The energy challenge: Data centers, under pressure from AI, are driving up electricity demand

By Kumar Mitra, Executive Director, Central Asia Pacific & Australia and New Zealand, Lenovo Infrastructure Solutions Group

The reliance on generative AI across industries is increasing globally — and so is the demand for data centers.

Across ASEAN+, 96% of organisations plan to increase AI investments by an average of 15% in 2026, signalling sustained infrastructure expansion to support AI workloads.

As enterprises move from experimentation to execution, the growth of AI workloads is directly accelerating demand for data center capacity across Malaysia and the wider ASEAN region.

AI accelerates the energy challenge

Kumar Mitra (pic) shared that concerns about energy demand are rising in tandem with the rapid expansion of AI-enabled data centers.

As AI adoption scales, infrastructure requirements are shifting. The Lenovo CIO Playbook 2026 highlights that inferencing costs can be up to 15 times higher than training over a model’s lifecycle, with 75% of AI compute expected to be dedicated to inferencing by 2030.

Paving the way to sustainability.

Lenovo executives note that data centers, as critical IT infrastructure, must operate sustainably. Therefore, governments and IT leaders should participate in reducing their carbon footprint by exploring ways to reduce energy consumption and improve energy efficiency in data centers.

The first step towards sustainability in data centers is to adopt innovative cooling solutions. Data centers use approximately 40-50% of their energy for cooling, making efficient cooling systems crucial for environmentally friendly data centers.

Furthermore, as server density increases, traditional air-cooling systems begin to face cooling limitations, while coolants such as water have significantly higher thermal conductivity than air.

Therefore, implementing an advanced liquid-cooling system can significantly reduce energy consumption for heat management and improve energy efficiency compared to using fans or air-cooling systems.

Finding the right balance for growth and innovation.

In addition to developing innovative cooling systems, the adoption of AI-driven computing and energy management infrastructure can further elevate data centers into sustainable, energy-efficient, and environmentally friendly infrastructure.

Leveraging the power of AI to analyze real-time data for resource management, workload distribution, and energy efficiency optimization can enhance operational efficiency while reducing carbon emissions.      

Furthermore, the use of a virtualized server environment reduces the number of physical servers required, resulting in lower energy consumption, reduced cooling system load, and decreased hardware requirements. This also minimizes the need for network and power infrastructure, ultimately lowering operating costs.  

Furthermore, another factor that businesses can consider is investing in environmentally friendly construction materials and the use of renewable energy, which are crucial for building sustainable data center infrastructure in the long term.

As AI investment accelerates across ASEAN+, infrastructure decisions made today will directly influence both operational resilience and long-term sustainability outcomes.

It is clear that adopting advanced cooling technologies, hybrid AI architectures, and AI-driven infrastructure management not only enhances sustainability — it also strengthens the foundation for scalable innovation across Malaysia and the ASEAN region.

Cover photo by Matthew Henry on Unsplash