By Satya Ramamurthy, Chair, Public Sector, Structured Finance Asia Pacific, SMBC, in collaboration with Ecosperity.
Although the field of Artificial Intelligence (AI) has been evolving for decades, the widespread public interest and rapid development of AI large language models accelerated in 2023 following OpenAI’s November 2022 launch of ChatGPT.
The past two years have seen Big Tech companies and startups alike raising increasingly large sums for AI development and research, with the support of investors also racing to capitalise on what is considered by some as the next industrial revolution.
As AI adoption accelerates, the demand for the physical infrastructure, energy and natural resources necessary to power AI development is also growing at an unprecedented pace. The emergence of AI has fuelled concerns that AI-driven data centre growth could exacerbate the global climate crisis and water scarcity in vulnerable regions.
In their 2025 report Energy and AI, the International Energy Agency highlights that data centres currently account for 0.5% of combustion emissions globally – i.e., indirect emissions resulting from data centre electricity consumption. Over the course of the next decade, this may grow by up to 2.5 times in the Lift-Off Case which assumes stronger growth in AI adoption and represents the largest emissions growth among all sectors.
At the same time, there are high expectations for AI’s potential to open the door to a new generation of technological breakthroughs aimed at reducing energy and water consumption.
In Asia Pacific, data centre growth is amplified by the needs of a young and tech-savvy consumer base, coupled with rapid digitalisation. The expansion of AI and digital services in Southeast Asia is driving significant expansion in data centre infrastructure, with the region’s internet users already exceeding 460 million, and a digital economy expected to nearly triple to US$1 trillion by 2030.
Markets in South-East Asia have benefitted from growing internet penetration, favourable policies, and investments from global tech companies – collectively positioning the region as a dynamic and fast-growing data centre hotspot.
Johor, Malaysia, is a notable example. The state is rapidly becoming an AI data centre hub in the region, with plans to expand data centre capacity from the current estimated 580 MW to 5,800 MW by the end of the decade. Capitalising on abundant land, affordable energy, and water resources, Johor aims to establish itself as a regional digital innovation leader.
Across the causeway, in Singapore, the government plans to increase data centre capacity by more than one-third to meet rising demand driven by digitalisation and AI services. However, the island nation faces challenges from limited land availability and clean energy supply constraints. Singapore’s data centre expansion strategy focuses on balancing its role as a regional digital hub with its sustainability goals.
In Australia, there is close to 1.5 GW of built-out capacity, and this is expected to double in the next decade. Australia’s annual data centre investments reached A$23 billion in 2024 and are projected to grow to A$40 billion by 2028.
At Ecosperity Week 2025, SMBC engaged a cross-section of experts in a roundtable session to help shed light on the key challenges faced by ecosystem stakeholders in balancing AI growth with sustainability objectives in the Asia Pacific. The roundtable also touched upon the potential role of financial institutions in enabling a more sustainable AI future.
Participants in the roundtable session included technology players – such as hyperscalers, data centre operators and semiconductor players – alongside suppliers of clean energy and water, as well as investors and financiers.
AI’s rapid growth poses significant challenges to environmental sustainability, necessitating comprehensive strategies that consider a host of interconnected factors in the drive for a more sustainable AI future:
AI’s rise has significantly increased the demand for data centre financing. According to a study by McKinsey, data centres will require a cumulative US$6.7 trillion in capital expenditure worldwide by 2030 to keep pace with the demand for compute power. Data centres built for AI processing are expected to need US$5.2 trillion in capital expenditures, whereas those supporting traditional IT applications are projected to require US$1.5 trillion.
To help meet this growing demand, financial institutions have a critical role to play in supporting the sustainable and responsible growth of AI and data centres by providing innovative and tailored financing solutions.