Data centers powering AI models are massive water consumers, with a new study led by Shaolei Ren, an associate professor at the University of California, Riverside, indicating that U.S. data centers could require additional water capacity comparable to New York City's daily supply by 2030. The research, publicly available on arXiv, highlights that limited public water capacity is becoming a critical bottleneck for data center growth.

Data centers rely heavily on water-intensive liquid cooling techniques to prevent overheating from dense concentrations of servers and IT infrastructure. While companies often claim to use "closed-loop" systems, even these can consume vast amounts of water due to reliance on evaporative cooling towers, potentially exceeding 1 million gallons per day during peak demand, and up to 8 million gallons for some planned facilities.

The study projects that if current water use intensity persists, U.S. data centers will need between 697 million and 1.45 billion gallons per day of new peak water capacity by 2030, with construction costs ranging from $10 billion to $58 billion. A significant portion of this financial burden is expected to fall on the communities hosting these data centers.

Ren and his colleagues propose requiring data centers to report their peak water demand, not just annual usage, and fostering corporate-community partnerships to fund infrastructure upgrades to prevent residents from shouldering the entire cost. Insufficient water capacity could directly impact the feasibility, efficiency, and operational reliability of data center projects, leading to increased costs, delays, and scalebacks.

👉 Read the full article on Gizmodo

👉 Read the full article on Gizmodo

Netflix has partnered with Ben Affleck and acquired his AI tech company, InterPositive, to encourage more filmmakers to utilize AI tools in their productions. Affleck, who quietly founded InterPositive in 2022 to develop AI-powered tools for filmmakers, will join Netflix as a senior adviser following the acquisition.

Affleck explained his motivation for founding InterPositive stemmed from observing early AI models falling short in film production. He aimed to build a workflow that captures on-set realities, using vocabulary familiar to cinematographers and directors, and providing the consistency and controls they expect.

He emphasized that the tool is "not about text prompting or generating something from nothing." Instead, filmmakers can build their own models using their movie's footage, then apply it in post-production for tasks like removing stunt wires, creating missing shots, or adjusting backdrops, colors, and lighting.

Netflix had previously announced plans to expand its use of AI, stating in a letter to shareholders that it aims to "empower creators with a broad set of GenAI tools to help them achieve their visions." This new partnership signals a further integration of AI into Netflix productions.

👉 Read the full article on Gizmodo