Artificial intelligence (AI) has emerged as a significant consumer of resources, and new reports reveal the environmental impact of Big Tech’s AI ambitions.
Google and Amazon released their sustainability reports recently, revealing concerning figures. Despite their commitments to achieve net-zero carbon emissions in the near future, the rise of AI has complicated their efforts. Google’s carbon emissions have surged by 25% compared to last year, while Amazon’s emissions have increased by 16%.
A detailed analysis of these reports indicates that both companies may need to undertake substantial and potentially expensive changes to meet their net-zero objectives.
While neither company explicitly blames AI for the uptick in emissions, there are hints that point in that direction.
AI at the Core
Amazon and Google both noted a significant rise in their energy consumption over the past year, coinciding with the increased use of AI. They referenced carbon intensity, which measures emissions generated per dollar of revenue, a metric often used in climate discussions. Interestingly, both companies also dedicated portions of their reports to highlight how AI can positively impact environmental efforts, possibly indicating a defensive stance.
A deeper dive into the data reveals that both firms are managing their carbon emissions related to energy purchases relatively well, thanks to years of investment in renewable energy. However, this could change as they begin to invest in natural gas power to meet the energy demands of AI.
The majority of the emissions increase for both companies can be attributed to Scope 3 emissions, which cover pollution from sources they do not directly control, such as the services and products they acquire or sell. For Amazon and Google, this includes items like GPU purchases and the energy consumption of their devices.
Google combines two categories of Scope 3 emissions—it acknowledges that emissions from sold products are minor. Most of Google’s hardware consists of smaller devices that don’t consume extensive energy. Consequently, data centers are likely the primary contributor. Over the past year, Google’s Scope 3 emissions surged by 2.1 million metric tons, effectively doubling since their 2019 baseline.
For Amazon, the increase in Scope 3 emissions stems primarily from capital goods and energy sources. Significant growth in data centers partly explains why Amazon’s Scope 3 emissions have climbed higher than Google’s. Notably, the company reported that in 2025, it added more data center capacity globally than any other firm, exceeding 1.2 gigawatt (GW) in the final quarter alone.
Facing Challenges
Such financial investment explains the newfound difficulties in decarbonization efforts. Traditionally, the most significant source of carbon emissions originated from energy used in offices and smaller data centers, which could be offset with renewable energy purchases.
However, AI has disrupted this balance. While tech companies could rely on renewable sources and batteries to power their data centers, they are increasingly reverting to fossil fuels. This trend complicates their net-zero commitments, although it is not an irreversible path.
Additionally, the emissions linked to constructing and equipping data centers pose challenges. The steel and cement sectors are notorious for their high pollution levels, and while some startups attempt to create low or zero-carbon alternatives, these solutions are not yet available at the required scale for tech firms.
Further complicating matters are the GPUs and memory chips essential for the AI boom. Manufacturing semiconductors consumes considerable energy, and many state-of-the-art factories are based in Asia, where energy sources mainly consist of fossil fuels. Moreover, the chemicals utilized in these factories are powerful greenhouse gases that can significantly contribute to global warming. The surge in chip demand has likely added to both Amazon’s and Google’s carbon footprints.
Though these challenges are daunting, they are not insurmountable. Amazon, Google, and their counterparts will need to significantly increase their renewable energy acquisitions, invest in advanced manufacturing processes for steel and cement, and purchase millions of tons of carbon removal credits. While achieving their net-zero commitments remains feasible, the integration of AI has certainly added to the complexity of the task.
