What’s the Environmental Cost of Your Online Activities? From AI Queries to Streaming Video
by bernt & torsten
Scrolling through social media, searching on Google, or chatting with an AI feels like a small, effortless action. But behind every click, there’s a hidden environmental cost. These activities consume water and energy and produce carbon emissions. While a single user’s impact might seem negligible, multiply it by millions - or even billions - and the numbers become astronomical.
What Do We Mean by Water Usage and Carbon Emissions?
Before diving into the data, let’s clarify what these terms mean:
-
Water Usage refers to the amount of water used to cool the servers in massive data centers. These centers run 24/7 to process searches, stream videos, and store photos. To prevent overheating, water is often used in cooling systems, either evaporated or circulated to remove heat.
-
Carbon Emissions are the greenhouse gases released into the atmosphere from the electricity used to power data centers, transmission networks, and user devices. Most electricity still comes from burning fossil fuels, which produces CO₂ (carbon dioxide), a major contributor to climate change.
Environmental Impact of Everyday Online Activities
Below is a table summarizing the resources used by one person’s action for each activity.
Activity | Water Usage (Per User) | Carbon Emissions (Per User) | What’s Happening Behind the Scenes |
---|---|---|---|
Google Search | ~0.003–0.005 liters | ~0.2 g CO₂ | Search engines use data centers to process and retrieve results in milliseconds. |
AI Chat (e.g., ChatGPT) | ~0.02–0.05 liters | ~2–4 g CO₂ | AI models require more computational power than standard searches, increasing energy use. |
Uploading a Photo (1 MB) | ~0.01 liters | ~1 g CO₂ | Storing and transferring photos involves multiple data center processes and energy use. |
Uploading a Video (100 MB) | ~1–2 liters | ~40–60 g CO₂ | Videos are large files, requiring more energy for upload, storage, and streaming. |
Streaming Video (1 Hour, HD) | ~40–60 liters | ~300–400 g CO₂ | High-definition streaming relies on constant data flow from servers to users. |
Posting on Social Media | ~0.01–0.02 liters | ~0.5 g CO₂ | Sharing a post involves minimal energy use for transferring small amounts of data. |
Streaming Music (1 Hour) | ~10–15 liters | ~80–100 g CO₂ | Streaming music uses fewer resources than video but still consumes energy for data transfer. |
Video Chat (1 Hour) | ~20–40 liters | ~150–300 g CO₂ | Real-time video calls require continuous data upload and download, adding strain to data center infrastructure. |
Sending an Email (No Attachment) | ~0.002 liters | ~0.2–0.5 g CO₂ | Emails are small text files, but their storage and transmission require data center resources. |
Sending an Email (With Attachment) | ~0.02–0.1 liters | ~5–50 g CO₂ | Large attachments increase storage and transfer energy demands, especially for cloud backups. |
Notes:
-
Google Search: Each search query utilizes data center resources, resulting in minor water usage and carbon emissions. mdpi.com
-
AI Chat (e.g., ChatGPT): Engaging with AI models involves significant computational processes, leading to higher energy consumption. theguardian.com
-
Uploading Media: The environmental impact increases with file size due to greater data transfer and storage demands.
-
Streaming Video and Music: Higher-quality streams (e.g., HD video) consume more resources than lower-quality or audio-only streams. thesun.ie
-
Video Chat: Continuous data transmission in video calls leads to higher energy use than audio-only communication.
-
Email: While sending a simple email may seem harmless, the environmental impact increases with factors such as storage, transmission, and attachments. Emails require energy for routing through servers and data centers, even without attachments. Larger attachments (e.g., images and documents) demand more resources for storage and transfer, especially when backed up to the cloud. Over time, the accumulation of emails - both sent and stored - adds up, contributing to water usage and carbon emissions in ways we often overlook.
The Bigger Picture: What Happens When Millions of Us Do This?
