Gemini is Google’s personal artificial intelligence (AI) assistant. The company boasts that it can help you brainstorm, study, generate images for you, and connect with Google Maps to help you navigate. Gemini is even expected to integrate with Google Home soon.
However, as climate change impacts our planet, there is growing concern about the AI boom. Companies are scrambling to introduce the technology into their workflows or consumer products to become more profitable. As other efforts to combat climate change — such as electric vehicles and reducing manufacturing carbon footprints — become increasingly popular, the rising use of AI has brought concerns that we’re undoing all the progress we’ve made.
In light of these concerns, Google released a report in August 2025 showing what it claims is the low energy use for the average text-based Gemini prompt. Reportedly, a single prompt uses the same energy as watching less than nine seconds of TV. The report also emphasizes a message of sustainability through future innovation. It’s a hopeful message, but the reality of creating and managing AI is complex, and overall, the numbers for the future don’t seem to be promising.
Google Gemini’s energy use per prompt
Google analyzed the median amount of energy used for text-based prompts on its Gemini app. Note that this does not account for voice prompts or other media, like the new Gemini photo feature coming to Android soon. Take note that medians are the middle amount, which does not account for outliers like low- or high-energy prompts.
The resulting technical report showed that per text prompt, Gemini uses 0.24 watt-hours of energy, emits 0.03 grams of carbon dioxide, and uses 0.26 milliliters of water. To put it into a more relatable perspective for daily human consumption, the report says it would be the equivalent of watching television for less than nine seconds. So if you are watching “Doctor Who: The Well,” that would be the same as over three hundred Gemini prompts. That doesn’t seem like so much for an individual, but it’s also important to note that Google Gemini has about 47 million active users in 2025.
Still, Google is quick to point out improvements that have been made in its energy usage, highlighting that over 12 months, the app’s total carbon footprint has dropped 44 times. Google is also focusing on decreasing the amount of water needed for its data centers to cool, helping with the global water consumption issue.
The bigger issue about AI energy use
Of course, AI energy usage is a very complex picture to look at as a whole. There are concerns that it can make your power bill go up even if you don’t use AI. Training AI models before launch requires huge amounts of energy. Then, after they’re launched, they must be consistently updated, have bugs fixed, and be expanded into different areas, such as Gemini being integrated into Android Auto.
There is also the energy consumption of data centers that run AI models like Gemini, as well as the hardware manufacturing and shipping requirements needed for these models. Looking at the energy used per the average text prompt is a small part of the whole. According to research from the Massachusetts Institute of Technology (MIT) News, North American data centers needed 2,688 megawatts in 2022, but in 2023 that increased to 5,341 megawatts, of which AI was a contributor.
Despite the positive messaging behind Google’s report, MIT authored a report in 2024 titled “The Climate and Sustainability Implications of Generative AI,” evaluating its current and future impact. Noman Bashir, the paper’s lead author and postdoctoral fellow in the Computer Science and Artificial Intelligence Laboratory, spoke to MIT News about its findings. “The demand for new data centers cannot be met in a sustainable way,” he said. “The pace at which companies are building new data centers means the bulk of the electricity to power them must come from fossil fuel-based power plants.”
