Copyright OilPrice
Will climate negotiations move beyond voluntary pledges to establish clear regulatory mandates and carbon pricing signals that reflect the environmental costs of computation? Global leaders will be compelled to adopt a dual perspective – acknowledging AI’s profound potential as a climate-action accelerator while confronting its rapidly expanding environmental footprint and the need for a robust global governance framework. AI is the single biggest driver of new electricity demand With both foremost world powers leading the surge in data center power demand, their burden to reduce emissions will likely be a focal point of conversation at COP30. The development and deployment of advanced AI, particularly large generative models such as ChatGPT, is the primary force driving the exponential growth in global power demand for data centers. Rystad Energy projects that power consumption from data centers will more than double by 2030 and escalate to 2,700 terawatt-hours (TWh) globally by 2050, from approximately 400 TWh in the early 2020s, with the US and China leading this expansion. These models, designed to generate original content by learning from colossal datasets, require an ever-increasing computational load. Power sources matter for data center workloads Data centers require round-the-clock power to maintain core digital services. A brief power failure can corrupt data, cause physical equipment damage, halt billions of transactions, and violate the demanding five-nines (99.999%) uptime guarantees agreed upon with customers. To ensure this essential reliability, facilities utilize redundancy layered systems, which start from grid connectivity and are often connected to multiple lines from different substations. They also deploy uninterruptible power supply (UPS) systems, which are large banks of batteries that provide an instantaneous electrical bridge of typically five to 15 minutes when grid power fails. Finally, they have backup sources such as diesel and gas generators that take over from UPS systems until grid power is restored. Data center demand for backup power generation is impacting the supply chain of gas power generation. Significant capital expenditure escalation, with price inflation exceeding 300%, and prolonged lead times of over five years for high-efficiency gas turbines, also known as combined-cycle gas turbines (CCGT), pose a critical concern for hyperscale data center deployment. Most importantly, this sharp price increase is simultaneously eroding the economic viability of new thermal base-load power plants, particularly impacting developing economies that are attempting to expand their thermal generation capacity. 24/7 clean power is the key to ensuring progress isn’t sacrificed for compliance Major tech companies, such as Google, Amazon, and Microsoft, are among the world’s largest corporate purchasers of renewable energy, often utilizing power purchase agreements (PPAs) to achieve 100% renewable energy on an annual, global basis, thereby driving private sector investments in the clean energy transition. Simply buying enough renewable energy annually to match global consumption does not mean the data center is running on clean energy every hour. When the sun sets or the wind stops, the data center continues to draw power from the local grid, which often includes sources of fossil fuels. The true challenge is to match a 24/7 carbon-free energy supply. This means that every kilowatt-hour consumed each hour of the day must come from local, carbon-free sources. Achieving this requires significant investment in new, diverse renewable and clean base-load generation capacity, improved grid flexibility, and advanced energy storage solutions. Data centers and their water-positive strategy Beyond electricity, data centers are major water consumers, primarily for cooling the heat-generating information technology (IT) equipment. This demand intensifies resource competition in drought-prone and water-stressed communities, and climate hazards further threaten data center resilience. To mitigate this impact, some developers have adopted a "water positive" strategy, meaning that a company returns more freshwater to communities and the environment than it withdraws. These strategies typically involve prioritizing water-efficient cooling technologies, utilizing reclaimed wastewater or closed-loop recirculation, or selecting a more climate-conscious site. The rapid growth of AI, particularly in large data center clusters, presents significant challenges for climate negotiations. The economic efficiency promised by generative AI comes at the expense of grid stability and sustains a carbon lock-in. The rapid increase in load pressures power infrastructure to extend the life of high-carbon energy assets and pass billions in upgrade costs onto consumers. These issues will be central to discussions at COP30. By Rystad Energy More Top Reads From Oilprice.com API Reports Sizeable Crude Oil Build Libya Announces New Onshore Oil Discovery HSBC Maintains Bearish Oil Price Forecast Despite OPEC+ Hikes
