Developers are rethinking data center design as the AI backlash makes its way to Asia

Thailand’s recent suspension of numerous data center projects points to growing opposition to the AI boom in Asia, following similar concerns in the United States. In September, Thai authorities halted construction of 49 data centers while regulators work on new guidelines concerning resource usage and safety. The AI boom had previously lured Asian governments with incentives like tax breaks, reduced land costs, and easy access to power and water.

Major tech firms such as Microsoft and Amazon have committed billions to develop AI infrastructure across the continent. However, construction delays and capacity challenges are emerging as companies scramble for space, according to Kevin Guan, chief investment officer at Bridge Data Centers. "We're seeing significant delays and challenges" with delivery, Guan stated at the Fortune Leaders Forum in Macau on September 8.

John Taylor, a global critical facilities leader at Gensler, argued that data centers must transition from being viewed as industrial infrastructure to becoming integral parts of the community. In densely populated Asian cities, these facilities are being constructed upwards of ten or eleven stories high and situated near residential zones.

Some companies are even repurposing excess heat generated by data centers for winter swimming pool use. Data centers necessitate large quantities of freshwater for cooling, but operators like Bridge Data Centers in Johor, Malaysia, are implementing wastewater treatment plants to reuse municipal wastewater and reduce strain on local water supplies.

However, the modular nature of hardware within data centers may soon become crucial as AI chips have shorter lifespans than the 20-year lifespan of data center buildings. This mismatch in timelines is prompting designers to devise ways for quick hardware replacements, similar to "Lego brick" components. Despite sustainability efforts, data centers still rely heavily on fossil fuel-generated electricity, accounting for about 56% of worldwide computing energy consumption, according to the International Energy Agency.

This reliance poses significant hurdles for corporate climate commitments. Gensler has pledged to achieve zero carbon emissions by 2030, yet John Taylor admitted that the AI boom presents substantial challenges to attaining this goal.