Google reveals Gemini AI uses 0.24 watt-hours per query

Google has become the first major tech company to publicly release detailed energy consumption data for its AI systems, revealing that an average Gemini text prompt uses 0.24 watt-hours of energy and emits 0.03 grams of carbon dioxide equivalent. The transparency milestone comes as AI’s environmental impact faces increasing scrutiny, with 61% of Americans expressing concern about AI electricity usage according to recent polling.

What you should know: Google’s methodology provides the most comprehensive view of AI energy consumption to date, accounting for factors typically overlooked in public estimates.

• The company tracked not just active computing power but also idle machines, host CPUs, RAM, data center overhead, cooling systems, and water consumption.
• A single Gemini query consumes “0.26 milliliters (or about five drops) of water,” which Google compared to “watching TV for less than nine seconds.”
• Google estimated Gemini had 350 million monthly users in March, representing almost half of ChatGPT’s user base.

The big picture: While Google’s numbers are higher than simplified calculations, they’re “substantially lower than many public estimates” of AI resource consumption.

• Google’s comprehensive approach revealed 0.24 Wh energy usage compared to a “non-comprehensive” estimate of 0.10 Wh for the same query.
• “The energy consumption, carbon emissions, and water consumption were actually a lot lower than what we’ve been seeing in some of the public estimates,” said Savannah Goodman, head of Google’s advanced energy labs.
• Over a recent 12-month period, the energy and carbon footprint of median Gemini text prompts dropped “33x and 44x, respectively.”

Industry context: Google’s disclosure contrasts sharply with the opacity of other AI companies regarding their environmental impact.

• OpenAI CEO Sam Altman claimed in June that ChatGPT queries use “about 0.34 watt-hours” but provided no methodology or supporting data.
• Meta’s data centers reportedly consume massive amounts of water, but the company hasn’t shared specifics.
• Major players like Anthropic have remained silent on their AI systems’ resource consumption.

Why this matters: The lack of concrete AI energy data has fueled environmental concerns as the industry rapidly expands infrastructure and power demands.

• Google’s own energy usage has more than doubled in four years despite a 12% reduction in data center emissions.
• President Trump recently pledged $92 billion toward AI infrastructure in Pennsylvania, extending the $500 billion Stargate initiative announced in January.
• The Trump administration’s AI Action Plan aims to “reject radical climate dogma” and expedite environmental permitting for new data centers.

What they’re saying: Google emphasized the importance of comprehensive measurement in understanding AI’s true environmental footprint.

• “We believe this is the most complete view of AI’s overall footprint,” the company stated in its report.
• However, Google acknowledged that neither the data nor claims had been vetted by independent third parties.

Looking ahead: Google’s disclosure could establish industry standards and create competitive pressure for other AI companies to share similar environmental data.

• More transparency could help users and businesses factor emissions into their AI model selection decisions.
• The effectiveness of such measures depends on whether the industry prioritizes renewable energy investment amid rapid growth and shifting regulatory priorities.