OpenAI's $20B Georgia data center: what it means for developers and the Savannah region

OpenAI plans to build a $20 billion data center campus in Effingham County, Georgia, near Savannah, with phased opening starting in 2028. The project, known as Project Camellia, will deliver 3.2 gigawatts of power and create up to 1,000 jobs by 2032. For AI builders, this signals a major expansion of compute infrastructure on the U.S. East Coast, with implications for power availability, regional policy, and supply chain planning.

What happened

OpenAI intends to build a 1,400-acre data center campus in an industrial park in Effingham County, about a 30-minute drive from Savannah's downtown. The facility is projected to employ 400 workers in six-figure jobs when it opens in 2028 and grow to roughly 1,000 employees by 2032. The Effingham County Industrial Development Authority approved a 50% property tax abatement for 15 years for the project. Even with that tax break, OpenAI is projected to become the county's largest taxpayer.

The project is part of a broader data center boom in Georgia. Metro Atlanta has become the fastest growing market for server farms in the U.S., and this coastal expansion positions Savannah as a new node in that network. Local officials describe it as one of the largest private economic development investments in coastal Georgia history.

Why AI builders should care

For teams building AI products, infrastructure scale directly affects model training costs, inference latency, and service reliability. OpenAI's $20B commitment signals that the company expects sustained demand for compute capacity well beyond current levels. The 3.2 GW power requirement is roughly equivalent to the output of three large nuclear reactors, highlighting the enormous energy needs of frontier AI workloads.

Georgia's data center expansion also influences regional power planning and industrial policy. Builders who rely on OpenAI's API or cloud services may see indirect effects: if power constraints delay infrastructure, it could affect capacity availability or pricing. Conversely, a well-executed buildout could improve East Coast latency for users in the Southeast and Mid-Atlantic.

The project is the first salvo in OpenAI's broader $750B infrastructure spending plan, which includes partnerships with Nvidia for up to $100B in data center investments. For developers, this means OpenAI is betting heavily on owning its compute stack rather than relying solely on cloud providers.

Practical implications

Power delivery is scheduled in phases from 2028 to 2032. The 3.2 GW capacity will be built out over several years, so the timeline for additional compute availability is gradual. OpenAI has not disclosed specific power sourcing plans, but the project will draw from the regional grid and may include on-site generation or renewable energy agreements.

The tax abatement reduces OpenAI's operating costs but also affects local government revenue. County officials project that the project could eliminate homestead property taxes for residents and boost school funding. However, data centers also raise concerns about water usage, land use changes, and strain on local infrastructure.

For AI builders, the key takeaway is that large-scale compute infrastructure is moving beyond traditional hubs like Northern Virginia. Coastal Georgia offers available land, power capacity, and incentives, making it a location to watch for future capacity planning.

Caveats

All project details are preliminary and subject to change. The announcement comes from local news sources and OpenAI's own statements; final scope, timelines, and hiring numbers may shift as permitting, construction, and negotiations proceed. The 3.2 GW figure is reported by multiple outlets but has not been independently confirmed by regulatory filings. Power sourcing and water usage plans have not been detailed. Builders should treat this as a directional signal rather than a firm timeline.