Saturday, July 25, 2026
DarkSubscribe
AI Infrastructure · News & Analysis
HomeCapital MarketsReport
Capital Markets · Report

OpenAI revised its long-term compute spending plan to $750 billion through 2030 and committed to $20 billion on building its own data centers, signaling a shift from cloud provider reliance to captive infrastructure.

Signals OpenAI's bet on captive datacenter infrastructure and sustained AI compute demand, requiring unprecedented power and grid buildout.
Trade pressSlicast · July 23, 2026 · US · Source: Google News
importance 95

OpenAI has increased its 2030 computing expenditure forecast to approximately $750 billion, up from $600 billion earlier this year. The company attributed this escalation to sustained demand for AI model development and a series of new deals with cloud providers securing massive computing resources.

To support this growth, OpenAI announced a $20 billion investment to construct "Project Camellia," a data center in Savannah, Georgia. This marks the first time OpenAI has led the design and construction of its own data center, signaling a strategic transition from compute tenant to infrastructure developer. Previously, the company obtained computing power primarily by renting chip capacity from providers such as Oracle and Amazon Web Services (AWS).

The spending increase requires context from earlier announcements. CEO Sam Altman had previously stated publicly that OpenAI planned to spend $1.4 trillion on computing power—a statement that prompted investor concern. CFO Sarah Friar subsequently clarified to investors that the actual projected spending through 2030 was approximately $600 billion. The newly announced figure represents a $150 billion increase from that estimate.

Sachin Katti, OpenAI's Vice President of Compute Strategy, explained the rationale for building proprietary infrastructure: "After going through two or three generations of compute iterations, we have learned many lessons on how to design these facilities to both cut costs and speed up construction. That is why we are stepping up to lead the design and the entire process."

OpenAI's broader vision extends beyond its own operations. The company plans to standardize its AI data center designs and publicly release them for use by other organizations—open-sourcing hardware, networking, and data center designs similar to the Open Compute Project co-founded by Meta in 2011.

To execute this ambitious agenda, OpenAI hired Brent Mayo as Head of Data Center Construction and Delivery. Mayo departed from xAI, where he served as a core architect of Musk's data center construction projects and led the development of xAI's Colossus supercomputing campus in Memphis, managing rapid procurement, commissioning, and deployment of large-scale AI chip clusters. At OpenAI, Mayo will oversee schedule compliance with cloud partners while advancing the new Georgia facility.

Read the original
OpenAI revised its long-term compute spending… · Slicast