AI represents one of the most significant technological transformations of our time, which will become more clear in the coming decade. Applied to fields such as medicine, energy, autonomous systems and quantum calculation, AI is ready to help people tackle major societal challenges, whether to help students learn, diagnose cancer in the past, make complex transport and cyber security systems more secure or even predict the path to wild fires to the first respondents.
Realizing the potential of AI will require robust energy infrastructure, more efficient energy consumption and even new innovative technology solutions. We are approaching this from many angles investment in new infrastructure, technically smarter and more resilient grids and scaling both mature and next generation sources of pure energy. At the same time, we are also focused on maximizing the efficiency of each layer of our operations from the design of our custom-built hardware for the software and the models running in our data centers.
To improve AI’s energy efficiency, a clear and comprehensive understanding of AI’s environmental footprint is important. So far, extensive data on the energy and environmental impact of the AI ​​-Inference has been limited.
Today, we help close this gap by releasing a comprehensive method of measuring energy, water and carbon emissions of Google’s AI models. Critically, our work with model efficiency delivers rapid progress. During the period of a 12-month period, while the supply of higher quality responses, median energy consumption and carbon footprint fell. Gemini-app-text prompt with factors of 33x and 44x respectively. Based on our recent analysis, we found that our work with efficiency proves to be efficient and the energy consumed per year. Medianprompt, equivalent to watching TV for less than nine seconds. These progress are based on our many years of commitment to the effectiveness of data center. In 2024, for example, we reduced our data center energy emissions by 12%, even when electricity consumption grew by 27% year-over years, driven by the expansion of our business and services.
As we continue to invest in the technology and innovation needed to accommodate significant new energy needs, transparency is the key to progress. We hope that this study contributes to ongoing efforts to develop effective AI at this critical time of energy, sustainability and scientific discovery – for the benefit of everyone. Read more about how we got to the math and our full power approach to energy innovation and dive into our technical report.