The rapid development of artificial intelligence is putting increasing pressure on the climate commitments of major technology corporations as their greenhouse gas emissions rise. Four leading providers of AI and cloud services recorded an increase in emissions by up to 239% between 2020 and 2024.
According to the latest report, 'Greening Digital Companies 2026,' prepared by the International Telecommunication Union (ITU) of the United Nations and the World Benchmarking Alliance (WBA), technology companies are making progress in climate reporting and renewable energy use, but the pace of emission reduction is still insufficient to meet global climate goals.
This report analyzes 200 large digital companies worldwide, examining their greenhouse gas emissions, energy consumption, use of renewable electricity, climate targets, and plans for transitioning to low-carbon operations. The findings highlight a growing contradiction between the massive computational demands of AI and the industry's climate ambitions.
While four major AI and cloud service providers demonstrated a sharp rise in operational emissions, the report showed that 14 large telecommunications companies managed to reduce their emissions by 11% over the same period.
The ITU Secretary-General, Dorin Bogdan-Martin, noted: 'Despite the enormous potential of digital technologies to combat climate change, we cannot ignore their growing energy needs and emissions. Environmental sustainability must be integrated into how we design, power, and scale the technologies that shape our shared digital future.'
Rising Energy Demand from AI
Out of the 200 companies included in the study, 301 million tons of operational greenhouse gas emissions were registered in 2024, equivalent to 0.8% of global energy-related emissions. These emissions increased by 1.2% compared to the previous year.
Electricity consumption was significantly higher: 163 companies reported a cumulative consumption of 494 terawatt-hours in 2024, accounting for about 1.7% of global electricity consumption. Among the companies that provided data, China Mobile was the largest electricity consumer with 63 TWh, followed by Alphabet and Samsung at 32 TWh each, and Microsoft used 30 TWh. Amazon did not provide electricity data for 2024.
The ten largest electricity consumers together used 269 TWh, demonstrating the scale of infrastructure required to support the digital economy. ITU forecasts that energy demand will continue to grow as AI, cloud computing, and other digital infrastructure expand.
However, AI has the potential to reduce emissions in other areas through applications such as energy optimization, renewable energy forecasting, and efficiency improvement. The main challenge, however, is ensuring that these benefits are not offset by the energy required to train and operate increasingly powerful AI systems.
Implications for South Africa
This issue is becoming increasingly relevant for South Africa as the country prepares for a significant expansion of data center capacity. According to the South African Development Bank's 'Digital Infrastructure Investment Study,' the IT load of data centers in South Africa is projected to grow from 435 MW in 2024 to 829 MW by 2029, representing a compound annual growth rate of 17.5%. This growth is partly driven by the increased adoption of AI, cloud computing, and the demand for data processing and storage.
The projected increase will lead to a substantial new energy requirement while the country focuses on securing reliable and clean power sources. The DBSA study indicates that AI workloads require more energy and higher rack density than traditional IT workloads, meaning local data centers must upgrade and expand existing infrastructure to accommodate these loads.
As energy demand grows, data center operators in South Africa are increasingly turning to renewable energy projects and 'power-wheeling' schemes to supplement grid supply and achieve sustainability goals.
Climate Goals Not Being Met
The global report found that the technology sector is advancing in renewable energy and climate commitments, but implementation remains uneven. Only 25 out of 200 assessed companies stated they use 100% renewable energy.
Meanwhile, 151 companies set short-term emission reduction targets. However, only 85 of these were deemed aligned with the plan based on current progress, and only 81 companies had comprehensive plans to achieve their climate goals. The report also revealed significant gaps in emissions reporting. While most companies disclosed Scope 1 and Scope 2 emissions, less than half reported emissions across their entire value chain.
This is significant because a large portion of the technology sector's environmental impact lies outside the electricity used in its own offices and data centers. Jebrand Haverkamp, Executive Director of the World Benchmarking Alliance, stated: 'Digital companies must engage with suppliers and address the emissions in the products and services they rely on.'
Only the electronics sector accounts for 53% of the reported emissions across the three subsectors assessed by WBA, highlighting the importance of considering not just the tech companies themselves, but also their supply chains.
Who Is Performing Best?
Swisscom ranked first in the assessment and was the only company to receive a perfect score. Accenture, Deutsche Telekom, Vodafone, Capgemini, and Telefonica are also among the strongest participants. On the other hand, 18 companies did not receive scores because they did not provide sufficient information, including Elon Musk's X and SpaceX. Among the lowest-ranked companies that received scores were Huawei, Spotify, Nintendo, Weibo, Xiaomi, Zoom, and Toshiba TEC.
These findings present a complex balance for the technology industry: AI and digital services can help other sectors reduce their carbon footprint, but the infrastructure required to support this growth consumes ever-increasing amounts of electricity and generates more emissions itself.