As artificial intelligence (AI) adoption accelerates globally, the demand for robust infrastructure has reached unprecedented levels. At ITW Asia’s keynote panel, “Scaling Infrastructure for AI Across Asia,” industry leaders delved into the unique challenges and opportunities associated with AI-driven growth across the region.

Moderated by Vivek Dahiya, managing director and head of Cushman & Wakefield’s data centre advisory team for Asia Pacific, the panel featured perspectives from Jeyakumar Janakaraj, CEO of AdaniConnex, Tony Rossabi, co-founder of OCOLO, Charmond Tsang, chief commercial officer of DCCONNECT GLOBAL, and David Dzienciol, chief customer and commercial officer at NextDC.

• Vivek Dahiya, managing director, head, data centre advisory team, Asia Pacific - Cushman & Wakefield (moderator)
• Jeyakumar Janakaraj, CEO - ADANICONNEX
• Tony Rossabi, co-founder and managing member - OCOLO
• Charmond Tsang, chief commercial officer - DCCONNECT GLOBAL
• David Dzienciol, chief customer and commercial officer - NextDC

Opening the panel, Dahiya asked the panel about the current trends and transactions shaping the Asia-Pacific region.

Janakaraj identified India as a pivotal market, attributing its rise to geopolitical tailwinds, technological leapfrogging, and an intrinsic demand amplified by the AI boom. “We’re seeing large-scale infrastructure requirements, from 500 megawatts to gigawatt-scale projects,” he explained.

Highlighting projects like the 1.2 GW IT load site in Navi Mumbai and Hyderabad’s burgeoning data centre market, Janakaraj emphasised that India’s positioning is not only advantageous for domestic needs but also for offering solutions to the broader APAC region.

Rossabi expanded on the discussion by comparing demand patterns in the Asia-Pacific with those in the US and Europe. He noted that the Asia-Pacific region is still in the developmental phase but poised for significant growth.

“If you’d asked me three years ago about growth, I’d have said two or three more years. Now, with AI, we’re looking at six to ten years of extensive growth,” he remarked. Rossabi, though. also flagged the environmental, regulatory, and power availability challenges that could hinder this growth trajectory.

A recurring theme during the panel was the disconnect between demand and delivery timelines. Dzienciol pointed out that the gap between signing contracts and activating capacity can be two to four years or longer. “We’re seeing impatience,” he said, citing an example where a firm deployed 2,000 GPUs and sold only 30% within three months, despite high expectations.

He also described a shift in customer expectations, noting that traditional phased deployments are giving way to immediate demands for large-scale capacity. “Customers now want 10 MW switched on in phase one, followed by another 10 or 20 MW—all to support GPU loads much faster,” Dzienciol explained.

Tsang, representing a carrier perspective, highlighted infrastructure-related hurdles specific to AI data centres. “AI centres require high-quality fibre infrastructure with diverse routes,” he noted. While developed markets like Singapore and Australia meet these needs, other Southeast Asian nations face challenges in providing adequate diversity and redundancy.

“Customers demand four to six diverse routes, but physically, there might be only two bridges or paths available,” he explained, illustrating the constraints that network providers face in meeting expectations.

Addressing power infrastructure, Dzienciol emphasised the complexity of ensuring not just availability but also sustainability. “Price is still a factor, but the type of power—particularly green energy—is critical,” he said. He added that access to GPUs and skilled resources remains a bottleneck, as does long-lead procurement for large-scale projects.

Janakaraj elaborated on India’s integrated approach to energy and infrastructure, noting how projects like AdaniConnex’s renewable energy site—a 30 GW hybrid project spanning 750 square kilometres—underscore the country’s capacity for scale.

However, he acknowledged the inherent risks in planning massive projects, particularly regarding customer commitment and execution timelines. “The interplay of green energy supply, infrastructure, and data centres is complex. The question is, who takes the risk?” Janakaraj posed.

Rossabi touched on regulatory uncertainties, particularly regarding AI’s potential societal impact. “Governments are wary of AI—whether it’s predictive analytics in financial markets or healthcare applications. Regulatory frameworks will likely evolve to address these concerns,” he predicted.

Looking ahead, the panellists shared their visions for the next five years. Dzienciol emphasised the importance of strategic land banking and aligning land acquisitions with power availability.

“It’s not just about buying land; it’s about pairing it with power infrastructure and planning for long-term deployments,” he said. He also underscored the need for a balanced approach between speculative builds and securing customer commitments.

Tsang advocated for more flexible, on-demand network services to accommodate fluctuating AI workloads. “The current on-demand models work for 1G or 10G services, but AI workloads require terabit-scale solutions. We need to develop this capability to meet the industry’s evolving needs,” she stated.

Janakaraj, meanwhile, highlighted the importance of early customer engagement in master planning. “It’s about co-developing infrastructure—from energy to routing to redundancy—to ensure readiness for gigawatt-scale demands,” he said. However, he stressed the need for customers to commit to long-term contracts to enable such large-scale investments.

Throughout the discussion, the panellists emphasised the necessity of collaboration across the ecosystem. From operators and carriers to customers and regulators, achieving AI’s potential in Asia-Pacific will require collective effort.

“This isn’t just about building data centres,” Dzienciol concluded. “It’s about rethinking infrastructure, power, and networks to support a revolution in how we work, live, and innovate.”