Google and Amazon disclosed a significant rise in their greenhouse gas emissions for 2025, with both companies attributing the increase directly to the rapid expansion of artificial intelligence data centers. The figures, published in their latest environmental reports, mark a notable setback for the firms' publicly stated climate pledges.
The surge underscores a growing tension between the tech sector's aggressive AI ambitions and its sustainability goals. As AI model training and deployment require immense computational power—and thus electricity—the carbon footprint of these operations is ballooning faster than efficiency gains or renewable energy purchases can offset.
Specific numerical details, such as percentage increases or total emissions in metric tons, were not provided in the reports. However, both companies acknowledged that the growth in emissions outpaces their internal reduction targets, which rely heavily on renewable energy credits and carbon offsets.
This development raises broader questions about the environmental viability of the current AI arms race. Major industry players, including Microsoft and Meta, face similar scrutiny over their energy consumption, as regulators and investors increasingly demand transparency and concrete action on decarbonization.
Critics argue that voluntary pledges lack teeth without mandatory carbon reporting standards. The companies maintain that their long-term investments in clean energy and carbon removal technologies will eventually close the gap.