AI’s Hidden Energy Costs: Why Experts Stay Silent
▼ Summary
– Sam Altman stated that an average ChatGPT query uses 0.34 watt-hours of energy, comparable to an oven running for just over a second or a lightbulb for a few minutes.
– Experts criticize Altman’s figure for lacking context, such as how OpenAI defines an “average” query or whether it includes energy used for training models and server cooling.
– AI’s growing energy use raises concerns about increased carbon emissions, but research is hindered by limited environmental disclosures from major AI companies like OpenAI.
– A study found that 84% of large language model usage in May 2025 involved models with no environmental transparency, leaving consumers unaware of their impacts.
– Claims about AI energy use, such as ChatGPT queries consuming 10 times more energy than Google searches, often lack reliable sources but are widely repeated in media and policy discussions.
The hidden energy costs of AI tools like ChatGPT are raising serious environmental concerns, yet tech companies remain tight-lipped about their actual impact. While OpenAI CEO Sam Altman recently shared that a single query consumes roughly 0.34 watt-hours, comparable to an oven running for just over a second, experts argue this figure lacks critical context. Without transparency around how “average” queries are defined or whether server cooling and model training are included, these numbers risk being misleading.
The lack of standardized reporting makes it nearly impossible to gauge AI’s true carbon footprint. Sasha Luccioni, climate lead at Hugging Face, dismisses Altman’s estimate as speculative, pointing out that 84% of large language model usage in May 2025 involved models with zero environmental disclosures. This opacity leaves consumers and regulators in the dark, even as AI adoption skyrockets.
Misinformation spreads easily in this vacuum of data. One widely cited claim suggests ChatGPT requests use 10 times more energy than Google searches, a statistic traced back to an offhand comment by Alphabet’s chairman, John Hennessy. Despite its shaky origins, this comparison has been repeated uncritically in media and policy discussions.
The parallels to other industries are striking. While cars come with fuel efficiency ratings and appliances list energy consumption, AI operates without similar accountability. Luccioni argues that mandatory emissions reporting should be a priority for regulators, especially given the urgency of the climate crisis. Until then, speculative figures and corporate silence will continue to obscure AI’s growing environmental toll.
(Source: Wired)
