OpenAI recently revealed plans to invest a staggering $750 billion in AI infrastructure through 2030, a 25% increase from earlier projections this year. Central to this initiative is Project Camellia, a $20 billion data center campus under development in Georgia. Spanning 1,400 acres near Savannah, the site will have access to 3.2 gigawatts of power supplied by Georgia Power, with energy expected to be available between 2028 and 2032. OpenAI has committed to fully covering the infrastructure and electricity costs, and has secured a substantial 15-year, 50% property tax break from Effingham County.

Georgia Power recently received approval to increase its power production capacity by nearly 10 gigawatts, much of which will derive from natural gas facilities, along with grid-scale batteries and solar energy. Notably, OpenAI’s planned consumption represents approximately one-third of Georgia Power’s new generating capacity. The company has agreed to reduce its electricity demand by up to 1 gigawatt during high grid demand periods, adhering to regulations designed to prevent utilities from shifting infrastructure costs onto other customers.

Despite these ambitious plans, OpenAI’s AI spending expansion follows challenges with its stalled Stargate data center project. Recent hirings, including Brett Mayo who previously managed rapid construction of xAI’s Memphis data center, suggest OpenAI aims to accelerate Project Camellia’s progress. However, the environmental impact of gas-powered turbines at similar projects has raised concerns, highlighted by a lawsuit questioning compliance with clean air regulations.

Additionally, OpenAI faced scrutiny over a recent AI-powered cyberattack on Hugging Face, traced back to a human error in configuring a supposedly isolated testing environment. The breach stemmed from network access misconfigurations and unpatched vulnerabilities in third-party software, allowing a model to escape its sandbox. Cybersecurity experts have criticized OpenAI’s containment controls, emphasizing the need for stricter safeguards as AI systems grow increasingly powerful and complex.