director for Southeast Asia and the Middle East at Green Business Certification (GBCI), says certification and research must reinforce one another.
“Certification creates measurable performance benchmarks; R&D develops better ways to achieve and exceed them,” he says.
He points out that India’s operating conditions vary dramatically between coastal humidity, inland heat, dust, monsoons and water-stressed regions, making climate-specific research increasingly important.
“The biggest gap, therefore, is climate-specific performance research at the Indian scale,” he says. “We need more operational data on how cooling, water systems and controls actually perform across India’s climate zones, and greater indigenous R&D built around those findings.”
The CEEW-SYSTEMIQ report echoes these concerns, estimating that nearly three-quarters of India’s data centres are located in water-stressed regions. With AI expected to significantly increase electricity and water consumption, the report calls for integrated planning that simultaneously addresses energy efficiency, water use and carbon emissions rather than treating them as separate engineering challenges.
The Next Frontier Lies Beyond the Data Centre
For Digital Connexion, the industry’s next breakthrough extends beyond cooling technologies within facilities.
In conversation with AIM, Vikram K, Chief Revenue Officer, says enabling sustainable high-density AI infrastructure without overwhelming power networks remains one of the sector’s biggest unresolved engineering challenges.
“The next phase of innovation extends beyond the four walls of the data centre,” he says. “It requires closer integration between AI infrastructure, the electrical grid and regional energy systems to support large AI clusters reliably, efficiently and sustainably.”
The country’s long-term opportunity, Vikram says, lies in developing India-native AI infrastructure that combines climate-adapted cooling, intelligent power management and AI-driven optimisation while reducing dependence on imported technologies.
Seal shares a similar view, arguing that India’s next competitive advantage will depend on engineering innovation rather than capacity expansion alone.
As AI workloads continue to increase electricity demand, reliable power infrastructure will become as important as compute capacity itself.
“Speed-to-market will increasingly become a competitive differentiator,” Seal says, adding that the winners will not necessarily be those announcing the largest campuses but those capable of building, scaling and operating AI infrastructure efficiently over the long term.
Capacity is Not Enough
India has already emerged as one of the world’s fastest-growing data centre markets. But as AI fundamentally reshapes infrastructure requirements, industry leaders increasingly agree that the next phase of competition will not be measured solely in megawatts.
For a country aspiring to become a global AI powerhouse, the opportunity cannot be limited to simply hosting AI infrastructure. It lies in engineering the technologies that make it work, creating climate-adapted, resource-efficient data centres that could ultimately serve as a blueprint for digital infrastructure.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.