HÀ NỘI – The race to dominate the artificial intelligence (AI) landscape is heating up, and with it, the demand for robust data center infrastructure. Asia, including Việt Nam, is witnessing a surge in data center development, driven by the explosive growth of generative AI.

Savills Hà Nội reports strong leasing demand in established tier-1 markets like Seoul, Tokyo, Hong Kong, Kuala Lumpur, Singapore, and Bangkok. However, rental value growth is slowing, indicating a late-stage upswing. Meanwhile, tier-2 markets like Johor, Hà Nội, HCM City, and Bengaluru are gaining momentum, experiencing rising rental values and positioned in the early stages of their upswing cycle.

This competitive environment is compressing yields for investment-grade properties in tier-1 locations, averaging around five percent. Emerging markets, however, offer more attractive yields between 9.5 and 10.5 percent, drawing significant investor interest.

A critical challenge facing the industry is the insatiable power demands of AI. Traditional data centers pale in comparison to their AI-driven counterparts, which require two to five times more power. This surge in demand is straining power grids, with some regions facing potential shortages. Ireland's recent moratorium on new data center grid connections in Dublin until 2028 highlights this growing concern. McKinsey predicts global data center capacity demand could reach a staggering 171 to 219 GW by 2030, emphasizing the urgency of addressing power constraints.

"AI-driven data centers require two to five times more power than traditional cloud-based facilities, prompting a fundamental shift in design and site selection approaches," says Thomas Rooney, Associate Director, Industrial Services, Savills Hà Nội. "Việt Nam needs to invest in infrastructure development, ensuring grid power and cybersecurity, raising data awareness and promoting data sharing to capitalize on the transformative power of AI and data centers. A skilled workforce will also be crucial for attracting investment."

Much of Asia's existing data center infrastructure is ill-equipped to handle the demands of AI servers. The majority are co-location facilities with IT loads of 10-40MW, concentrated in mature markets. This underscores the intensifying competition to develop cutting-edge, AI-ready data centers.

The continued rise of AI in 2025 will further strain infrastructure, forcing the market to navigate land availability and power constraints. Southeast Asian nations, recognizing the economic potential of data centers, are implementing national AI strategies, like Singapore's NAIS 2.0. Việt Nam's 2030 strategy also identifies AI as a key driver of the 4th industrial revolution, highlighting its importance for national competitiveness and sustainable economic growth. The coming years will be crucial for the region's ability to balance the promise of AI with the practical challenges of powering its growth.

[Copyright (c) Global Economic Times. All Rights Reserved.]