Preparing For Quantum Computing Means Building Beyond Current Needs

Artificial intelligence (AI) workloads are reshaping the architecture of data centres in India. With demands for higher compute density, lower latency, and greater energy efficiency, operators are moving toward what are being termed ‘AI-grade’ facilities. In an interview with BW Businessworld, Anbu Shanmugam, Chief Operations Officer at Digital Connexion- a joint venture between Reliance, Brookfield Asset Management and Digital Realty, outlined how the sector is adapting to these shifts, the role of sustainability, and what comes next as quantum computing enters the equation.<!--? echo set_value('description') ?-->

AI-grade facilities represent a break from conventional data centre setups. “These facilities are engineered to accommodate rapid escalation in computing power required by AI workloads,” Shanmugam said. Racks in such centres may draw as much as 150 kilowatts, far above standard levels, which requires technologies such as liquid cooling systems, higher-voltage substations and more resilient electrical layouts.

The priority, he explained, is ensuring scalability, reduced latency and consistent performance as enterprises accelerate AI deployments.

GPU-as-a-Service And Wider Accessibility
One trend shaping adoption is GPU-as-a-Service (GPUaaS), which provides businesses with on-demand access to high-performance GPU clusters. “GPUaaS is a powerful enabler for startups, SMEs and MSMEs that may not be able to invest in dedicated infrastructure,” Shanmugam said.

Read More