Q1. Could you start by giving us a brief overview of your professional background, particularly focusing on your expertise in the industry?  

I am a seasoned executive with more than 25 years of experience in data center infrastructure operations and enterprise delivery. Currently serving as President – Global Data Center Operations, Enterprise Delivery & Program Management at CtrlS Datacenters, I oversee the management of India’s largest mission-critical campuses, driving operational excellence, scalability, and innovation for enterprise and cloud ecosystems.

My expertise spans:

Data center operations & risk management – ensuring resilience, compliance, and uptime across large-scale facilities.

Cloud & colocation infrastructure delivery – leading enterprise deployments and hybrid cloud integrations.

AI-ready infrastructure – enabling GPU-based deployments to support next-generation workloads.

Leadership in industry forums – active contributor to DatacenterDynamics and NASSCOM Special Interests Group for Data Centers, shaping thought leadership in sustainability and digital transformation.

Previously, I held senior leadership roles at Microsoft Corp., Tata Communications, and Colt Data Centre Services, where I managed critical infrastructure supporting cloud adoption and enterprise growth. My career reflects a unique blend of technical depth, operational leadership, and strategic vision, positioning me as a key voice in India’s evolving data center ecosystem.

Q2. The new ₹2,000Cr Safe Harbor threshold for IT/GCCs is live. Are you seeing an immediate influx of "High-IP" workloads (model training) into DCs, or are GCCs still keeping the "brain" of the AI in the US?

I am seeing early movement of AI workloads into India, especially inference and export-oriented compute, but the “brain” of AI (model training) is still predominantly US-based. The Safe Harbor threshold and tax incentives are strong levers, but it will take a few years before GCCs fully entrust their most IP-sensitive workloads to India.  

The recently launched DPDP policy and its implementation would improve this equation, which would eventually favour India.

Q3. Microsoft’s India South Central region is set to go live in mid-2026 as its largest hyperscale site. From your vantage point, will this massive capacity injection create a temporary colocation glut in Hyderabad, or is the $100B trade-driven demand for AI inferencing already outpacing this new supply?

From what I understand from publicly available information, this capacity will be used to provide Indian and global customers with a variety of products that Microsoft is best at, like additional capabilities of Co-Pilot, etc. This capacity would be consumed in no time, and additional large capacity shall be needed, as for any other hyperscale data center operator in India.

Q4. How does the shift from 15kW racks (traditional cloud) to 100kW+ racks (AI Factory) impact the terminal value of legacy Tier-3 data centers in India? Are they becoming "stranded assets"?

100kw racks are still being explored, and it would take some time to make it a generalised specification for a large cloud player or within the Enterprise customer community. During this period, a typical Tier-3 DC will continue to serve the much-needed demand of 6-10KVA customers. Having said that, multiple providers are retrofitting the DCs to suit higher capacities in India. Many are keeping the design "flexible" enough to suit a certain customer base or to customize for a particular anchor customer.

Q5. With the $90B total DC investment committed to India, according to you, where is the most critical supply chain "choke point" today? And why?

Major metros (e.g., Mumbai, Chennai) already show peak demand strain. Without grid reinforcement and distributed generation/storage, operators may resort to expensive backup systems (such as diesel generators and batteries), increasing OPEX and the carbon impact. Some additional key bottlenecks would be:

• Shortages and long lead times for AI accelerators (GPUs, TPUs, etc.)
• Limited local ecosystem for high-density switchgear, PDUs, and power electronics that meet hyperscale specs.

Q6. Beyond power and chips, is the inter-campus optical fiber density in India sufficient to support the massive east-west traffic of distributed AI training?

India’s overall optical fibre deployment has been expanding due to 5G backhaul, FTTH growth, and data centre demand — industry estimates indicate significant fibre installation in the coming years. Neutral interconnection and DCI services (e.g., provided by DE-CIX India across Mumbai, Delhi, Chennai, Hyderabad, and Kolkata) are beginning to provide high-speed optical links between major data centre clusters. These platforms are foundational for the "east–west" traffic. Across regions east–west and north–south: The optical backbone is not yet fully sufficient for massive, distributed AI workloads — it is in "rapid build-out" mode, but could become a bottleneck unless capacity upgrades and transport-level investments accelerate.

Q7. If you were an investor looking at companies within the space, what critical question would you pose to their senior management?

Some of the key questions I would ask:

• Is your cooling model scalable to 100 kW+ without becoming water- or regulation-constrained?    
• Are AI workloads improving your EBITDA/MW — or just inflating your capex/MW?  
• Do you control your DCI economics, or are you dependent on third-party long-haul providers?  
• Do you design in-house or rely on 3rd-party design houses/engineering firms?