.png){kind=logo}
AI-generated images consume electricity, with usage varying by model and generation process.
Millions of image generations can significantly increase AI’s overall environmental footprint.
Higher resolutions can require substantially more computing energy and resources.
The latest AI-powered nostalgia trend is taking social media back to the 1980s. Users are turning ordinary selfies into retro-style portraits, complete with period clothing, hairstyles, lighting, and film effects, using AI tools such as ChatGPT and Gemini. The trend has spread rapidly, with celebrities and public figures joining in.
What appears to be a simple digital makeover, however, depends on physical infrastructure, including GPUs, data centers, electricity, and cooling systems. Each AI-generated image carries an environmental cost, and the impact grows when millions of people join the same trend.
Generating an AI image can consume between 0.01 and 0.29 kilowatt-hours of electricity, depending on the model and generation process, according to research cited by Hindustan Times.
For an individual user, that amount may appear small. The concern lies in the sheer number of images being generated. Users rarely stop after producing a single image. They may generate several versions, change prompts, request different styles, or repeatedly regenerate an image before settling on one they want to share.
When you multiply those individual requests across millions of users, the energy demand becomes much harder to ignore.
The United Nations University estimates that generating one AI image can require roughly 1,450 times the energy of a basic text-classification task. Its research also says AI’s day-to-day inference workloads account for roughly 80–90% of total AI energy demand.
That distinction matters since AI’s environmental footprint isn’t limited to training increasingly powerful models. Every time users interact with those models and request new content, computing resources are required to process the request.
Also Read: Instagram’s Viral ‘80s Trend is Everywhere! Here’s How You Can Try
However, this is not the whole story. Water use by AI facilities could include fresh water for cooling, along with indirect use through energy generation and semiconductor production.
According to UNU’s 2026 report, AI facilities worldwide could require 945 TWh of electricity by 2030. At the same time, their water consumption could match the annual domestic needs of 1.3 billion people in sub-Saharan Africa.
This helps explain why the conversation around the environmental impact of generative AI is growing. The question isn’t simply how much electricity a single AI use would take, but how quickly demand grows and what infrastructure will be needed to meet it.
The higher computational demands of image generation can also contribute to higher resource use.
The quality of an AI-generated image can also affect its environmental footprint.
One recent report notes that doubling image resolution can increase computing energy requirements by 1.3 to 4.7 times.
For users participating in the 80s image trend, this matters when an image is repeatedly regenerated or produced at increasingly high resolutions. A user may not notice the additional computing happening behind the scenes, but higher-quality output can require substantially more processing.
The difference becomes especially relevant when large numbers of users create multiple high-resolution versions of the same type of image.
Also Read: Top Generative AI Trends to Watch in 2027
The more immediate lesson is that AI-generated content carries weight regardless of its appearance on a smartphone screen. Behind every generated portrait is an infrastructure network involving computing power, electricity, cooling systems, and hardware.
Users can reduce unnecessary demand by avoiding repeated generations, limiting iterations, and choosing lower resolutions when high resolution isn’t required. But individual choices are only one part of the issue.
AI companies and the infrastructure supporting these services bear more of the responsibility. Improving model efficiency, increasing the use of cleaner electricity, and developing less water-intensive cooling systems could reduce the environmental burden as AI adoption continues to expand.
The 80s trend may take only seconds to produce a nostalgic portrait. The infrastructure required to deliver those seconds, however, carries a considerably larger resource cost.
As AI-generated images become a routine part of social media culture, that hidden cost is worth considering before the next trend takes off.
How much electricity does generating an AI image consume?
Generating an AI image can consume between 0.01 and 0.29 kilowatt-hours, depending on the model and generation process.
Why does AI image generation affect the environment?
AI image generation requires computing power, electricity, and cooling, while data centers and hardware manufacturing also create environmental impacts.
How much energy does one AI image require compared with text?
Generating one AI image can require roughly 1,450 times the energy of a basic text-classification task.
Does higher image resolution increase AI’s environmental impact?
Yes. Doubling image resolution can increase computing energy requirements by 1.3 to 4.7 times, according to one report.
What can users do to reduce AI image generation’s environmental impact?
Users can avoid unnecessary generations, limit repeated iterations, and choose lower resolutions when high-quality images are not required.
