Sustainability is now the headline, not a footnote

Sanjay Agrawal, Head Presales and CTO at Hitachi Vantara India and SAARC, opines that the conversation is moving beyond headline metrics like PUE toward a broader view of how data lifecycle management and infrastructure efficiency reduce the overall environmental footprint. Let’s see why and how- while also touching upon adjacent (or not-so-adjacent) factors like redundancies, availability and AI-readiness.

What is changing about data centre infrastructure in India- is it resonating with global trends and evolution?

India’s data centre infrastructure is evolving in step with global trends, moving away from siloed, asset-heavy environments toward adaptive, data-driven platforms. The emphasis is no longer just on scale, but on enabling seamless movement of data across core data centres, hybrid cloud and edge environments with consistent performance, governance and availability.

The emphasis is no longer just on scale, but on enabling seamless movement of data across core data centres.

This transition is being driven by the rise of AI-enabled, cloud-native and real-time workloads, which require infrastructure that is software-defined, automated and resilient by design.

Any of your own examples and observations on the industry movements in areas like High-density design choices, capacity planning, power availability and PPAs/Captive models?

Across India’s data centre ecosystem, design and investment decisions are increasingly being shaped by the convergence of AI workloads, energy constraints, and regulatory expectations. These forces are driving a fundamental rethink across infrastructure design, capacity planning, and data governance models. There is a shift to Platform-Based Infrastructure. AI-driven and high-density workloads are pushing organisations to move beyond siloed compute, storage and network layers toward integrated data platforms that scale performance, protection and availability together, rather than as separate technology domains.

We also see Elastic, Consumption-Based Capacity Planning. Infrastructure planning is increasingly focused on flexible, consumption-based models that allow capacity to expand or contract in line with business demand, reducing the risk of rigid, over-provisioned builds.

What about data sovereignty/localisation and integrated stacks?

Localisation requirements are shaping architectures that enable local control of sensitive data with global visibility, supporting AI, analytics and business operations through hybrid integration without compromising compliance.

Overall, the move toward integrated stacks is being driven by the need to reduce operational complexity, strengthen cyber resilience, and ensure continuous access to trusted data across hybrid and distributed environments, especially as AI and edge use cases make data availability and governance increasingly critical for business.

How crucial and intentional is sustainability, especially with water management, cooling mechanisms, energy sources, and PUE?

For many enterprises, sustainability is becoming a core design principle, not an operational afterthought. We believe that modern infrastructure must be high-performance, resilient, and energy-efficient by default, ensuring environmental responsibility is built into the platform rather than added later. This is reflected in the way platforms are designed to optimise power usage through efficient storage architectures, intelligent data placement, and automation. From a Hitachi Vantara perspective, we also notice the conversation moving beyond headline metrics like PUE toward a broader view of how data lifecycle management and infrastructure efficiency reduce the overall environmental footprint.

Water and cooling strategies are also becoming more intentional, aligning infrastructure performance with regional resource realities. As a result, sustainability is increasingly linked to long-term operational viability, regulatory alignment, and brand trust, rather than carbon reporting alone.

Are data centres moving forward well on issues like outages, downtime, redundancies, security loopholes, AI-readiness, etc?

The industry has made meaningful progress in moving from a traditional focus on uptime toward a broader model of resilience and recoverability. Today, leading data centres are designed around multi-layer redundancy, automated failover and cyber-resilient architectures that treat availability, security and data protection as part of a unified platform rather than isolated functions.

AI-readiness is also becoming a foundational design principle. This includes planning for higher power density, advanced cooling, high-throughput networking and data platforms that can support large-scale analytics and AI pipelines without compromising governance or performance. As a result, data centres are increasingly being evaluated not just on how rarely they go down, but on how quickly and cleanly they can recover, adapt and continue operating under disruption.

Is AI compute demand making data centre investments and strategies myopic and one-dimensional? Any worries about AI bubbles and idle infrastructure?

AI compute demand is undeniably shaping investment priorities, but more mature strategies are avoiding one-dimensional build-outs by focusing on flexible, software-defined and multi-workload platforms rather than fixed, AI-only environments. This approach allows the same infrastructure to support AI, cloud, enterprise and edge workloads as demand patterns evolve.

From our experience at Hitachi Vantara, the real risk is not AI itself, but inflexible capacity. To hedge against potential cycles of overbuild or underutilisation, operators and enterprises are increasingly adopting consumption-based models, modular design and integrated data platforms that can be repurposed as workloads shift. In this sense, strategic agility is emerging as the key safeguard against idle infrastructure and market hype, rather than sheer scale.

pratimah@cybermedia.co.in