The viral ChatGPT 80s trend is turning ordinary selfies into retro portraits across Instagram. The transformation may take only seconds on a screen, but the computing behind a single AI-generated image can consume between 0.01 and 0.29 kilowatt-hours of electricity, depending on the system and generation process.

Instagram feeds have found their latest makeover.

Over the past few days, users have been turning photographs into stylised 1980s portraits using ChatGPT, adding the familiar visual language of the decade to otherwise ordinary selfies. The process feels almost effortless from the user’s side: upload a photograph, wait briefly and share the result.

What happens during that wait is much less visible.

An AI-generated picture has to be processed inside a data centre, where specialised computing hardware performs the calculations required to produce the final image. Those machines use electricity, produce heat and depend on cooling systems to continue operating.

That raises a more difficult question behind the Instagram trend: what does generating just one AI image actually cost the environment?

ChatGPT 80s trend: One AI image can use up to 0.29 kWh of electricity

Research cited from Carnegie Mellon University and Hugging Face estimates that generating an AI image can consume anywhere between 0.01 and 0.29 kilowatt-hours of electricity.

There is no single figure that applies to every image. The amount depends on the AI system being used and the image-generation process itself.

At the lower end, the energy requirement for a single image can appear small. Even the upper end is easy to overlook when the result arrives almost instantly on a phone.

The environmental question changes when one image becomes five, five become millions, and a viral trend encourages users around the world to repeatedly regenerate the same photograph until they get the version they want.

That scale is what makes trends such as the ChatGPT 80s makeover worth examining.

There is an important distinction here. The available figures do not mean that every image created through ChatGPT consumes 0.29 kWh. They describe a range observed across AI image-generation systems and processes. The precise electricity use of an individual 80s-style ChatGPT image cannot be determined from the information available.

The image on your phone is being created somewhere physical

AI can feel unusually detached from physical infrastructure. There is no visible factory, engine or machine operating next to the person generating the picture.

But the infrastructure is real.

Data centres contain processors that handle AI workloads, and those processors require electricity while also producing heat that must be managed. AI’s wider expansion is already increasing pressure on data-centre infrastructure.

The International Energy Agency expects electricity consumption from data centres worldwide to more than double by 2030. The material also cites MIT figures showing that data-centre power requirements in North America increased from 2,688 megawatts at the end of 2022 to 5,341 megawatts a year later.

An Instagram photo trend is only one small part of that much larger picture, but it demonstrates how rapidly individual digital actions can multiply.

What about water use?

Electricity is only one part of the environmental footprint.

Data centres can also require water for cooling. But putting a precise water figure on a single ChatGPT-generated image would be misleading.

Research from the University of California, Riverside cited in the material estimated that training GPT-3 could require around 700,000 litres of freshwater. The researchers also estimated that roughly 20 to 50 ChatGPT questions and answers could correspond to around 500 ml of water use.

Those figures, however, relate to different kinds of AI activity and cannot simply be converted into a fixed water cost for one 80s-style image. Water requirements can vary depending on location, cooling technology, electricity source and workload.

In other words, there is currently no universal amount of water that can be attached to every ChatGPT prompt or AI image.

That caveat matters, particularly when viral discussions reduce AI’s environmental footprint to a single dramatic number.

The environmental cost begins before you generate the picture

The electricity consumed while creating an image is also not the entire story.

AI data centres rely on GPUs, servers, networking equipment and semiconductor components. Producing that hardware requires raw materials, manufacturing and complex supply chains.

The broader footprint therefore stretches beyond the moment a user presses “generate”.

It can include mineral extraction, chip production, transportation, construction of data centres and eventually electronic waste.

That makes measuring the environmental cost of one viral AI portrait far more complicated than simply looking at the electricity consumed during a few seconds of image generation.

So, how damaging is one ChatGPT 80s photo?

On its own, one retro AI portrait is not presented by the available evidence as an environmental disaster.

The more significant issue is repetition at scale.

Viral AI trends are designed, socially if not technically, around participation. One person generates an image, others see it and create their own. Users may also request several versions of the same picture before choosing one to post.

The source material makes this point directly: five versions of a selfie may appear insignificant to an individual user, but repeated across millions of people they represent a much larger amount of computing.

That is perhaps the most useful way to understand the environmental debate surrounding Instagram’s ChatGPT 80s trend.

The problem is not that a digital portrait secretly carries one fixed environmental price. It is that every generation requires physical computing resources, and apparently tiny demands become substantial when repeated often enough.

The 80s aesthetic will eventually give way to another online trend. The data centres, processors, cooling systems and hardware supporting AI generation will still be there.

For users, the most accurate takeaway is also the simplest: an AI image may look weightless on a screen, but creating it is not resource-free.

Also Read: What Is Instagram’s 80s Trend? How to Make It With ChatGPT