Google says datacenters, AI cause its carbon emissions to rise sharply

Google LLC said increased electricity demand driven by artificial intelligence and its growing fleet of data centers has caused the company's greenhouse gas emissions to soar by 48% above its 2019 baseline, creating a challenge for the tech giant in meeting its carbon neutrality goals by 2030.

The company disclosed the findings July 2 in an annual environment report detailing how it is failing to meet its goal of achieving net-zero emissions by 2030. Instead, the Alphabet Inc. company showed that its emissions have only grown since setting the target, rising 13% in 2023 compared with 2022, and 48% since 2019, its baseline year.

The report comes as the International Energy Agency has projected electricity demand could double by 2026 because of exponential load growth from data centers. The US Energy Information Administration, which launched a data center tracking pilot in 2018, showed that data centers use 10 to 40 times more energy per square foot than the typical office building and account for almost 2% of the electricity use in the US.

The growth in data center electricity demand, and the fact that it is being met with fossil generation, is also affecting the sustainability goals of other technology firms. Indeed, data centers have caused Microsoft Corp.'s emissions to increase by 29% above a 2020 baseline, the company said in a May sustainability report.

The trend among data center operators could result in greenhouse gas emissions increasing by 2030. In its environment report Google explained that although the company has managed to make headway in powering most of its facilities with renewable and zero-emissions energy, its primary challenge comes from "hard-to-decarbonize" areas in the Asia-Pacific region.

"In addition, we often see longer lead times between initial investments and construction of clean energy projects and the resulting GHG reductions from them," the report added.

To overcome this issue, Google recently introduced "a clean transition rate," which brings together customers with utilities to help finance new clean energy projects in the US. The company has also unveiled new investments in Asia, including building 1 GW of new solar capacity in Taiwan.

"We know that scaling AI and using it to accelerate climate action is just as crucial as addressing the environmental impact associated with it," the report said. To address the emissions from AI, Google has been transitioning its data center servers to more efficient processors, while also looking at more advanced ways of cooling their operations.

The company's research shows that it can reduce the energy required to train AI by up to 100 times, and the emissions associated with AI by up to 1,000 times. It further said that even with the company falling behind on its climate goals, AI still "holds immense promise to drive climate action" and has the potential to help mitigate 5% to 10% of global greenhouse gas emissions by 2030.

For example, the report said Google products aim to reduce 1 gigaton of carbon equivalent emissions annually for individuals, cities and other partners, and that "we'll continue to develop technologies that help communities adapt to the effects of climate change."

Even with the jump in emissions, Google remains committed to using only carbon-free sources of electricity, with 10 of its operations' grid regions achieving at least 90% carbon-free energy. Google said it has maintained a global average of 64% carbon-free energy. It has also asked its largest manufacturing suppliers to commit to a 100% renewable energy match by 2029.

Google set a goal in 2020 to rely on carbon-free sources of electricity 24 hours a day and seven days a week by 2030. It defines carbon-free energy as any type of electricity generation that does not directly emit carbon dioxide, including but not limited to solar, wind, geothermal, hydro and nuclear power.

The report noted that the use of sustainable biomass-based power with carbon capture and storage would be considered on a case-by-case basis in situations when it is considered carbon-free energy.