Big tech representatives say they do what they can to reduce emissions from their in-house data centers. But a new analysis by The Guardian has shown that the true costs of technology might be far higher.
According to the calculations of the British news organization, from 2020 to 2022 the real emissions from the company-owned data centers of Google, Microsoft, Meta, and Apple are likely about 662% (7.62 times) higher than officially reported.
Amazon is actually the largest emitter of the “big five,” but the corporation has been excluded from the calculations because its business model makes it difficult to isolate the necessary data.
Still, all five tech giants have been claiming carbon neutrality, even though Google admitted in its 2020 environmental report that the company’s emissions surged nearly 50% compared to 2019. The company’s total data center electricity consumption grew 17% in 2023 alone.
Researchers say creative accounting allows firms to claim carbon neutrality. Companies usually purchase Recs (renewable energy certificates), but the problem is that renewable energy doesn’t need to be consumed by their actual facilities.
In other words, tech companies use Recs to calculate “market-based” emissions. But “location-based” emissions – the actual emissions generated from the area where the data centers are located – are much more precise, The Guardian says.
According to the analysis, the massive differences in location-based and official emissions numbers showcase just how carbon-intensive data centers really are and how deceptive firms’ official emissions numbers can be.
For instance, Meta reported its official emissions for 2022 as 273 metric tons of CO₂. However, under the location-based accounting system, the number jumps to more than 3.8 million metric tons of CO₂ equivalent for data centers alone.
The numbers are similar for other tech giants, too, and it’s also important to note that they rent a large portion of their data center capacity from third-party operators.
The Guardian's data is from 2020-2022. This means that the energy-hungry AI boom hadn’t even started, and AI is far more energy-intensive for data centers.
According to Goldman Sachs, a single ChatGPT query needs nearly 10 times as much electricity to process as a Google search. The investment bank also says that data center power demand will grow 160% by 2030.
The US government is now paying attention to the concerns. After meeting top tech and power company executives last week, the White House announced a new task force to deal with the growing needs of AI infrastructure.
Your email address will not be published. Required fields are markedmarked