Artificial intelligence is reshaping the data centre industry from a supporting infrastructure layer into a strategic constraint on growth. AI demand accelerates faster than energy systems, water supply, and regulatory frameworks can adapt. Data centre operators face rising costs, higher stranded-asset risk, and increasing resistance from local communities. What was once a predictable asset class is now being redefined by AI-driven volatility.
According to Meiske Sompie, Asia Partner at TBH, the widening gap between AI workload growth and energy infrastructure deployment is the most critical risk facing the sector. AI workloads require far higher rack densities and power consumption, accelerating equipment depreciation and shortening the lifecycle of chips and networking assets. This shift is forcing operators to move away from static, upfront investment models toward full lifecycle planning that allows for modular upgrades and phased capital expenditure.
Construction timelines are also becoming a decisive factor. AI-driven data center projects are highly time-sensitive, often targeting delivery windows of 12 months or less. Delays during construction directly erode profitability, making upfront planning and execution discipline essential to capturing demand in a fast-growing market.
Beyond power, water scarcity has emerged as a structural constraint, particularly across Asia. Regions such as Johor, parts of India, and Peru are already under significant water stress. Thus, prompting authorities in places like Johor Bahru to delay new data centre approvals until an adequate supply can be secured. In response, operators are increasingly investing in on-site water recycling plants and dedicated treatment facilities to reduce community impact and regulatory risk.
Another growing concern is asset obsolescence. Facilities that are not modular or AI-optimized risk becoming stranded within the next decade. Early warning signs include escalating operating costs, outdated cooling systems, and limited flexibility to support higher-density workloads. Technologies such as direct-to-chip liquid cooling, immersion cooling, and closed-loop systems are becoming increasingly central to the long-term viability of data centers.
Key takeaways for operators and investors:
- AI demand is outpacing energy and water infrastructure growth
- Asset lifecycles are shortening, increasing stranded-asset risk
- Modular, lifecycle-based planning is replacing static investment models
- Sustainability and community impact are now board-level priorities
Sompie notes that responsibility for future-proofing data centres is shared across operators, governments, utilities, investors, and regulators. Examples such as Singapore’s green data centre standards demonstrate how coordinated policy can promote sustainability while enabling growth. As AI intensifies pressure on infrastructure, sustainability is no longer just a compliance goal—it has become a core determinant of resilience, financing, and long-term competitiveness.
Source:
https://www.itnews.asia/news/ai-is-forcing-a-redesign-of-the-data-centre-business-model-623112
