Discover your AI carbon footprint and learn how to offset it through nature-based solutions like tree planting or environmental cleanup.
Adjust your usage and click calculate to see your environmental impact
A typical ChatGPT query is about 0.8 Wh. A median Gemini text prompt is about 0.24 Wh. Reasoning and long-context calls, such as GPT-5.5 thinking or Claude Opus on a long document, can use roughly 10–20 times more.
A mature tree absorbs approximately 21 kg (48 lbs) of CO₂ per year, helping offset carbon emissions.
Removing 1 kg of plastic from the ocean prevents ~3 kg of CO₂ equivalent emissions over its lifecycle.
Global data center electricity consumption is projected to reach 1,050 TWh by 2026, with AI as a primary driver of growth.
Our methodology is based on peer-reviewed research on AI energy consumption
We estimate energy per query from published figures: Google's Gemini measurement, OpenAI's 2025 ChatGPT average, and independent estimates for GPT-5.5, Claude, and reasoning models. A long prompt or a thinking trace uses more than a short chat.
We apply average global grid carbon intensity (~0.4 kg CO₂/kWh) to convert energy usage to carbon emissions.
We calculate equivalent trees needed (21 kg CO₂/tree/year) or trash collection impact (3 kg CO₂ prevented per kg plastic).
These calculations are estimates based on publicly available research. Actual energy consumption varies based on data center efficiency, renewable energy usage, and specific model implementations.
Start planning your next marine wildlife adventure with Voyatlas. Join the waitlist to get early access to the mobile app the moment it launches. Have a question? Reach out on social media or email info@voyatlas.com.
