Google, in partnership with Kairos Power and the Tennessee Valley Authority, has selected Tennessee as the site for an advanced small modular nuclear reactor expected to supply electricity to its data centers in the U.S. Southeast by 2030. Concurrently, Google released a technical paper detailing the environmental impact of its Gemini AI model, revealing that a median Gemini Apps text prompt consumes 0.24 watt-hours of energy, 0.26 milliliters of water, and emits 0.03 grams of CO2. This energy use is equivalent to watching television for less than nine seconds and matches the energy consumption of a Google search in 2008. Over the past year, Google has achieved a 33-fold reduction in energy consumption and a 44-fold reduction in carbon footprint per Gemini prompt, attributed to improvements in model architecture, use of Tensor Processing Units (TPUs), quantization techniques, and renewable energy-backed data centers. Comparatively, ChatGPT uses slightly more energy and water per prompt, while training large models like GPT-4 requires substantially higher energy. Google’s efforts highlight a growing focus on transparency and efficiency in AI operations, with the nuclear-powered data center initiative further underscoring its commitment to sustainable infrastructure.