The Digital Revolution and Its Invisible Carbon Footprint
Every morning, millions of Indonesians reach for their mobile phones before even touching a glass of water. Notifications stream in—messages, news, shopping recommendations, entertainment content. In seconds, that data request travels through fibre optic networks, stops at a server in a large building called a data centre, and returns to our screens in the form of pleasing pixels. It all feels clean, wireless, and traceless. In reality, it is not.
Behind that convenience, large machines run non-stop, electricity consumption continues to swell, and piles of used electronic devices await their turn to be discarded. Indonesia now stands at a crossroads: become an advanced digital nation, or become a nation that silently pays the ecological price of that progress.
Data centres are the invisible backbone of digital civilisation. Thousands of servers are lined up inside them, working without pause, storing and processing data from all corners of the world. Their energy needs are startling: according to a report by the International Energy Agency (IEA), global data centre electricity consumption reached 415 terawatt-hours (TWh) in 2023, equivalent to 1.5 percent of total global electricity demand. This figure is projected to soar to 945 TWh by 2030, surpassing Japan’s annual electricity consumption.
This surge is largely driven by the explosion in the use of artificial intelligence (AI). By the end of 2024, AI had accounted for 20 to 24 percent of total electricity use in global data centres, and that figure is expected to rise to 49 percent by the end of 2025. In the Southeast Asian region, including Indonesia, data centre energy demand is projected to exceed 100 percent alongside increased digitalisation and cloud computing investment.
Indonesia itself requires 1.09 gigawatts (GW) of electricity consumption for data centres in 2025. The Ministry of Energy and Mineral Resources (ESDM) estimates this figure will increase fivefold by 2034, to 5.22 GW. Meanwhile, data centre capacity in Indonesia in the first half of 2024 alone had already reached 202 megawatts, with projections for AI-ready data centres soaring to 743 MW in the near future.
The problem is that most of the electricity supplying Indonesian data centres still comes from fossil fuel-fired power plants. Without a serious transition to renewable energy, this digital expansion is directly proportional to the expansion of the national carbon footprint. Indonesia’s internet penetration, which now touches 79.5 percent of the population, should be a momentum for, not a burden on, the energy transition.
There is another dimension of the digital ecological crisis rarely discussed in public: electronic waste. Based on the Global E-waste Monitor 2024 report published by the United Nations Institute for Training and Research (UNITAR), global e-waste production in 2022 reached 62 million tonnes, an 82 percent increase compared to 2010, while the recycling rate was only 22.3 percent.
Indonesia is not merely a spectator in this crisis. The same report places Indonesia as the largest producer of electronic waste in Southeast Asia, with a pile-up reaching 1.9 million tonnes in 2022. Java Island is the largest contributor with 56 percent of the national total, followed by Sumatra with 22 percent. This accumulation is projected to surge dramatically to 4.4 million tonnes by 2030—a figure that should make us pause our scrolling for a moment.
Behind these numbers lies a real health threat. Electronic devices contain lead, mercury, cadmium, and various other hazardous compounds. When disposed of carelessly in final disposal sites without sorting, these substances seep into the soil and water, poisoning the ecosystem and the humans living nearby. Ironically, Indonesia’s national e-waste recycling rate is only 17.4 percent, and its management is still dominated by the informal sector, which does not meet environmental safety standards.
This is not solely a matter of infrastructure. It is a matter of lifestyle choices and policy priorities. The increasingly short mobile phone replacement cycle, driven by the industry’s tempting marketing strategies, creates a continuous flow of e-waste without a clear destination. Consumers upgrade, manufacturers reap profits, and the earth bears the remainder. Artificial intelligence is now hailed as the answer to almost all problems, from city congestion to early cancer detection. Yet there is a price not yet listed in the technology’s promotional brochure: the ecological cost of the intensive computation that powers it.
A comprehensive study from Tsinghua University titled OpenCarbonEval (2024) reveals that AI’s carbon emissions increase proportionally with the computation required to train a model. The larger the model, the greater the energy consumption and the larger its carbon footprint. AI models that implement explicit reasoning processes have been shown to produce significantly higher carbon emissions compared to models with short responses.
It is not just the training process; the inference process—when AI is used to answer questions or perform tasks—also consumes energy continuously, especially when accessed by millions of users daily. Every question we pose to an AI chatbot consumes significantly more energy than a standard internet search.
Paradoxically, AI is also one of the most promising tools for monitoring and mitigating climate change itself, from climate modelling and building energy consumption optimisation to sensor-based precision agriculture. We are using fire to fight fire, and it is not yet clear which is faster.
No single solution is sufficient. What is needed is a systemic shift, from upstream to downstream. At the policy level, the government needs to actively push, not merely suggest, the use of renewable energy for data centres. Steps like those taken by Digital Edge Indonesia,