OpenAI’s GPT-5 Energy Use Rivals That of Small Countries

OpenAI's latest AI model, GPT-5, has drawn attention not only for its capabilities but also for its staggering energy consumption. Estimates indicate that GPT-5 requires approximately 16.4 terawatt-hours (TWh) of electricity each year, a figure comparable to the annual energy use of entire countries such as Slovenia or Cuba.

This vast energy requirement stems from the model's operational demands. Each response generated by GPT-5 consumes around 18 watt-hours of electricity. With daily processing of approximately 2.5 billion queries, this results in a total consumption of roughly 45 gigawatt-hours (GWh) each day. To put this into perspective, the model's daily power needs are equivalent to the output of two large nuclear reactors.

The energy consumption of GPT-5 starkly contrasts with its predecessor, GPT-4, which utilized only about 2 watt-hours per response. This means that GPT-5 consumes approximately eight to nine times more energy per query than GPT-4. Such a dramatic increase raises critical questions about the sustainability of deploying large-scale AI models in today's energy-conscious climate.

In contrast, competing AI models show significantly lower energy consumption rates. Anthropic's Claude 2 operates on several watt-hours per query, while Google's Gemini model has achieved remarkable efficiency, consuming only 0.24 watt-hours per response. This efficiency translates to a staggering 75 times less energy use compared to GPT-5. Google has reported that it has optimized Gemini's energy consumption by a factor of 33 over the past year, showcasing the potential for energy-efficient AI development.

Despite the concerning figures surrounding GPT-5's energy use, OpenAI has not publicly shared detailed data regarding its model's consumption, leaving analysts and researchers to rely on independent estimates. The implications of such high energy demands are significant, especially when considering ongoing discussions about the environmental impact of AI technologies.

As the AI sector continues to grow, the sustainability of energy use in AI development will be a critical area of focus. The discrepancy in energy consumption between leading AI models suggests room for improvement and innovation in creating more energy-efficient systems. With increasing scrutiny on the environmental footprint of technology, companies like OpenAI may need to address these concerns to maintain their competitive edge and public trust.