The cost of AI: 1 million Ghibli images equals water for 2,000 people

Social media is awash with AI-generated Studio Ghibli-style images, from whimsical selfies to reimagined historical scenes. Over 6 million posts using Ghibli-related hashtags have flooded Instagram alone, showcasing just how viral and beloved this aesthetic has become.

At the heart of this explosion is the soaring popularity of tools like ChatGPT, which reportedly added 1 million users in just one hour, according to OpenAI CEO Sam Altman. The hunger for AI-driven content creation has never been greater, fueled by curiosity and the chase for social media clout.

But behind the pastel-colored euphoria lies a stark reality: every generated image consumes real-world resources, including precious water, enough to sustain thousands in drought-stricken regions.

A Topview analysis reveals that generating just 1 million Ghibli-style images consumes approximately 40,000 liters of water, a volume that, according to the United Nations’ minimum water requirement of 20 liters per person per day, could sustain 2,000 people.

According to the report, with 34 million AI images generated daily, annual water use could fill 200 Olympic pools.

This alarming correlation underscores the environmental toll of its relentless demand for computational power and cooling.

Breaking Down the Numbers

AI image generation relies on power-hungry GPUs, which require vast amounts of water for cooling and energy production.

Here’s the breakdown:

Metric	Calculation	Source
Water per image	0.04 liters (Cooling GPUs + electricity generation)	Study by Hugging Face and Carnegie Mellon University
Energy per image	0.003 kWh (25% of a smartphone’s daily charge)	Study by Hugging Face and Carnegie Mellon University
CO2 per 1M images	1.17 kg CO2 × 1M = 1,170 kg Equal to driving 2,925 miles (4707 Km)	EPA Data (400 g CO2 per mile for an average gasoline car​)

Note: Actual figures vary based on factors like model size, data center design, local climate, and use of renewable energy. Efficient, green-powered centers can reduce water and carbon footprints, while fossil-fueled or drought-prone locations may increase them.

Research by Hugging Face and Carnegie Mellon University also highlights that large models can consume up to the energy of one smartphone charge per image, nearly four times the conservative water estimates shown above.

The Human Toll of AI’s creativity

Those tens of thousands of liters, used for cooling or indirectly for generating the required electricity, could have met critical water needs elsewhere, especially in drought-stricken regions of Africa or Southeast Asia, where just 20 liters per person per day can be a lifesaving resource.

Yet, the AI gold rush shows no signs of slowing.

Recall that OpenAI CEO Sam Altman even admitted the surge in demand “melted GPUs,” delaying free-tier access to the tool. He also announced that his team will introduce limits on the ChatGPT Ghibli image-generation feature.

“It’s super fun seeing people love images in ChatGPT. But our GPUs are melting. We are going to temporarily introduce some rate limits while we work on making it more efficient. Hopefully, won’t be long!” he expressed.

Meanwhile, data centers in arid regions like Oregon and Chile have faced community pushback and protests due to heavy water usage, further straining already scarce local supplies.

Studio Ghibli is a legendary Japanese animation studio known for its beautifully crafted and emotionally rich films. Founded in 1985 by Hayao Miyazaki and Isao Takahata, Studio Ghibli has produced masterpieces like Spirited Away, My Neighbor Totoro, and Princess Mononoke.

Based on this, the OpenAI model allows users to transform any image into this signature anime aesthetic, blending soft watercolor textures with expressive character designs.

After the unveiling of the image filter in ChatGPT, users have flooded social media with Ghibli-inspired (anime-style) versions of famous moments. The trend has surged from the iconic Dilwale Dulhania Le Jayenge train scene to Donald Trump’s post-assassination attempt photo.

Hayao Miyazaki, the co-founder of Studio Ghibli, has previously expressed strong disapproval of AI-generated animation. In a 2016 meeting where he was shown an animation demo, Miyazaki said:

“I am utterly disgusted. If you really want to make creepy stuff, you can go ahead and do it. I would never wish to incorporate this technology into my work at all.” He also said, “I strongly feel that this is an insult to life itself.”