Google Says Each AI Query Needs 5 Drops of Water

(MENAFN) Google's Gemini artificial intelligence chatbot consumes energy equivalent to merely nine seconds of television viewing per text query while utilizing approximately five water drops for cooling operations, new technical research reveals.

The tech giant outlined these findings in a Google Cloud blog post entitled "How much energy does Google's AI use? We did the math," delivering an exhaustive examination of artificial intelligence's environmental consequences.

The study illuminated Gemini's ecological impact, demonstrating that energy usage, carbon output, and water consumption registered substantially below earlier projections.

Each individual text query to Gemini averages 0.24 watt-hours of energy consumption, generates 0.03 grams of carbon emissions, and requires 0.26 milliliters of water usage.

Enhanced Efficiency and Diminished Environmental Impact
Google engineers revealed Gemini's environmental burden proves far smaller than previously calculated. Per-query energy consumption dropped by 33-fold, while carbon emissions plummeted by 44 times.

Critical developments including model enhancements and software refinements over the preceding year achieved 23-fold energy reductions. Superior resource management and renewable energy programs contributed an additional 1.4-fold decrease in carbon output.

The research noted that previous studies frequently inflated environmental impacts through limited analytical scope. Most earlier assessments examined only active AI processor energy consumption, whereas Google's investigation encompasses idle equipment energy usage, processor and memory power requirements, plus data center cooling infrastructure.

The analysis further disclosed that average water consumption linked to Gemini interactions measured hundreds of times lower than prior estimates, which projected 45 to 50 milliliters per exchange. Updated measurements revealed just 0.26 milliliters of water per text query—a dramatic reduction.

During 2023, Google also unveiled air-cooling systems for new data facilities in water-stressed regions, targeting near-zero water usage in these locations.

Although individual AI query environmental impact remains minimal, the report stressed the significance of evaluating cumulative global effects given billions of daily AI interactions worldwide.

Google's approach incorporates ultra-efficient data centers, custom-designed hardware, and carbon-neutral energy objectives, collectively minimizing AI's comprehensive environmental costs.

MENAFN30082025000045017169ID1109996304