Trump promotes AI power agreement: Amazon, Microsoft, and OpenAI commit to covering data center energy costs themselves

On March 5th, it was reported that a new AI energy agreement promoted by U.S. President Donald Trump has recently been signed by several major tech companies. According to the “User Protection Commitment” released by the White House, Amazon, Google, Meta, Microsoft, OpenAI, Oracle, and xAI agree to bear the costs of electricity and infrastructure needed for their AI data centers, avoiding passing these costs onto ordinary electricity consumers.

This commitment is a non-binding agreement, primarily requiring tech companies to “self-build, introduce, or purchase” the energy resources necessary for operating AI data centers, including new power facilities, transmission networks, and long-term power supply. The White House stated that this move aims to alleviate public concerns over the surge in electricity consumption by AI data centers and to ensure that residents’ electricity bills do not increase due to AI industry expansion.

During a roundtable with government and tech industry representatives, Trump said that the public generally believes large data centers will drive up electricity prices, but the government hopes to change this expectation through the agreement. He emphasized that building AI infrastructure requires greater public understanding and that companies will be responsible for the energy demands, rather than ordinary consumers.

In recent years, the demand for AI computing power has grown rapidly, leading to the construction of numerous data centers across the U.S. A report from Harvard Kennedy School previously indicated that the electricity demand from AI data centers is rising quickly, with some areas nearing grid capacity limits. The report predicts that by 2028, AI data centers could account for about 12% of the nation’s total electricity consumption.

Meanwhile, data from the U.S. Energy Information Administration shows that residential energy prices in the U.S. have already increased by about 6% in 2025, with further increases possible in the coming years. Against this background, the relationship between electricity prices and AI infrastructure expansion has become a key policy discussion topic.

According to the commitment, participating companies also agree to bear the electricity costs after connecting to the grid, even for power that is not actually used. Additionally, they will promote local employment and skills training, and, if necessary, connect backup generators to the grid to support during periods of electricity shortages.

However, it remains unclear how this commitment will be monitored and enforced. The White House has not yet announced specific regulatory frameworks or measures to address non-compliance. As the demand for AI computing power continues to grow, the energy costs of data centers and grid pressures are expected to remain important issues in U.S. tech policy.