On July 22, 2026, OpenAI announced that it plans to invest $75 billion in AI infrastructure before 2030, an increase of about 25% compared to earlier estimates this year. This move marks OpenAI's renewed focus on long-term computing power advantages through more aggressive capital expenditures, amid the stagnation of its flagship data center project "Stargate."

OpenAI

As the first project under this large investment plan, OpenAI intends to build a 1,400-acre "Project Camellia" data center campus in the northwest of Savannah, Georgia, with a cost of $2 billion. The campus has partnered with Georgia Power and is expected to obtain at least 3.2 gigawatts of electricity supply between 2028 and 2032, with all new infrastructure and power service costs borne by OpenAI. To reduce the impact on the local power grid, OpenAI has committed to reducing peak load by up to 1 gigawatt, and the project has received a 15-year 50% property tax exemption from Effingham County.

Despite the massive power consumption, neither party disclosed specific details about the transition to clean energy. Regulatory documents show that Georgia Power plans to add nearly 9.9 gigawatts of power generation capacity, with over half (about 5.8 gigawatts) relying on natural gas, while the rest will be supplemented by photovoltaic and energy storage systems, raising concerns in the market about the increased reliance on fossil fuels and carbon emissions due to computing power expansion. Notably, OpenAI recently hired Brett Mayo, who previously led the construction of xAI's Memphis Colossus data center, to oversee its data center operations, possibly indicating that the first GPU deployment nodes may be established before 2028.

At this critical stage of generative AI entering advanced model training and large-scale inference applications, the implementation of OpenAI's substantial infrastructure funding highlights not only the deep competition among leading manufacturers for computing power and electricity resources, but also the real challenges faced by ultra-large-scale AI data center construction in reshaping regional power grid structures and promoting energy structure transformation.