Every time you type a question into a search bar, a hidden chain of events begins in massive, humming data centres. While we often think of the internet as “weightless,” it has a very real physical cost in terms of electricity. With the rise of AI-powered tools like ChatGPT and Google’s AI Overviews, that cost is changing dramatically compared to the “classic” keyword search we have used for decades.
The Simple “Blue Link” vs. The “Deep Thinker”
A traditional Google search is remarkably efficient. It functions like a highly organised library index; when you search for a term, it retrieves pre-indexed information. Estimates suggest a standard search consumes roughly 0.3 watt-hours (Wh) of electricity—about the same as leaving a 10-watt LED lightbulb on for two minutes.
In contrast, an AI search tool doesn’t just find information; it “thinks.” It must process billions of parameters to generate a unique, conversational response in real time. Early analyses, such as those from Goldman Sachs, estimated that an AI query could require nearly 3 watt-hours—ten times the power of a standard search.
Efficiency vs. Scale
The good news is that tech companies are rapidly optimising. More recent data from 2025 suggests that newer models, such as GPT-4o, have dropped their energy cost significantly, with some estimates placing them closer to 0.3 Wh, nearly on par with traditional searches. Google reported a 33x reduction in the energy cost of its AI queries within just one year.
However, the “rebound effect” remains a concern. Because AI makes searching more intuitive and useful, we tend to use it more frequently, which can offset these efficiency gains and increase overall energy demand.
What Can You Do?
While individual searches have a small impact, they add up across billions of users. To be more eco-conscious, consider:
- Choosing the right tool: Use standard search for simple facts (e.g., “weather today”) and save AI for complex synthesis.
- Being specific: Better prompts reduce the need for follow-up “turns,” saving energy.
- Direct URLs: Navigating directly to a website instead of searching for it can reduce carbon impact by up to 35%.
As AI becomes the new standard, balancing its immense utility with the energy it requires will be one of the great challenges for the future of the web.
If you want to know more about how to reduce your business’s carbon footprint and still use AI-powered tools, please get in touch.