Activity (Global Scale) | Water Usage (Daily) | Carbon Emissions (Daily) |
---|---|---|
Google Searches (5.6 billion/day) | ~16.8–28 million liters | ~1,120 metric tons of CO₂ |
AI Queries (100 million/day) | ~2–5 million liters | ~200–400 metric tons of CO₂ |
Streaming Video (1 billion hours/day) | ~40–60 billion liters | ~300,000–400,000 metric tons of CO₂ |
Video Chat (500 million hours/day) | ~10–20 billion liters | ~75,000–150,000 metric tons of CO₂ |
Sending Emails (333 billion/day) | ~0.67–3.3 million liters | ~66,600–666,000 metric tons of CO₂ |
Note on Email:
With 333 billion emails sent daily worldwide, even simple emails contribute significantly to water usage and carbon emissions, especially when factoring in the storage and transmission required by email servers and data centers. Attachments increase energy demand, further increasing the overall environmental footprint.
Behind the Scenes – The Infrastructure Impact
Backend Activity | Water Usage (Liters per hour per user) | Carbon Emissions (g CO₂ per hour per user) | Explanation |
---|---|---|---|
Search Engine Operations | ~1-10L per 1000 searches | ~0.2-1g CO₂ per search | Constant website indexing, ranking algorithms, and data storage require significant computing power. |
AI Model Training | 500,000 - 5 million L per model | 300 - 1,500 tons CO₂ per model | Training large AI models consumes vast amounts of water and electricity, making it one of the most resource-intensive digital processes. |
Video Storage (e.g., YouTube) | ~2L per GB stored | ~50-200g CO₂ per hour streamed | Storing and streaming large video files require ongoing data center resources. |
Image Storage (e.g., Facebook, Instagram) | ~0.5L per image uploaded | ~20-50g CO₂ per 1000 images | While images are smaller than videos, the volume of daily uploads leads to significant resource use. |
Chat Services (e.g., ChatGPT, WhatsApp, Messenger) | ~1-5L per 1000 messages | ~10-50g CO₂ per 1000 messages | AI-driven chatbots require continuous processing, while messaging apps store vast amounts of text, images, and video files. |
Video Chat (e.g., Zoom, Google Meet) - 480p | ~2L per hour per user | ~150-300g CO₂ per hour per user | Lower-resolution video calls reduce energy consumption. |
Video Chat (e.g., Zoom, Google Meet) - 720p | ~6L per hour per user | ~400-600g CO₂ per hour per user | HD video calls increase data transmission rates and processing. |
Video Chat (e.g., Zoom, Google Meet) - 1080p | ~10-12L per hour per user | ~800-1,000g CO₂ per hour per user | Full HD calls require significantly more bandwidth and computing power. |
Video Streaming (e.g., YouTube, Netflix) - 480p | ~2L per hour per user | ~50-150g CO₂ per hour per user | Lower-resolution streaming is much less demanding on data centers. |
Video Streaming (e.g., YouTube, Netflix) - 720p | ~6L per hour per user | ~300-500g CO₂ per hour per user | HD streaming significantly increases energy consumption. |
Video Streaming (e.g., YouTube, Netflix) - 1080p | ~10-12L per hour per user | ~700-1,200g CO₂ per hour per user | Full HD streaming strains data centers and network infrastructure. |
Video Streaming (e.g., YouTube, Netflix) - 4K | ~20-30L per hour per user | ~2,000-4,000g CO₂ per hour per user | Ultra HD requires 4x more data than 1080p, leading to extreme resource use. |
Email (without attachment) | ~0.2L per 1000 emails | ~0.3-1g CO₂ per email | Basic emails require minimal resources, but storing millions of emails adds up. |
Email (with attachment) | ~1-5L per 1000 emails | ~10-50g CO₂ per email | Large attachments significantly increase storage and processing energy demands. |
Email (with large attachments and cloud storage backups) | ~5-10L per 1000 emails | ~50-100g CO₂ per email | High-resolution files and cloud storage replication consume more data center resources. |
Key Takeaways
- One User vs. Billions: A single Google search or AI chat may seem minor, but when billions of people do the same thing daily, the environmental impact skyrockets.
- Streaming is the Biggest Culprit: HD video streaming consumes far more energy and water than other digital activities due to the high data flow.
- The Hidden Cost of Infrastructure: Beyond user actions, the backend operations—like AI training and web crawling—are even more resource-intensive.
