Asia Pacific's Data Centres Race Against AI Demand: Is Adaptation Possible?

As artificial intelligence (AI) continues to gain traction across various sectors, the demand for data centers in the Asia Pacific is on an upward trajectory. Traditional facilities, designed for earlier computing generations, are now grappling with the energy consumption and cooling requirements posed by modern AI technologies. By 2030, workloads driven by graphic processing units (GPUs) could elevate rack power densities toward an astounding 1 megawatt (MW), rendering minor upgrades insufficient. Therefore, operators are actively exploring the construction of “AI factory” data centers, meticulously designed for this new era.

To delve deeper into this rapidly evolving landscape, we caught up with Paul Churchill, Vice President of Vertiv Asia, to discuss how the region is gearing up for this transition and what kind of infrastructure adaptations are necessary.

Growth at Lightning Speed

The AI-driven data center market is expected to skyrocket from $236 billion in 2025 to nearly $934 billion by 2030. Such explosive growth is largely fueled by the swift integration of AI technologies within industries like finance, healthcare, and manufacturing. These spheres require ultra-efficient computing environments powered by dense GPU clusters, which demand significantly more energy and cooling than conventional servers.

In Asia Pacific, this surge is further enhanced by government investments in technology, the rollout of 5G networks, and the rise of cloud-native and generative AI applications. “The uptick in computational needs is unprecedented for this region,” Churchill remarked.

However, addressing this soaring demand necessitates more than just expanding existing facilities. It calls for innovative infrastructure strategies that are both scalable and sustainable. Churchill emphasized that leaders in infrastructure must go beyond piecemeal improvements. “A vision for the future must include adopting infrastructure that's optimized for AI, combining high-capacity power systems, advanced thermal management, and integrated, scalable designs,” he explained.

Hot Topics: Cooling and Power

With projections showing rack densities soaring from 40kW to an impressive 130kW—and possibly reaching up to 250kW by 2030—the issues of cooling and power delivery have never been more critical. Traditional air cooling methods are becoming obsolete.

Vertiv is stepping up by developing hybrid cooling systems that cleverly mix direct-to-chip liquid cooling with air-based solutions. This innovative approach can adapt to fluctuating workloads while reducing energy consumption and ensuring reliability. “Our coolant distribution units enable direct-to-chip liquid cooling and maintain reliability, even in high-density settings,” said Churchill.

Redesigning Data Centers for AI

The rise of liquid-cooled GPU pods and high-density racks, as envisioned by tech giants such as AMD, Microsoft, Google, and Meta, indicates a substantial architectural transformation. Rather than merely retrofitting existing facilities, new data centers are being purposefully designed to cater specifically to AI.

Churchill asserts, “The future architecture of data centers will be hybrid; these infrastructures must prioritize liquid flow.” This involves rethinking floor layouts and upgrading coolant distribution and power systems.

The next generation of facilities aims to synchronize cooling, power, and monitoring from the chip level all the way to the grid. For Asia Pacific, where large-scale data centers are rapidly emerging, this kind of integrated design is vital to meet performance expectations and sustainability objectives.

From Upgrades to AI Factory Data Centers

By 2030, Asia Pacific is projected to surpass the United States in data center capacity, achieving nearly 24 GB of commissioned power. To accommodate this growth, enterprises are shifting away from reactive upgrades to comprehensive AI factory data centers.

According to Churchill, this transition should unfold in stages. The first stage is integrated planning, merging power, cooling, and IT management into one cohesive system. This streamlined approach facilitates deployment and creates a strong foundation for scaling.

Next, companies should consider adopting modular and prefabricated systems. These systems allow for phased capacity additions without major interruptions. “Organizations can deploy factory-validated modules alongside their current infrastructure, smoothly transitioning workloads to AI-ready systems without significant disruptions,” Churchill added.

DC Power: Newfound Relevance

Recently, Vertiv has debuted the PowerDirect Rack, a direct current (DC) power shelf aimed at AI and high-performance computing. Transitioning to DC power significantly reduces energy losses, minimizing the number of conversion steps from the grid to the server. This also aligns seamlessly with the increasing use of renewable energy and battery storage in markets like Vietnam and the Philippines.

Churchill asserted that DC power isn’t merely an efficiency move; it’s a crucial strategy for fostering sustainable growth.

Sustainability: An Urgent Priority

As AI accelerates energy consumption, data center operators face mounting regulations and problematic grid constraints, particularly in Southeast Asia where power reliability can be erratic.

Vertiv is collaborating with operators to incorporate alternative energy sources, including lithium-ion batteries and microgrids, that mitigate grid dependence and bolster resilience. Additionally, innovative solar-backed uninterruptible power supply (UPS) systems and advanced energy storage technologies are starting to gain traction, assisting in load balancing and cost management.

Modular Solutions for Quick Scalability

For many emerging economies in Asia Pacific, issues such as limited real estate and inconsistent power supplies pose significant barriers. In this context, modular and prefabricated data center solutions present a pragmatic answer.

These prefabricated modules can shorten deployment timelines by up to 50% while enhancing energy efficiency and scalability. Their design flexibility is crucial for AI workloads that can expand quickly and unpredictably.

By merging compact design with effective operations, modular systems provide a way to establish AI-capable capabilities swiftly and with reduced risks—an indispensable advantage as the digital economies of the region continue to evolve.

Preparing for a High-Pressure Future

The growing prominence of AI is fundamentally altering how data centers are constructed and managed across Asia Pacific. As workloads become more demanding and sustainability imperatives escalate, relying on outdated infrastructure simply won't cut it. The shift toward AI factory data centers, characterized by advanced cooling, DC power, and modular systems, reflects a proactive approach in preparing the region for the next wave of computing demands.