OpenAI’s Infrastructure War Chest Swells to $750 Billion Through 2030

OpenAI is significantly escalating its capital expenditure, raising its projected infrastructure spending to a staggering $750 billion by the end of the decade. This massive financial commitment signals a pivot toward aggressive hardware and energy acquisition as the race for AGI intensifies.

Project Camellia: The $20 Billion Georgia Powerhouse

As part of this renewed spending blitz, OpenAI has unveiled "Project Camellia," a massive $20 billion data center campus located on 1,400 acres northwest of Savannah, Georgia. This project serves as a critical cornerstone for the company's long-term compute strategy, especially following reports that its previous "Stargate" data center project has stalled.

The scale of Project Camellia is unprecedented. The campus is slated to draw at least 3.2 gigawatts of power from Georgia Power, with generating capacity expected to come online between 2028 and 2032. To mitigate the impact on the local community, OpenAI has agreed to a demand-response arrangement where it will reduce its power draw by up to 1 gigawatt during peak grid demand periods. Furthermore, OpenAI has committed to covering the full cost of the infrastructure and electric-service costs, adhering to recent Georgia Public Service Commission (PSC) rules designed to prevent utilities from passing massive infrastructure costs onto smaller consumers.

The Energy Dilemma: Natural Gas vs. Sustainability

The sheer energy requirements of modern LLM training and inference are pushing regional utilities to their limits. For OpenAI's Georgia project, the energy mix presents a complex environmental challenge. Regulatory filings indicate that Georgia Power expects to meet the massive demand largely through natural gas.

Specifically, the utility plans to build or procure approximately 5.8 gigawatts of natural gas generating capacity—roughly doubling its current fleet. While some capacity will be supplemented by solar energy and grid-scale batteries, a significant portion will come from more polluting simple-cycle turbines. This reliance on fossil fuels highlights a growing tension in the AI industry: the conflict between the exponential growth of compute requirements and the global push for carbon neutrality.

Scaling Compute at Record Speed

To ensure these massive investments translate into operational hardware, OpenAI is aggressively recruiting top-tier talent. The company recently hired Brett Mayo to lead data center construction, a move that brings direct expertise from xAI. Mayo previously oversaw the construction of the "Colossus" data center in Memphis, which was noted for its record-breaking build speed.

While the full electricity supply for Project Camellia is not expected until 2028, the hiring of construction specialists suggests OpenAI is looking to accelerate the timeline for when the first GPUs can be energized. This "build-at-all-costs" mentality is becoming the standard for frontier AI labs as they transition from software research to massive-scale industrial operations.

Key Takeaways

  • Massive CapEx Expansion: OpenAI has increased its infrastructure budget to $750 billion through 2030 to secure the compute and energy needed for future models.
  • Project Camellia: A new $20 billion, 1,400-acre data center campus in Georgia will demand 3.2 gigawatts of power, marking a significant leap in localized AI energy consumption.
  • Energy Infrastructure Shift: To support AI growth, utilities are pivoting heavily toward natural gas expansion, highlighting the significant environmental footprint of large-scale AI deployments.