OpenAI’s GPT-5 is a marvel of modern engineering, capable of tasks that were once the exclusive domain of humans. Yet, as the AI world celebrates this new leap forward, a quiet but serious concern is bubbling up: its massive energy consumption. While the company has provided no official data, independent researchers are finding that the power required to run the new model is far greater than its predecessors, raising alarms about the environmental sustainability of ever-more-powerful AI.
The numbers are striking. According to researchers at the University of Rhode Island’s AI lab, generating a medium-length response from GPT-5 uses an average of 18 watt-hours of electricity. This is a significant jump and places the new model among the most energy-intensive AIs ever benchmarked. When you consider that ChatGPT handles billions of requests a day, this isn’t just an abstract number. The collective energy use of GPT-5 could be enough to power 1.5 million US homes daily, an astonishing figure that brings the environmental impact of AI into sharp focus.
This surge in energy consumption is directly linked to the model’s scale. Although OpenAI has not disclosed the number of parameters for GPT-5, experts believe it is “several times larger than GPT-4.” A disclosure from the French company Mistral confirmed a “strong correlation” between a model’s size and its resource consumption, meaning that as models grow, their energy footprint grows in kind. This principle suggests that the quest for ever-larger and more capable models, a stated goal of companies like OpenAI, will come at an escalating environmental cost.
Adding to the power drain are GPT-5’s new features. Its advanced “reasoning mode” and multimodal capabilities, which allow it to work with video and images, require more computational effort and time. A professor working on AI energy consumption stated that using the reasoning mode could increase resource usage by “five to 10” times for the same answer. This suggests that while a “mixture-of-experts” architecture offers some efficiency, the new, more complex tasks are driving the overall energy footprint to new heights. The urgent call for transparency from AI developers is a direct response to this growing environmental concern.
