Project Camellia

Effingham County, GA

Project details

Capacity 2400 MW reported. Status planned. City . County Effingham. State GA. Grid .

Overview

OpenAI’s Project Camellia is a $20 billion AI data-center campus in Effingham County, Georgia, inland from Savannah. OpenAI announced the project on July 22, 2026. Effingham County publishes an official Project Camellia resource page that identifies the proposal as an OpenAI data center and collects county documents, work-session materials, and notices. The Project Camellia site says OpenAI is designing and developing long-term AI infrastructure in the county, is contracting with Georgia Power for 3.2 gigawatts of power expected in phases between 2028 and 2032, will pay the full energy costs to serve the project, will keep ongoing water use low with closed-loop cooling, and will report publicly through an annual independent audit. Beyond Tomorrow reported the $20 billion capital commitment, an $80 million community-benefits package directed toward schools, public safety, health care, workforce training, housing, veterans’ services, and small-business support, and about 400 permanent on-site jobs when the campus opens around 2028, rising toward 1,000 by 2032. Effingham County said the campus would use treated surface water only, not groundwater or the Floridan aquifer, plans no diesel backup generators after the project is complete, and includes a minimum 300-foot vegetative buffer along neighboring non-industrial properties. OpenAI committed to perimeter noise-monitoring equipment and regular noise-data updates. County materials project more than $43 million in annual local sales-tax revenue across SPLOST, TSPLOST, LOST, and ESPLOST by 2032, separate from property taxes and the $80 million community investment. The Board of Commissioners held a Project Camellia work session on August 29.

Other Georgia campuses include QTS Blakely in Early County, Project Arrowhead in Irwin County, and Pine Ridge Technology Center in Twiggs County.

The public map lists three planned map records: Building 1, Building 2, and Building 3. Campus capacity is 2,400 MW. All three mapped halls are planned.

Map records

Map entries are not a physical building count; campus and building scope may differ across sources.

Related

Sources