As global leaders gather for COP30 (the 2025 United Nations Climate Change Conference, or Conference of the Parties of the UNFCCC) in Belém, Brazil, to tackle climate change, the spotlight is not just on factories and vehicles but also on a less visible source of carbon emissions: our digital lives.
A recent study by TRG Datacenters, a private data centre provider based in Texas, USA, reveals that common internet activities like streaming Netflix or YouTube videos generate high amounts of carbon dioxide (CO₂), a leading driver of global warming.
This hidden climate cost illustrates that the digital footprint created by everyday technology plays a significant role in warming the planet. This study, which calculates and ranks the carbon footprint of 10 different digital applications, was released last week.
It measures energy consumption (in kilowatt-hours, kWh) and associated CO₂ emissions for various popular digital activities to show which leave the biggest environmental footprint.
The research team of TRG Datacenters has quantified how streaming, Zoom video calls, emails, Google search, and artificial intelligence (AI) queries each consume electricity and produce greenhouse gases, highlighting the urgent need for cleaner data centres alongside shifts in user behaviour (see Table: Carbon Footprint of Different Applications).
Ranking Of Digital Behaviour
As per the carbon footprint estimates, watching one hour of HD video on platforms like Netflix or YouTube consumes 0.12 kWh electricity and produces an average of 42 grams of CO₂. The recent TRG Datacenters study has calculated that if someone watches two hours of streaming daily, that adds up to more than 30 kilograms of CO₂ over a year.
Creating a short 6-10 seconds text-to-video clip generates on average 17.5 grams of CO₂ owing to the high computational power needed, ranking it third in carbon emissions among digital activities.
During the COVID pandemic, video calls helped people stay connected with their families and co-workers. TRG Datacenters has estimated that a one-hour Zoom video meeting emits roughly 17 grams of CO₂. Video conferencing requires 0.0486 kWh per hour, as it constantly encodes, transmits, and decodes video and audio in real time. This means that a daily hour-long work standup can create about 6 kilograms of CO₂ over a year.
Meanwhile, sending a basic short email without attachments produces around 4.7 grams of CO₂. Someone sending 50 electronic texts a day would create about 85 kilograms of CO₂ annually just from one email address alone!
The new study has also looked at artificial intelligence (AI) based activities. For instance, producing one AI-created image results in 1 gram of CO₂. Compared to video generation, this is 17 times less, but also 10 times higher than asking an AI chatbot a question, say the researchers.
Similarly, each query asked to AI voice assistants, such as Alexa, Siri or Google Assistant, emits 0.175 grams of CO₂, making voice queries relatively low in carbon footprint. The lower footprint comes from the fact that these systems process relatively simple commands and return quick responses. According to TRG Datacenters, a person needs to ask their voice assistant about 240 questions to match the carbon levels from streaming one hour of video.
According to TRG Datacenters, Google searches and AI chatbot queries share a similar low emission rate of about 0.1 gram of CO₂ per query. Among the lowest impact, generating two Gemini text prompts produces just 0.084 grams of CO₂.
“The tech sector added roughly 900 million tons of CO₂ to the atmosphere in the last year, comparable to the annual emissions of Germany. By the end of 2025, that number is expected to exceed 1.2 billion tons,” said the spokesperson of TRG Datacenters.
“By the end of 2025, that number is expected to exceed 1.2 billion tons. The issue isn't whether we use technology (that's inevitable) but how we power it. Right now, only about 30 per cent of datacentre energy comes from renewables. If we all gather efforts, and get that to 80 per cent or 90 per cent, we would cut the carbon footprint of every digital activity by more than half, without anyone changing their behaviour,” the spokesperson added.
About The Study
The TRG Datacenters has used various data sources for its recent study. These include peer-reviewed studies on digital infrastructure energy use, industry disclosures from hyperscale providers (Google, Microsoft, Open AI, Anthropic, Meta), benchmarking papers, lifecycle carbon inventories for emission factors, and independent analyses and media reporting.
