When people ask, “Who pays the electricity bill for AI data centers?” the simple answer is: first, the tech companies do. In practice, the most direct payer is usually the data-center operator or its parent company, which buys electricity from a utility or through long-term power deals and may also pay for new grid connections or even new power plants. This matters because AI is making data centers much bigger energy users. The International Energy Agency said data centers used about 415 terawatt-hours of electricity in 2024, and that number could rise to around 945 terawatt-hours by 2030. In the United States, data centers are expected to account for nearly half of electricity demand growth by 2030. (iea.org)

But that is not the end of the story. If power becomes more expensive, part of the cost can move downstream. Gartner warned that tighter power supplies can raise electricity prices for data centers and that these higher costs are likely to be passed on to AI and cloud service providers. In other words, businesses and users who pay for cloud computing, AI tools, or online services may indirectly pay part of the bill through higher prices. (gartner.com)

There is also a political fight over whether ordinary households should pay too. In parts of the U.S., regulators worry that if utilities build new transmission lines and power supplies for giant data centers, the cost could be spread across all customers. PJM, the grid operator in the Mid-Atlantic, said in March 2026 that large-load growth is already affecting both grid reliability and electricity costs, and it proposed a one-time backstop power procurement that would be paid for by large loads. Meanwhile, on March 4, 2026, the White House announced a “Ratepayer Protection Pledge” in which seven major tech and AI companies said they would pay for the full cost of the energy and infrastructure needed for their data centers. Some states are also creating special “large load” tariffs so very large customers, including many data centers, must pay minimum demand charges and long-term contract costs. (insidelines.pjm.com)

So, who pays? Officially, the bill should be paid by the companies building and using AI data centers. In reality, the final cost can be shared among tech firms, cloud customers, and—if regulators are not careful—everyone else on the grid. (insidelines.pjm.com)