The Internet’s Invisible Carbon Footprint
- Author: Xavier Harding
- Full Title: The Internet’s Invisible Carbon Footprint
- Category: articles
- Document Tags: #tech
- URL: https://foundation.mozilla.org/en/blog/ai-internet-carbon-footprint/
Highlights
- According to a study by Purdue University, a one-hour Zoom call produces between 150 to 1,000 grams in carbon dioxide. Comparatively, a gas-powered car burning a gallon of gasoline emits 8,887 grams of carbon dioxide, making a Zoom call potentially more green than driving to meet your colleague in person. (View Highlight)
- That same one-hour call also requires between two and 12 liters of water and an iPad Mini’s footprint worth of land mass (those data centers filled with servers for all those Zoom calls have to go somewhere). (View Highlight)
- you can reduce the footprint of your call by 96% by going audio-only (View Highlight)
- an average podcast episode emits 1.16 metric tons of carbon dioxide (that equates to the same amount of harm as around three barrels of oil per episode!). Where does most of that waste come from? Mainly electricity and how it’s produced. Some electricity is made using harmful sources like coal or natural gas. Or it can be created using less harmful sources like wind and solar. IRL Podcast took an average between those sources and arrived at 1.16 metric tons of carbon per episode. (View Highlight)
- According to Netflix in 2020, the company says one hour of streaming video on the service produces 100 grams of carbon dioxide equivalent (CO2e). One person watching four hours of Netflix is like driving a gas-powered car one mile. Netflix can produce around 1.1 million metric tons of carbon dioxide annually — the equivalent of 240,000 passenger cars (View Highlight)
- A study published in January counts annual carbon emissions at around 6.5 million metric tons of CO2e for YouTube. TikTok’s carbon footprint is more than double YouTube’s. TikTok produces around 14.7 million metric tons of CO2e per year. (View Highlight)
- The energy needed to teach an AI model like ChatGPT could power an average American’s home for hundreds of years — this is how Gizmodo sums up a research paper by Stanford’s AI division. Other estimations measure the amount of energy used to train ChatGPT at 1.3 gigawatt hours, or the amount of energy that 120 U.S. homes use in a year. (View Highlight)