Transition Finance Holds Key to AI Data Centers' ESG Paradox

Park Jae-heum, Leader of EY Hanyoung ESG Impact Hub

Artificial intelligence has evolved beyond individual corporate technology competition to become a core factor determining national competitiveness. With the spread of generative AI, data centers are being reorganized from simple IT infrastructure into strategic assets encompassing industry, finance, and security.

AI data centers face what is called the ESG paradox—enormous power consumption, carbon emissions, and local community opposition (NIMBY). The International Energy Agency (IEA) projects that global data center power demand will more than double by 2030 compared to 2024 due to AI proliferation. Microsoft also reported that its carbon emissions increased approximately 23% compared to 2020 due to AI and cloud expansion. This is why AI is perceived as both a symbol of innovation and a source of environmental burden.

From an ESG perspective, AI possesses "double materiality." While consuming massive amounts of energy, it can also serve as a tool to accelerate carbon reduction by improving efficiency across industries. The key is not whether to adopt AI, but how sustainably it can be built and operated.

Transition finance is emerging as a solution in this context. Unlike green finance, which supports only green activities, transition finance supports the process of carbon-intensive industries moving toward low-carbon structures. AI data centers are evaluated as representative transition assets—while their initial environmental burden is significant, they can reduce carbon intensity through operational efficiency improvements and renewable energy adoption.

In fact, global operator CyrusOne secured approximately $11.2 billion in sustainability-linked loans (SLL) in 2024. The structure adjusts interest rates based on achievement of Power Usage Effectiveness (PUE) and Water Usage Effectiveness (WUE) improvement targets. Capital markets have begun prioritizing "transition credibility" over "current eco-friendliness."

Domestic financial holding companies are also investing hundreds of billions of won, recognizing data centers as core infrastructure for the national AI transformation (AX). However, financial institutions need sophisticated frameworks that systematically evaluate long-term energy efficiency improvement pathways and carbon reduction targets (KPIs), and deploy capital in stages based on implementation performance. Companies must also demonstrate quantified efficiency improvement results through AI optimization algorithms and other means.

Social acceptability remains a challenge. To reduce local opposition during construction, data centers must be redefined not as closed facilities but as assets that share benefits. If structures are designed where residents participate in investment and returns—like "public participation infrastructure funds"—data centers can become engines of regional growth.

Ultimately, the key to solving AI data centers' ESG paradox lies in combining environmental transition and social acceptability with finance. Transition finance must become a lever driving corporate low-carbon innovation, while companies must simultaneously transform data centers into assets that coexist with local communities through benefit-sharing models. Only when capital's good intentions meet corporate social responsibility will we truly embrace the prosperous future that AI offers.