Suddenly, and thanks to FOMO or Fear of Missing Out, the 1980s are back on our digital screens. Big hair, bigger sunglasses, colourful shirts, dramatic poses and that unmistakable Bollywood-era glow are flooding Instagram, Facebook and X. No time machine required — just a selfie, a prompt and a few seconds of AI magic.
From Bollywood actors and influencers to politicians and ordinary social-media users, everyone seems to be taking a quick trip to the 1980s. Upload a photograph, ask AI to make it retro, and within seconds you have feathered hair, vintage clothes, grainy film textures and a version of yourself straight out of an old Bollywood poster.
Harmless fun? Perhaps. But the infrastructure powering this nostalgia is anything but weightless.
A June 2026 United Nations report puts the scale of the larger AI challenge in stark terms. Data centres powering AI could consume 945 terawatt-hours of electricity annually by 2030 — nearly three times the combined annual electricity use of Pakistan, Bangladesh and Nigeria, countries home to more than 650 million people. And electricity is only part of the footprint. Every unit of power consumed by data centres carries associated water and land footprints, through cooling, energy production, power infrastructure and supply chains.
The UN University research behind the report estimates that AI's associated water footprint could, by 2030, equal the basic annual domestic water needs of 1.3 billion people in sub-Saharan Africa, while its land footprint could exceed 14,500 sq km. It also challenges the way AI's environmental cost is commonly understood: while attention has focused on the energy required to train large models, day-to-day usage or inference accounts for roughly 80–90% of total AI energy demand.
The scale of everyday use is staggering. One widely used AI service is estimated to process around 2.5 billion prompts a day. Energy demand also varies dramatically by task: an AI-generated image can require around 1,450 times the energy of basic text classification, while video generation is considerably more resource-intensive.
The water story is equally sobering. Research from the University of California, Riverside, cited in the material reviewed for this article, estimated that Microsoft's training of GPT-3 consumed approximately 700,000 litres of freshwater. The researchers also estimated that a conversation involving 20–50 questions and answers could correspond to roughly 500 ml of water consumption. These are estimates, not universal figures: actual consumption varies with location, cooling technology, electricity source, season and workload.
The physical infrastructure is expanding too. MIT has reported that North American data-centre power requirements rose from 2,688 megawatts at the end of 2022 to 5,341 megawatts a year later. And the environmental footprint begins before the first prompt: GPUs, servers and networking equipment require mining, minerals, semiconductor manufacturing and transportation, eventually creating electronic waste.
This is where the current 1980s trend becomes more than a social-media curiosity.
A Firstpost report on the environmental impact of the viral AI-photo trend, citing research from Carnegie Mellon University and Hugging Face, estimates that generating a single AI image can consume 0.01–0.29 kWh of electricity, depending on the model, hardware and process. One image may seem insignificant. Millions of users generating multiple versions are another matter.
The timing is difficult to ignore. While our screens are recreating the 1980s, South Asia is confronting increasingly destructive floods and extreme weather. Nepal's recent disaster affected around 84,000 people, prompted a $49.6-million UN emergency appeal, and generated an estimated 2.2 million tonnes of debris, according to a preliminary UNDP assessment.
The answer is not to reject AI. It is to build a more sustainable one.
A recent ScienceDirect review, “The green paradox: The climate, environmental, and sustainability implications of artificial intelligence,” argues that efficiency alone may not be enough. It calls for low-carbon data-centre infrastructure, transparent environmental reporting, responsibly sourced hardware and circular-economy approaches. That means renewable energy, energy-efficient chips and data centres, water-efficient or closed-loop cooling, waste-heat recovery, and hardware designed for repair, disassembly, reuse and recycling.
This is the emerging idea of Green AI: making environmental efficiency a design principle rather than a corrective measure after deployment.
For India, where digital adoption and data-centre capacity are expanding rapidly with the current capacity of about 1.8 gigawatt, this is no longer a niche environmental discussion. AI policy, infrastructure planning and sustainability policy will increasingly intersect.
So, the next time we press “Create again”, perhaps the question is not whether we should enjoy the trend, but whether every digital impulse needs another computational cycle.
The 1980s may be making a comeback.
Our approach to AI cannot afford to leave the planet behind.
