Social media’s environmental footprint is larger than most people realize, stretching from the electricity powering massive data centers to the minerals mined for the phones in your pocket. Every scroll, every video, every ad auction running in the background consumes energy, water, and physical resources. The impact isn’t dramatic on a per-user basis, but scaled across billions of people using these platforms daily, it adds up fast.
The Energy Behind Every Scroll
Social media platforms run on data centers, warehouse-sized facilities packed with servers that store your photos, serve your feeds, and run the algorithms deciding what you see next. Globally, data centers consumed roughly 415 terawatt hours of electricity in 2024, about 1.5% of all electricity used worldwide. That figure is projected to double to around 945 terawatt hours by 2030, pushing data centers to nearly 3% of global electricity demand.
Social media companies share that infrastructure with cloud computing, streaming, and other online services, so isolating their exact slice is difficult. But platforms like TikTok, Instagram, and YouTube are among the most data-intensive services in existence, constantly processing video uploads, running recommendation engines, and delivering personalized content to billions of users simultaneously.
Streaming video offers a useful benchmark. The IEA estimates that one hour of streaming video produces roughly 36 grams of CO2 on average. That sounds small, but consider that TikTok users in many countries average over an hour per day on the app, and YouTube serves over a billion hours of video daily. Those grams multiply into a significant collective output.
AI Recommendations Are Raising the Bill
The shift toward AI-powered content is accelerating energy consumption in ways that weren’t a factor even a few years ago. Social media platforms increasingly rely on generative AI for content recommendations, automated moderation, image filters, and chatbot features. These AI workloads are substantially heavier than traditional computing. A generative AI training cluster consumes roughly seven to eight times more energy than a typical computing workload, according to researchers at MIT.
Even after a model is trained, using it consumes significant power. A single query to a generative AI system uses about five times more electricity than a standard web search. As platforms embed AI more deeply into their features, from auto-generated captions to personalized video summaries, the energy cost per user interaction rises. Researchers expect AI inference (the process of actually running these models for users in real time) to become the dominant source of electricity demand as the models grow larger and more complex.
Water Usage for Cooling
Data centers generate enormous amounts of heat, and most rely on water-based cooling systems to keep servers from overheating. This creates a less visible but significant environmental cost. Researchers at the University of California, Riverside estimated that a single 100-word AI prompt uses roughly 519 milliliters of water, about one standard bottle. Social media platforms processing billions of requests daily translate that into massive water withdrawal, often in regions already facing drought stress.
This water is typically evaporated during the cooling process, meaning it doesn’t return to local water systems. As data center construction booms to meet growing demand, communities near these facilities are increasingly raising concerns about competition for limited water supplies.
The Hidden Cost of Ad Auctions
Advertising is how social media companies make money, and the system powering those ads has its own environmental footprint. Nearly every ad you see on a social platform is placed through programmatic advertising, an automated system where multiple companies bid in real-time auctions to show you a specific ad. Each impression triggers multiple auctions, data calls, tracking pixels, and creative file loads, all of which require server processing and electricity.
Recent estimates suggest that an average digital advertising campaign produces approximately 5.4 metric tons of CO2. In the United States alone, programmatic ad activity may generate roughly 100,000 tons of carbon emissions per month. This is an invisible layer of energy consumption that operates entirely in the background. You never see the dozens of server-to-server transactions happening in the milliseconds before an ad appears in your feed, but each one draws power.
Devices, Mining, and E-Waste
Social media doesn’t just consume energy through the internet. It also drives demand for the physical devices people use to access it. Smartphones are the primary gateway to social platforms, and manufacturing them requires extracting lithium, cobalt, copper, nickel, and other minerals from the earth. Cobalt alone accounts for nearly 64% of the resource depletion impact of lithium-ion batteries, according to research published in Environmental Science & Technology. Copper contributes significantly to both human toxicity and ecological toxicity risks associated with battery production.
Mining these materials causes deforestation, water contamination, and soil degradation in extraction regions, many of which are in lower-income countries with limited environmental regulation. The demand cycle is relentless: social media platforms continuously release features optimized for newer hardware, encouraging more frequent upgrades and shortening the useful life of existing devices.
When those devices are discarded, they become electronic waste. The 2024 Global E-Waste Monitor found that small IT and telecommunications equipment, a category that includes smartphones, laptops, and routers, generated 4.6 million tonnes of e-waste globally. Only 22% of that was formally collected and recycled. The rest ends up in landfills or informal recycling operations, where toxic materials can leach into soil and groundwater.
What Tech Companies Are Promising
The major companies behind social media platforms have set ambitious climate targets. Apple, Google, and Meta have all pledged to reach net-zero carbon emissions by 2030. Amazon set its target for 2040. Microsoft went further, promising to become “carbon negative,” actively removing more CO2 from the atmosphere than it emits, by the end of this decade.
These goals are under serious scrutiny. Researchers have described some of these net-zero commitments as “verging on fantasy,” largely because the explosive growth in AI and data center demand is pushing emissions in the opposite direction. Google’s own environmental reports have shown its emissions rising in recent years despite massive investments in renewable energy. The gap between corporate pledges and operational reality is widening as these companies race to build more AI infrastructure.
Renewable energy purchases are the primary tool companies use to offset data center emissions, but buying renewable energy credits doesn’t always mean a data center runs on clean power in real time. A facility might draw from a coal-heavy grid while the company claims carbon neutrality through credits purchased from a wind farm in another region.
What Your Usage Actually Looks Like
For an individual user, the carbon footprint of social media use is relatively modest compared to driving a car or heating a home. An hour of scrolling video content might produce somewhere in the range of 36 grams of CO2, roughly equivalent to driving a car about 500 feet. The environmental problem with social media isn’t any single person’s usage. It’s the infrastructure required to serve billions of people simultaneously, the mining needed to build billions of devices, and the e-waste generated when those devices are replaced every two to three years.
If you want to reduce your personal contribution, the most impactful steps are keeping your phone longer, reducing video autoplay, and choosing Wi-Fi over mobile data (Wi-Fi transfers data more efficiently). But the meaningful changes sit with the companies building and operating this infrastructure, not with individual scrolling habits.