The rapid ascent of artificial intelligence is transforming the global economy, but lurking beneath this technological marvel is a mounting crisis: the insatiable energy and water appetite of data centers powering AI. From the bustling tech corridors of Singapore and Tokyo to the sprawling deserts of the American Southwest, the AI boom is pushing electricity grids and natural resources to their limits, creating both lucrative opportunities and profound challenges for investors, policymakers, and everyday citizens.
In 2024, data centers worldwide consumed about 1.5% of all electricity, a share projected to double by 2030, according to the International Energy Agency. The United States, a prime battleground in this race, is seeing investor-owned utilities plan over US$1.1 trillion in capital expenditure between 2025 and 2029 just to keep up. Goldman Sachs Research estimates that by the end of the decade, U.S. data centers could gulp down as much as 8% of the nation’s total electricity—an astonishing leap that’s reshaping the energy market and straining the grid. A single high-density AI data center can require as much power as 100,000 households, and the largest planned facilities are set to demand 20 times that amount.
But the story doesn’t end with power. Water, often overlooked, is becoming a flashpoint. In 2023, U.S. data centers used 75 billion gallons for cooling, a figure that’s only set to rise as AI’s heat-intensive workloads multiply. Singapore, a global digital hub, faces its own crisis: by 2030, its data centers are projected to consume 65.55 billion liters of water annually, a 36% jump from 2025 levels. The city-state’s government has already imposed a moratorium on new data center projects, citing environmental concerns and a mere 2% vacancy rate as of 2025. China, meanwhile, saw its data centers consume 15.7 billion cubic meters of water in 2022, with water-stressed provinces like Gansu and Inner Mongolia bearing the brunt of ambitious projects such as the Eastern Data and Western Computing Project.
Asia’s other tech powerhouses aren’t immune. Tokyo, with a 13% data center vacancy rate in early 2025, is grappling with land shortages, soaring construction costs, and persistent power constraints. South Korea is racing to integrate offshore wind and battery storage, but its AI workloads are driving energy needs ever higher. Across the ASEAN region, data centers could account for 2-30% of national electricity demand by 2030, potentially outpacing the deployment of renewable energy.
For investors, these trends are both a warning and an invitation. Traditional valuation models for data centers have focused on metrics like Power Usage Effectiveness (PUE) and proximity to cloud platforms. But as environmental pressures mount, energy source carbon intensity and water scarcity costs are coming to the fore. A data center in Malaysia, for example, might boast a PUE of 1.3, but with 81% of its electricity still derived from fossil fuels, the environmental footprint remains immense.
Singapore-listed companies are already moving to seize the moment. Geo Energy Resources, a coal mining and infrastructure firm, reported a 68% year-on-year revenue surge to US$166 million in the first quarter of 2025, delivering 3.5 million tonnes of coal—double the previous year’s volume. Its subsidiary, PT Marga Bara Jaya, is building integrated infrastructure set for completion by June 2026, which will ramp up production capacity at Tanjung Raya Asphalt by over 20 million tonnes per year. Management is bullish, particularly as U.S. demand for coal technology and exports rises alongside AI-driven energy needs.
Seatrium Limited, specializing in maritime and energy engineering, saw its revenue climb 34% year-on-year to S$5.4 billion in the first half of 2025, with net profit soaring 301% to S$144 million. Its oil and gas segment alone brought in S$3.6 billion, a 26% jump, fueled by global energy demand from data centers and AI. Meanwhile, the offshore wind segment more than doubled revenue to S$1.1 billion, thanks to project completions and high demand in Europe and Asia Pacific. Notably, Seatrium’s April 2024 partnership with A*STAR aims to develop hydrogen and ammonia fuel solutions and integrate AI into maritime infrastructure, a move designed to diversify its energy portfolio and tap into the AI-driven energy surge.
Keppel Infrastructure Trust (KIT) is also riding the wave, with distributable income up 31% year-on-year to S$119 million in the first half of 2025. Its City Energy segment—supplying piped town gas—saw a 43.4% leap to S$30.1 million, expanding into liquified petroleum gas and electric vehicle charging. On June 25, 2025, KIT’s parent, Keppel Ltd, announced a S$1.5 billion partnership with the Asian Infrastructure Investment Bank for sustainable infrastructure and technology-enabled projects, directly addressing the power demands of rapid urbanization and the AI sector.
Yet, the AI revolution’s environmental costs are rapidly catching up. Google reported a 48% rise in greenhouse gas emissions since 2019, largely due to data center expansions for AI. In the U.S. alone, AI data centers now contribute over 100 million tons of CO2 annually—the equivalent of adding 23 million cars to the roads. Water woes are worsening, especially in drought-prone regions like the American Southwest, where overbuilding leaves facilities running at just 20-30% capacity, wasting both energy and precious water. Consumers are feeling the pinch, too: electricity rates for households and small businesses have spiked by as much as 36% in some U.S. states, with utilities prioritizing Big Tech’s massive power deals over public needs.
Governments are scrambling to respond. In the U.S., the Trump administration declared a "national energy emergency" in 2025, fast-tracking permits and loosening environmental rules to accommodate the data center boom. Critics, however, warn this is a short-term fix that fails to address long-term sustainability. The EPA is drafting plans to curb power plant emissions even as AI demand surges. In Asia, regulators are tightening environmental rules for data centers, forcing developers to rethink strategies.
Still, innovation is lighting a path forward. Microsoft, AWS, and Google are investing in renewable energy Power Purchase Agreements (PPAs), advanced cooling technologies like liquid immersion systems, and modular designs to slash resource use. Singapore’s Equinix and Digital Realty are piloting AI-driven water conservation systems, cutting usage by up to 15%. In Malaysia, Microsoft’s cloud regions are designed for 100% carbon-free energy, and Google’s Indonesian projects integrate solar and wind power. Thai and Vietnamese developers are turning to modular designs and edge computing for greater efficiency.
Financing is evolving, too. Green bonds, sustainability-linked loans, and project finance tied to resource reduction targets are becoming the norm. Investors now scrutinize not just compliance with ESG standards, but genuine innovation at the intersection of energy, water, and digital infrastructure. As one industry insider put it, "The next decade will belong to the firms that treat sustainability not as a cost but as a competitive advantage."
The AI boom has set a high-stakes challenge: balance explosive digital growth with environmental stewardship. The winners—whether in Singapore, the U.S., or beyond—will be those who build resilience and sustainability into every watt and every drop.
