Google's Greenhouse Gas Emissions Soar Due to AI
TapTechNews July 3rd news, the greenhouse gas emissions of tech giant Google have soared by nearly 50% in the past five years, mainly due to the staggering power consumption of its giant data centers for artificial intelligence computing. According to Google's 2024 environmental report released on Tuesday, the company seems to be渐行渐远 from its goal of achieving carbon neutrality by 2030.
The report shows that In 2023, Google's carbon dioxide emissions were 14.3 million tons, an increase of 48% compared to 2019 and a 13% increase from the previous year. In the report, Google attributes this increase in emissions to increased energy consumption in data centers and increased emissions in the supply chain.
As we incorporate more AI technology into our products, it is expected that future emissions will remain high due to the energy demand from infrastructure investment, and achieving the emission reduction target faces challenges, the report writes.
TapTechNews observes that Google's report reveals the environmental impact brought about by the explosive development of artificial intelligence. Tech giants such as Google, Microsoft, Amazon, Meta, and Apple are planning to invest billions of dollars in artificial intelligence research and development, but training artificial intelligence models requires an extremely huge amount of electricity. In addition, the use of artificial intelligence functions itself also consumes a large amount of electricity. In 2023, researchers from the artificial intelligence start-up HuggingFace and Carnegie Mellon University found that the amount of electricity consumed to generate a single image using artificial intelligence is enough to fully charge a smartphone. According to the Financial Times, Bernstein analysts said that artificial intelligence will double the growth rate of US power demand, and the total power consumption may exceed the current supply level in the next two years.
Last month, Microsoft, which also promised to achieve carbon negative emissions by the end of the 2030s, also released its environmental report, which shows that its greenhouse gas emissions have increased by nearly 30% since 2020 due to the increase in data center construction.
In Google's report, due to the increase in artificial intelligence workloads, the amount of water used by its data centers to keep cool has also risen significantly. In 2023, Google's data center water consumption increased by 17% from the previous year to 6.1 billion liters, which is equivalent to the annual irrigation water consumption of approximately 41 golf courses in the southwestern United States.
As our business and industry continue to develop, we expect that the total greenhouse gas emissions will increase before they decline in order to finally achieve the established absolute emission reduction target, Google's report writes, but does not explain the specific reasons for the decline in emissions. Predicting the future environmental impact of artificial intelligence is complex and constantly changing, and our past historical trends may not fully capture the future development trajectory of artificial intelligence. As we deeply integrate artificial intelligence into our product portfolio, the distinction between artificial intelligence and other workloads will become less important.