OpenAI Commits $750 Billion to Infrastructure Through 2030
The company's first major move is a 1,400-acre Georgia data center campus requiring 3.2 gigawatts of power, raising questions about energy sourcing and grid impact across the U.S. Southeast.
A Quarter-Trillion-Dollar Escalation
OpenAI has raised its infrastructure spending commitment to $750 billion through the end of this decade, marking a 25% increase from estimates disclosed just months ago. The revised figure arrives as the company's Stargate data center initiative has encountered obstacles, prompting a strategic pivot toward new regional deployments.
The first concrete expression of this capital deployment is Project Camellia, a $20 billion data center complex planned for 1,400 acres northwest of Savannah, Georgia. The facility will require a minimum of 3.2 gigawatts from Georgia Power, the regional utility, with generating capacity scheduled to come online between 2028 and 2032. OpenAI has committed to covering the full cost of infrastructure and electric service for the campus.
At DailyTechWire, we've tracked the escalating power demands of frontier AI labs across Asia and North America. This Georgia deployment represents one of the largest single-site commitments to date and signals a broader shift in how compute-intensive AI development is reshaping regional electricity grids.
Grid Economics and Local Incentives
Georgia Power's agreement with OpenAI includes a demand-response provision requiring the AI company to reduce consumption by up to 1 gigawatt during periods of peak grid stress. This arrangement reflects growing utility concern over the stability implications of hyperscale data centers, particularly in regions where summer cooling loads already strain transmission infrastructure.
Effingham County has approved a 50% property tax abatement for 15 years to secure the project. The incentive structure mirrors deals struck by other Southern states competing for data center investments, though the scale of the OpenAI commitment dwarfs most prior agreements.
Georgia adopted regulations in 2023 preventing utilities from socializing costs for new customers drawing more than 100 megawatts. The rule was designed to shield residential and small commercial ratepayers from the infrastructure expenses associated with hyperscale compute facilities. OpenAI's commitment to self-fund grid connections appears to satisfy this threshold, though the long-term rate impact of adding gigawatts of baseload demand remains an open question.
The Fossil Fuel Question
Neither OpenAI nor Georgia Power has disclosed the energy mix that will supply Project Camellia. Regulatory filings offer some clues. In December, Georgia Power received approval to add 9,885 megawatts of generating capacity, with expectations that all capacity would be contracted by the end of 2026. The OpenAI deal accounts for roughly one-third of that total.
Documents submitted to the Georgia Public Service Commission indicate that approximately 5.8 gigawatts of the new capacity will come from natural gas plants, including about a quarter from simple-cycle turbines, which are less efficient and more polluting than combined-cycle alternatives. The expansion would more than double Georgia Power's natural gas fleet. The remainder of the new capacity is expected to come from grid-scale batteries and solar installations.
The reliance on fossil generation for AI infrastructure has become a flashpoint in climate policy circles. While OpenAI and its peers frequently emphasize commitments to renewable energy, the physics of baseload power and the timing of renewable buildouts often result in natural gas serving as the default solution for meeting immediate demand.
Construction Timeline and Leadership
OpenAI has not specified when the first GPU clusters at Project Camellia will become operational, though electricity is expected to begin flowing in 2028. The company recently hired Brett Mayo to oversee data center construction. Mayo previously led the Colossus data center project for xAI in Memphis, a facility notable for its accelerated construction timeline.
Colossus was completed in a matter of months, a pace that drew both admiration and scrutiny. The Memphis facility has since been named in a lawsuit filed by the NAACP and the Southern Environmental Law Center, which alleges that dozens of unpermitted natural gas turbines have been operating on site, with xAI claiming exemption from federal clean air regulations. The suit also raises concerns about local air quality impacts.
Mayo's appointment suggests OpenAI may be prioritizing speed over extended permitting and environmental review processes, though the company has not publicly outlined its construction philosophy for Project Camellia.
Stargate Stalls, Regional Bets Rise
The $750 billion spending plan comes as OpenAI's Stargate project, a joint initiative with partners including Microsoft and Oracle, has reportedly lost momentum. Stargate was conceived as a multi-site network of data centers designed to support the next generation of large language models and multimodal systems.
The pivot to standalone regional campuses like Project Camellia may reflect shifting economics around data center partnerships, or it may signal OpenAI's desire for greater operational control. The company has not commented publicly on the status of Stargate or the strategic rationale for pursuing independent facilities.
For utilities and policymakers across Asia and North America, the OpenAI spending trajectory underscores a broader challenge: how to accommodate exponential growth in AI compute demand without destabilizing electricity grids or locking in decades of fossil fuel dependence. Southeastern U.S. states, with relatively low power costs and available land, have become focal points for this infrastructure buildout. Yet the environmental and rate implications are only beginning to surface in public debate.
What Comes Next
OpenAI's revised spending commitment of $750 billion through 2030 dwarfs the capital expenditure of most technology companies, placing it in the same league as national infrastructure programs. The company has not disclosed how many additional sites it plans to develop beyond Project Camellia, nor has it detailed the geographic distribution of future investments.
As AI labs race to secure compute capacity, the energy question will likely become more acute. The gap between renewable energy ambitions and the immediate availability of low-carbon baseload power remains wide. For regions like Georgia, the arrival of multi-gigawatt data centers represents both economic opportunity and a test of grid resilience and environmental policy.
The decisions made in Effingham County over the next few years will offer a case study in how local governments, utilities, and AI companies negotiate the terms of this new industrial era. Whether that negotiation produces a sustainable model or simply a faster route to fossil fuel lock-in remains to be seen.
