Google Unveils New Data on AI’s Environmental Impact

Amid mounting global concern about the environmental footprint of artificial intelligence (AI), Google has taken a step by publishing new data on the energy, water, and carbon costs of its Gemini AI service. According to the company, sending a single text prompt through Gemini consumes as much energy as watching television for nine seconds

Economic Growth Versus Sustainability

Economists argue that AI will boost productivity and drive global growth, with Goldman Sachs projecting it could expand global GDP by 7%, or $7 trillion, over the next decade. However, scientists caution that these gains may be undermined by AI’s immense resource demands, which are still not fully understood. For instance, AI relies on energy-hungry data centers, some of which consume between 100 and 1,000 megawatts enough to power 80,000 to 800,000 homes, according to the U.S. Government Accountability Office.

Water and Hardware Costs

Beyond power, AI data centers use enormous volumes of water to cool servers, raising concerns over strained municipal supplies and ecological disruption. MIT research warns that the rise of generative AI has also increased demand for high-performance computing hardware, adding indirect environmental costs from manufacturing and transport.

Calls for Transparency

Ben Gomes, Google’s senior vice president of learning and sustainability, highlighted the need for better data. “In order to improve the energy efficiency of AI, a clear and comprehensive understanding of AI’s environmental footprint is important. To date, comprehensive data on the energy and environmental impact of AI inference has been limited,” he said in a blog post.

Gemini’s Environmental Profile

According to Google, a Gemini text query uses 0.24 watt-hours of energy, emits 0.03 grams of carbon dioxide equivalent, and consumes 0.26 milliliters of water about five drops. By comparison, OpenAI’s ChatGPT queries average 0.34 watt-hours and one fifteenth of a teaspoon of water, CEO Sam Altman has noted. Google also reported progress in cutting Gemini’s footprint. Over the past year, the energy consumption and carbon emissions of the median text prompt dropped by factors of 33 and 44, respectively.

Currently, no regulations require tech firms to disclose how much energy or water their AI models consume. But with adoption growing rapidly, Google’s findings underscore the urgent need for transparency and innovation to ensure AI’s future is not only intelligent but also sustainable.