Building artificial intelligence requires massive amounts of digital heavy lifting, and that work takes a physical toll on the electricity and water supplies in the places where it happens.
California Governor Gavin Newsom has signed a series of new laws that change how AI companies operate their data centers, which are the giant buildings filled with thousands of computers used to train and run AI models. These buildings require intense cooling systems that consume massive amounts of water and constant electricity. The new rules force these companies to pay for upgrades to local power grids and water systems, rather than passing those utility costs onto regular residents. Additionally, developers must now share detailed plans about how much water they will use and how they plan to manage energy efficiency before they can get approval to build.
The hidden costs of the cloud
AI models are trained on massive amounts of data, which requires a process that relies on specialized chips working around the clock. These chips run hot and need constant cooling to function, which is why data centers use enormous quantities of water. Because these centers need so much consistent electricity to stay powered, they can strain local energy grids, potentially causing price spikes for the surrounding community. Up until now, these facilities often benefited from existing infrastructure, but these laws treat data centers as high-demand industrial users that must help pay for the necessary upgrades to the grid they are leaning on.
This reflects a growing tension between the rapid expansion of AI and the physical limits of our environment. For years, the impact of AI was treated as a purely digital concern, but as the scale of these models has ballooned, the physical resource demands have become unavoidable. By requiring companies to internalize the cost of their energy and water usage, California is signaling that the era of treating AI infrastructure as a free rider on public utilities is ending. It raises a larger question for the rest of the country: as we choose to integrate AI into more parts of our lives, who should be responsible for the massive hardware infrastructure required to keep it running?
Liked this one? The next lands at breakfast.
Every story in tomorrow's AI news, rebuilt in plain English — five minutes, sources linked, free forever.
By joining you agree to receive Article's daily newsletter — unsubscribe in one click. Privacy