In June 2025, a regional workshop on energy transition in Malaysia brought together an eclectic mix of governments, international organizations, industry leaders, and academics. The urgent topic on everyone’s mind? How Southeast Asia can meet its surging energy demand in the digital age—an era defined by the exponential growth of data centers and artificial intelligence (AI). According to reporting from the South China Morning Post, consensus at the table was clear: sustainable and alternative energy solutions are needed, and quickly, to keep up with the region’s digital transformation. But, as the workshop focused on energy, one looming issue was largely overlooked: the staggering water consumption tied to this technological boom.
Across Asia, data center capacity is set to nearly double over the next five years, reaching a colossal 25,000 megawatts. The financial stakes are equally impressive, with annual co-location rent for data centers projected to hit US$44 billion in the same period. Massive construction is already underway, and more than 72 percent of this new capacity is concentrated in countries such as Japan, mainland China, Australia, India, and Malaysia. The drivers are familiar to anyone following the tech world: AI workloads are multiplying, cloud adoption is accelerating as businesses and governments go digital, and stricter data sovereignty laws are prompting a surge of local data center development.
But here’s where it gets tricky. The very facilities powering Asia’s digital future are thirsty—very thirsty. Cooling systems prevent data centers from overheating, and these systems are notoriously water-intensive. According to the South China Morning Post, a single one-megawatt data center can gulp down almost 26 million liters of water annually just for cooling. Regionally, the scale is jaw-dropping: cooling water use by data centers in Asia is projected to reach nearly 1 trillion liters in 2025, ballooning to 1.7 trillion liters by 2030. Even more concerning, drinking water accounted for 47.5 percent of data center water consumption in 2024, highlighting a heavy reliance on municipal water grids.
This is happening in a region where water insecurity is already a dire problem. More than 75 percent of Asia is classified as water insecure, and countries home to over 90 percent of the continent’s population are facing imminent shortages. The collision of digital growth and resource scarcity is a recipe for tension, and it’s a challenge that can’t be solved by energy reforms alone.
At the heart of this transformation is a revolution in artificial intelligence. The year 2024 marked a turning point for the industry. As Project Syndicate reports, the age of "bigger-is-better" AI models—where only the richest tech giants could build the most powerful systems—came to a close with the launch of open-source models such as DeepSeek. This democratization of AI development has ushered in a new era, where high-performing models are now emerging from all corners of the globe. Chinese companies like Alibaba (with its Qwen model) and Moonshot AI (with Kimi), as well as Japan’s Sakana AI and the US giant Meta with its Llama program, have all contributed to a rapidly diversifying and competitive landscape.
The shift is profound. No longer is AI innovation the exclusive domain of Silicon Valley or Beijing. Instead, the future lies in tightly integrating AI models with real-world applications. Generic "foundation models" are giving way to systems designed for complex, interdependent tasks—think loan evaluations or production scheduling—where precision, reliability, and context matter more than sheer computational muscle.
This evolution is not just technical; it’s geopolitical. Governments, especially in the US and China, view AI as a national security priority, investing on a scale reminiscent of the Manhattan Project. Yet, as Project Syndicate notes, the quest for "sovereign AI"—reducing reliance on foreign technology while cultivating domestic expertise—has become a global phenomenon. The aim is not just self-sufficiency, but also the projection of soft power: building AI models that others want to adopt, thereby shaping international norms and dependencies.
Trustworthiness and human-centered design are now at the forefront of the AI race. Public acceptance hinges on more than technical prowess; it depends on whether AI systems can be trusted to respect privacy, avoid bias, and enhance—rather than replace—human potential. Concerns about surveillance, coercion, and digital trade deficits are real, and countries that can offer trustworthy, transparent AI stand to win the confidence of the Global South and beyond.
Southeast Asia, in particular, is racing to harness these technologies at scale. According to a new IDC InfoBrief commissioned by UiPath and published in August 2025, the region is experiencing a dramatic shift from AI experimentation to large-scale implementation. Agentic automation—combining robotic process automation (RPA), AI models, and human expertise—has become the backbone of enterprise transformation. In 2025, 75 percent of surveyed organizations in Southeast Asia said agentic AI supports better decision-making, while 72 percent reported increased productivity.
Financial services, manufacturing, and retail are leading the charge, deploying agentic AI for customer support automation, risk management, fraud detection, and productivity enhancement. But the road to widespread adoption is not without obstacles. The IDC InfoBrief highlights risk management (22 percent), a shortage of skilled professionals (18 percent), and high infrastructure costs (18 percent) as major constraints. Data privacy, security vulnerabilities, and the potential for unintended consequences from autonomous AI actions remain persistent worries for business leaders.
To navigate these challenges, the IDC InfoBrief recommends building transparent human-agent ecosystems, robust governance frameworks, and strict compliance with data security and privacy standards. Technology leaders are urged to select agentic tools that integrate seamlessly with existing systems and scale with organizational needs. Meanwhile, policymakers across Southeast Asia are called upon to strengthen regulatory frameworks, establish clear ethical guidelines, and promote public-private partnerships to ensure responsible AI deployment.
Yet, for all the promise of AI and digital transformation, the region’s infrastructure faces a daunting paradox. The very data centers enabling this leap forward are intensifying pressure on already scarce resources—especially water. Without careful planning, the race to the digital future could deepen existing inequalities and environmental strains, rather than alleviate them.
As Southeast Asia stands at this crossroads, the choices made today—about energy, water, AI governance, and ethical design—will ripple for decades. The digital revolution is here, but its true success will depend on whether the region can balance technological ambition with sustainability and trust. The story is far from over, and the next chapter will be written not just by engineers and executives, but by policymakers, communities, and, ultimately, the people whose lives are being transformed by these powerful new tools.
