OpenAI is planning a data center in Georgia carrying a price tag of roughly $20 billion, and according to a report from The Atlanta Journal-Constitution (ajc.com), the project came together quietly in Effingham County, a largely rural area in the southeastern corner of the state near Savannah.
The AJC's framing is the most notable part of the story: the newspaper describes how the plan "secretly took root" in Effingham. Large data center deals are frequently negotiated under code names and non-disclosure agreements, with local development authorities handling land assembly and incentives before the identity of the end user becomes public. The AJC report indicates this project followed a similarly low-profile path before OpenAI was named.
Beyond the location, the company, and the $20 billion figure, the available reporting does not spell out a construction timeline, the facility's power requirements, the number of permanent jobs, or what tax incentives, if any, were attached. Those details typically surface later, through county records and state filings.
Why it matters: $20 billion is an enormous amount of capital to commit to a single site, and it is a concrete measure of how much physical infrastructure the current wave of artificial intelligence actually requires. AI models are not abstractions floating in the cloud; they run on buildings full of chips that consume electricity, land, and water, and those buildings have to go somewhere specific. When they land in a county like Effingham, the consequences are local, in the form of grid demand, property tax bases, water use, and construction traffic, even though the decisions are made in boardrooms far away. The secrecy the AJC highlights also raises a question communities across the country are now confronting: how much should residents know about a project of this scale before it is effectively a done deal?