top of page
Search
Writer's pictureSylvain Richer de Forges

Analysis of the energy consumption associated with generative AI searches

Understanding the Energy Footprint of AI-Powered Search vs. Traditional Google Search




In our quest for knowledge, we often overlook the energy footprint of our online searches. With the rise of AI-powered search engines, it's crucial to understand how they compare to traditional searches in terms of sustainability.



AI-Powered Search:


AI search engines, like those leveraging GPT-4, consume significantly more energy due to the computational power required. A single AI search query can consume up to 10 times more energy than a traditional search. According to a study by OpenAI, the energy consumption of training a large AI model can be as much as 284,000 kWh .



Traditional Search:


In contrast, traditional searches are more energy-efficient. Google has reported that the energy used for one search query is approximately 0.3 watt-hours (Wh), which equates to about 0.0001 kWh . Given Google's vast infrastructure and ongoing improvements in data center efficiency, this number continues to decrease.



By making informed choices, we can leverage the power of AI responsibly and sustainably. Let’s lead the change towards a more sustainable digital future!




0 views0 comments

Comments


bottom of page