OpenAI increased its AI infrastructure spending estimate to $750 billion through 2030, up about 25% from earlier figures, according to company announcements made July 22, 2026 [1, 2]. The company unveiled plans for a $20 billion data center campus in Georgia, named Project Camellia, covering 1,400 acres and drawing 3.2 gigawatts of power [1, 2].
The data center campus is expected to come online with available power capacity between 2028 and 2032, marking a significant expansion of OpenAI’s computational resources [1, 2]. Georgia Power will supply the electricity for the facility, with most power sourced from natural gas plants, grid-scale batteries, and solar energy [1].
OpenAI confirmed it "will pay the full cost of the infrastructure and electric-service costs for the new data center," and plans to reduce power draw by up to 1 gigawatt during peak demand periods to aid grid stability [1].
The company secured a 50% property tax abatement for 15 years from Effingham County to support the development at the Georgia site [1]. OpenAI hired Brett (Brent) Mayo, formerly with Elon Musk’s xAI, to lead construction of the data center infrastructure [1, 2].
Project Camellia’s scale and power footprint signal OpenAI’s commitment to expanding AI training and deployment capacity dramatically in the coming years. The new campus represents one of the largest dedicated AI compute facilities announced to date in the United States.
OpenAI’s increased infrastructure investment aims to support growing demand for AI services globally. The company’s enhanced budget and facility plans were revealed alongside the disclosure that overall AI spending targets have risen 25% compared to prior estimates [1, 2]. The Georgia data center is a cornerstone of that expansion and is scheduled to be operational with full power capacity no later than 2032 [1, 2].