- Video Chat’s Heavy Footprint: Remote meetings are essential, but large-scale video calls require vast water and energy.
- Small Changes Can Help: Lowering video resolution, opting for audio-only calls, and reducing unnecessary uploads can make a real difference.
Do Small Changes Matter?
Our actions online may seem small, but when multiplied by billions of users worldwide, the environmental impact is immense. Our digital habits carry real-world consequences, from the water needed to cool data centers to the carbon emitted by servers.
It might feel like your contribution is a drop in the ocean, but when multiplied by millions or billions of users, these changes become astronomical. For example:
- Streaming video in SD instead of HD: If just 10% of global users switched, it could save billions of litres of water and tons of CO₂ annually.
- Turning off your video during one weekly meeting: This simple action could prevent thousands of litres of water use over a year across an organization.
What can we do? Small changes, like reducing video streaming quality or limiting unnecessary uploads, can make a difference. By being mindful of how we use technology, we can collectively reduce our digital footprint and lessen our environmental impact.
How to Reduce Your Digital Carbon Footprint
Here are some practical tips to help reduce your digital carbon footprint without sacrificing convenience:
1. Be Smart About Streaming
✅ Lower the resolution – Watching in 480p instead of 1080p can reduce CO₂ emissions by 70%.
✅ Download instead of streaming – Downloading saves bandwidth and energy if you're rewatching something.
✅ Turn off autoplay – This prevents unnecessary streaming of content you don’t intend to watch.
✅ Use Wi-Fi instead of mobile data – Mobile networks are much less efficient than wired connections.
2. Optimize Video Calls
✅ Use audio-only mode when possible – It cuts emissions by up to 90% compared to video.
✅ Reduce resolution – 720p or even 480p is often good enough for meetings.
✅ Turn off your camera when not needed – Every minute of video uses energy.
3. Search Efficiently
✅ Use bookmarks – Instead of searching for the same website multiple times, save frequently visited pages.
✅ Use direct URLs – Typing a site’s address (e.g., youtube.com) instead of searching saves energy.
✅ Use eco-friendly search engines – Some alternatives, like Ecosia, use revenue to plant trees.
4. Be Mindful of Cloud Storage
✅ Delete old files – Unused files in cloud storage take up server space, requiring constant cooling and power.
✅ Compress images and videos – Smaller file sizes reduce data center storage needs.
✅ Use local storage for important files – Instead of keeping everything in the cloud, store essential files on your device.
5. Reduce Social Media’s Impact
✅ Limit video autoplay – Many platforms automatically play videos, which drains energy.
✅ Upload fewer unnecessary images & videos – Each upload adds to data center processing.
✅ Use dark mode – It reduces screen power consumption, especially on OLED devices.
6. Be Aware of AI Usage
✅ Use AI chatbots wisely – Each query uses energy equivalent to a Google search multiplied by 10.
✅ Refine your questions – The more precise your query, the fewer resources wasted on unnecessary processing.
7. Upgrade to Energy-Efficient Devices
✅ Choose efficient hardware – Newer devices are often optimized for power savings.
✅ Extend device lifespan – Keeping a phone or laptop for one extra year reduces its overall carbon footprint by up to 30%.
8. Reduce Email’s Environmental Impact
✅ Unsubscribe from newsletters you don’t read – Every spam or unread email still requires storage and energy.
✅ Delete old emails – Emails, especially with attachments, take up data center space and require ongoing power.
✅ Send lighter emails – Avoid large attachments when possible; use compressed files or cloud links instead.
✅ Use efficient subject lines – Clear, specific subjects help recipients find emails easily without extra searches.
✅ Batch your emails – Instead of multiple short emails, try to consolidate information into fewer messages.
The Human Brain and Its Role in Creativity, Memory, and AI
The human brain is a marvel of evolution, a universe within, teeming with thoughts, memories,...
What’s the Environmental Cost of Your Online Activities? From AI Queries to Streaming Video
Scrolling through social media, searching on Google, or chatting with an AI feels like a...
A Poem: Too Old to Be Seen, Too Wise to Be Broken
They saw my years, not the code I wrote,
Dismissed the knowledge I carried, remote.