OpenAI is building its next major AI data center in rural Georgia. The company announced Project Camellia in Effingham County, according to OpenAI, pairing the new site with commitments to responsible energy use, local investment, jobs, and community access to Codex, its AI coding tool. This is the latest sign that the race to build AI infrastructure has moved out of Silicon Valley and into small American counties with land and power to spare.
What OpenAI committed to
OpenAI framed Project Camellia as more than a warehouse full of chips. According to OpenAI, the project comes with four specific promises:
- Responsible energy. A pledge to source power in a way that accounts for the strain these facilities put on local grids.
- Community investment. Money directed back into Effingham County, not just spent on servers.
- Jobs. Construction and operations roles tied to the build-out.
- Codex access. Residents get access to OpenAI’s coding assistant, a nod toward local skills and workforce development.
What stands out here is the community angle. Most data center announcements lead with capacity and speed. OpenAI led with energy responsibility and local benefit. That’s a deliberate choice, and it tells you the company knows these projects have become politically sensitive.
Why this matters
AI data centers are enormous. They consume land, electricity, and water on a scale that reshapes the towns around them. Communities across the U.S. have started pushing back on projects that raise power bills, drain water tables, or deliver few local jobs once construction ends.
OpenAI is trying to get ahead of that resistance. By naming energy responsibility and community investment upfront, the company is writing a playbook for how to enter a small county without triggering a fight. If Project Camellia works, expect this to become the template for future sites.
There’s also a strategic reason OpenAI needs the space. Training and running frontier models takes staggering amounts of compute, and the company has been expanding its physical footprint aggressively. Project Camellia fits a pattern of OpenAI locking down land and power capacity well ahead of demand, betting that the need for compute keeps climbing.
The bigger context
The status quo used to be simple: tech giants leased space in existing cloud regions run by Amazon, Google, and Microsoft. That’s changing fast. AI companies now want purpose-built facilities they control, sited near cheap power. Georgia, with its available land and growing energy infrastructure, has become a magnet for exactly this kind of project.
Effingham County sits just inland from Savannah, a region already drawing heavy industrial investment. Adding a flagship AI facility raises the stakes for the local grid and workforce, which is likely why OpenAI attached so many community commitments to the announcement.
The Codex piece is worth flagging on its own. Handing local residents access to an AI coding tool is a small gesture next to a data center, but it signals where OpenAI wants the conversation to go. The pitch isn’t just “we’re building here.” It’s “we’re bringing skills and opportunity too.”
What to watch next
- Energy sourcing details. “Responsible energy” is a promise until there are contracts and numbers behind it. Watch for how OpenAI actually powers the site and what it does to local rates.
- Job counts. Construction jobs are temporary. The real test is how many permanent operations roles the facility creates.
- Community follow-through. Investment pledges are easy to make and easy to quietly shrink. Local reporting will show whether the money arrives.
This announcement is significant because it marks a shift in how AI companies approach the towns they build in. The compute race is now a land-and-power race, and the winners will be the ones who can expand without turning communities against them. OpenAI is betting that goodwill, energy responsibility, and local access are part of the cost of doing business at this scale.
Full details on Project Camellia and OpenAI’s commitments are available at the original source.