Home
5-Second Video Consumes Power of 10 Mobile Phones; AI Industry Devours Global Water and Electricity

The rapid growth of the AI industry is placing unprecedented strain on global energy and water supplies. Earlier today, the head of the National Energy Administration, while presenting the "14th Five-Year Plan" for the new energy system, noted that generating a 5-second high-definition video using AI consumes as much electricity as charging ten smartphones.
Since early February, global energy markets have seen sharp fluctuations, with many countries experiencing varying levels of energy shortages and supply constraints. The steady increase in AI computing power has only worsened this challenge.
It’s not only about electricity—asking AI 10 questions uses as much water as a bottle of mineral water.
The AI industry’s high water consumption is equally concerning. At the 2026 Summer Davos Forum, Wu Xuchu, vice chairman of KPMG China, noted that a 5-minute conversation with AI requires roughly 500 milliliters of cooling water. Experts from the National Geographic Society estimate that asking AI 10 questions uses about the same volume of water as a standard bottle of mineral water.
It's important to note that AI’s water usage extends far beyond data center cooling towers. According to a World Economic Forum report, water consumption in chip manufacturing and power generation is significantly higher than in data center operations themselves. Across the entire AI value chain—from chip fabrication and power supply to computing maintenance—annual freshwater consumption totals 23 billion cubic meters, accounting for 3.7% of global industrial freshwater use. While this percentage may appear modest, freshwater is difficult to transport over long distances, making its scarcity far more acute than electricity, which can be readily redistributed across regions.
40% of data centers are situated in water-scarce regions, worsening the crisis due to geographic mismatches.
Even more alarming is the severe disconnect between industry location and water resource availability. Data indicates that roughly 40% of data centers and nearly a third of chip fabrication plants are located in water-scarce areas. When selecting sites, companies tend to prioritize electricity costs and land prices, relegating water capacity to a secondary factor.
Related article
Apple Safari Preview Integrates MCP Service AI Agent to Streamline Web Development and Debugging
On July 1, the Apple WebKit team introduced the MCP server feature in Safari Technology Preview 247, streamlining front-end development and debugging through AI agent integration.Understanding the MCP ProtocolMCP, or Model Context Protocol, is an ope
Meta AI Agent Muse Secretly Leaks User Address and Arranges Meetings with Buyers, Unaware of the User
Meta’s AI agent Muse, launched in the US on September 22, has recently faced serious allegations of overstepping its boundaries. According to The Guardian, the agent shared a user’s Toronto address with a Facebook Marketplace buyer without authorizat
Amazon Invests 10 Billion Euros in European Logistics, Unveiling AI Robot Proteus for Next Year Launch
On June 5, Amazon hosted its "Future of Delivery" event at the Dartford distribution center in Kent, UK, unveiling an AI-enhanced version of its Proteus warehouse robot. This advanced system, designed to understand and respond to voice commands, repr
Related Special Topic Recommendations
Comments (0)
0/500

The rapid growth of the AI industry is placing unprecedented strain on global energy and water supplies. Earlier today, the head of the National Energy Administration, while presenting the "14th Five-Year Plan" for the new energy system, noted that generating a 5-second high-definition video using AI consumes as much electricity as charging ten smartphones.
Since early February, global energy markets have seen sharp fluctuations, with many countries experiencing varying levels of energy shortages and supply constraints. The steady increase in AI computing power has only worsened this challenge.
It’s not only about electricity—asking AI 10 questions uses as much water as a bottle of mineral water.
The AI industry’s high water consumption is equally concerning. At the 2026 Summer Davos Forum, Wu Xuchu, vice chairman of KPMG China, noted that a 5-minute conversation with AI requires roughly 500 milliliters of cooling water. Experts from the National Geographic Society estimate that asking AI 10 questions uses about the same volume of water as a standard bottle of mineral water.
It's important to note that AI’s water usage extends far beyond data center cooling towers. According to a World Economic Forum report, water consumption in chip manufacturing and power generation is significantly higher than in data center operations themselves. Across the entire AI value chain—from chip fabrication and power supply to computing maintenance—annual freshwater consumption totals 23 billion cubic meters, accounting for 3.7% of global industrial freshwater use. While this percentage may appear modest, freshwater is difficult to transport over long distances, making its scarcity far more acute than electricity, which can be readily redistributed across regions.
40% of data centers are situated in water-scarce regions, worsening the crisis due to geographic mismatches.
Even more alarming is the severe disconnect between industry location and water resource availability. Data indicates that roughly 40% of data centers and nearly a third of chip fabrication plants are located in water-scarce areas. When selecting sites, companies tend to prioritize electricity costs and land prices, relegating water capacity to a secondary factor.
Apple Safari Preview Integrates MCP Service AI Agent to Streamline Web Development and Debugging
On July 1, the Apple WebKit team introduced the MCP server feature in Safari Technology Preview 247, streamlining front-end development and debugging through AI agent integration.Understanding the MCP ProtocolMCP, or Model Context Protocol, is an ope
Meta AI Agent Muse Secretly Leaks User Address and Arranges Meetings with Buyers, Unaware of the User
Meta’s AI agent Muse, launched in the US on September 22, has recently faced serious allegations of overstepping its boundaries. According to The Guardian, the agent shared a user’s Toronto address with a Facebook Marketplace buyer without authorizat
Amazon Invests 10 Billion Euros in European Logistics, Unveiling AI Robot Proteus for Next Year Launch
On June 5, Amazon hosted its "Future of Delivery" event at the Dartford distribution center in Kent, UK, unveiling an AI-enhanced version of its Proteus warehouse robot. This advanced system, designed to understand and respond to voice commands, repr