Its calculations include server-side electricity consumed per activity, datacentre overhead via Power Usage Effectiveness, and grid carbon intensity using a global average.
Power Usage Effectiveness (PUE) is a metric for a data centre's energy efficiency, calculated by dividing the total power consumed by the data centre by the power delivered to its IT equipment. A PUE of 1.0 is the ideal, meaning all power is used for computing, but in reality, a PUE will always be greater than 1.0 due to energy used by cooling, lighting, and other infrastructure. A lower PUE value indicates greater efficiency.
The TRG Datacenters team has excluded some factors from the study for comparability. These include — end-user device energy (screen, speakers, laptop power), network transmission (WAN, CDN, local Wi-Fi), and embodied hardware manufacturing and disposal. The researchers inform that these “exclusions are noted but not included to avoid double-counting and to maintain focus on server-side impacts”.
Ecological Footprint of Data Centres
However, data centres have a large ecological footprint too. Artificial intelligence (AI) model training and deployment occur mainly in data centres. Data centres are facilities used to house servers, storage systems, networking equipment and associated components that are installed in racks and organised into rows. This IT equipment, and a range of auxiliary equipment required to keep it in working order, comprise servers, storage systems, networking equipment, cooling and environmental control, uninterruptible power supply (UPS) batteries and backup power generators.
Today, electricity consumption from data centres is estimated to amount to around 415 terawatt hours (TWh), or about 1.5 per cent of global electricity consumption in 2024. It has grown at 12 per cent per year over the last five years, as pointed out by the International Energy Agency (IEA).
The rise of AI is accelerating the deployment of high-performance accelerated servers, leading to greater power density in data centres. According to IEA, global electricity consumption for data centres is projected to double to reach around 945 TWh by 2030 in the Base Case, representing under 3 per cent of total global electricity consumption in 2030.
Data centres are also high in water consumption as they consume large amounts of water to cool their processor chips, so as to avoid overheating and potential damage.
According to the Washington DC-based Environmental and Energy Study Institute, large data centers can consume up to 5 million gallons per day, equivalent to the water use of a town populated by 10,000 to 50,000 people. With larger and new AI-focused data centers, water consumption is increasing alongside energy usage and carbon emissions.
Environmental and Energy Study Institute quotes scientists at the University of California, Riverside, who claim that each 100-word AI prompt is estimated to use roughly one bottle of water (or 519 milliliters). This may not sound like much, but billions of AI users worldwide enter prompts into systems like ChatGPT every minute. Large language models (LLM) require many energy-intensive calculations, necessitating liquid cooling systems. LLMs are AI programmes trained on vast amounts of text data to understand and generate human-like text.
Instead of long hours of Netflix or YouTube streaming, which produce high emissions, people can choose healthier and more eco-friendly activities, which are good for their body and mind too. For instance, spending time outdoors walking, cycling, or gardening use no digital energy, improve physical fitness, and boost mental health.
Offline hobbies like cooking, painting, or handmade crafts nurture creativity without digital electricity use. Despite the convenience of Kindles and e-books, there’s an unmatched magic in holding a book, flipping through its pages, and inhaling the comforting, earthy scent of paper that makes reading a deeply personal and timeless pleasure.
Meeting friends face to face builds deeper bonds through genuine connection, eye contact, and shared moments, which nourishes both our mental health and happiness in ways no screen can replicate. People can also optimise their digital habits. Download content for offline viewing, lower streaming quality, limit video calls, and opt for text or audio over video when possible to reduce emissions.
The future of our planet hinges on the everyday choices we make—how we use technology, how much energy we consume, and how mindful we are of its impact on both the environment and our own well-being.
While technology is an essential part of modern life and progress, we have the power to choose a path that is gentle on the earth and nurturing to our mental and physical health. By embracing sustainable digital habits, supporting cleaner energy for data centres, and balancing screen time with real-world interactions, we can help ensure a healthier planet and a healthier self.
Have you liked the news article?