The surge in generative artificial intelligence is significantly boosting revenue for major tech companies, but the high energy demand is turning these companies into substantial contributors to climate change.
Earlier this month, Google reported a 48% rise in its carbon emissions since 2019, primarily due to the energy consumption by its data centers and supply chain emissions. Google’s 2024 Environmental Report revealed a 13% year-over-year increase in carbon emissions for 2023. Despite a goal set in 2021 to achieve net-zero emissions by the end of the decade, Google disclosed that from 2023 onwards it would no longer maintain operational carbon neutrality. Instead, the company plans to focus on other “carbon solutions and partnerships” to meet its net-zero objective.
Similarly, Microsoft had announced in 2020 its aim to become “carbon negative” by the decade’s end. However, its 2024 Environmental Sustainability Report showed a nearly 31% rise in carbon emissions since 2020, largely due to the expansion of data centers for AI workloads and hardware such as semiconductors and servers.
Microsoft acknowledged the challenges, highlighting the need for greener concrete, steel, fuels, and chips. Their statement emphasized that their challenges in part reflect the global hurdles in developing and using more sustainable materials.
In April, Ami Badani, chief marketing officer of British chip designer Arm, noted that AI chatbots like OpenAI’s ChatGPT account for 2% of global electricity consumption, warning that such demand could eventually hinder AI advancements. A Goldman Sachs study indicated that a ChatGPT query requires nearly ten times more electricity than a Google search.
According to the Electric Power Research Institute, data centers could consume up to 9% of U.S. electricity by 2030, more than twice the current usage. A third of U.S. nuclear power plants are reportedly negotiating deals with tech companies to provide electricity. Furthermore, OpenAI CEO Sam Altman invested in Exowatt, a startup that is developing modules to store energy as heat and generate electricity for AI data centers.