OpenAI has reportedly committed $300 billion to Oracle for 4.5 gigawatts of cloud data center power by 2027, signaling a massive growth in AI infrastructure.

The staggering new cost of generative AI's increasing demand for data center computing power and electricity is $300 billion. According to reports, OpenAI, the company behind ChatGPT, has agreed to invest that amount with Oracle over five years to maintain its early success. On Wednesday, The Wall Street Journal and The New York Times both revealed the transaction, which would rank among the biggest contracts for cloud computing.

According to the Journal, the contract with Oracle would require the delivery of up to 4.5 gigawatts of electrical capacity, which is approximately equivalent to the capacity of two Hoover Dams or the amount used by four million residences. According to the reports, the agreement is expected to take effect in 2027.

The number of data centers in the United States almost doubled between 2021 and 2024, according to research studies, and the need for additional data centers is expected to increase by 9% annually until 2030. By 2035, the electricity consumption of such facilities is predicted to double from its current level. OpenAI used to only use Microsoft Azure, but it has started to expand its range of cloud services.

The establishment of the Stargate Project, a new business that will invest $500 billion over four years to construct AI infrastructure, or data centers, with partners including Oracle, Microsoft, Nvidia, and Softbank, was announced by OpenAI in January. Beyond the 10 gigawatts pledged in the January launch, OpenAI and Oracle stated in July that they would work together to generate 4.5 gigawatts of data center capacity for Stargate.