OpenAI raises 2030 computing expenditure forecast to $750 billion but CFO expresses uncertainty on financing and repayment viability.
OpenAI's compute bill has climbed to $750 billion through 2030, and the crucial detail isn't only the size. Its own finance chief has reportedly questioned whether the company can grow fast enough to pay for it.
The Wall Street Journal reported on July 22 that OpenAI has lifted its planned cloud and compute spending to $750 billion through 2030, up from about $600 billion earlier this year. That is the headline. The real warning is sharper: CFO Sarah Friar has raised concerns that OpenAI may not be able to support future computing contracts if revenue growth doesn't keep pace.
The spending spans deals with Microsoft, Oracle, AWS, and CoreWeave, plus OpenAI's own Project Camellia in Effingham County, Georgia. On July 22, OpenAI announced that Camellia will be designed and developed in the Savannah Gateway Industrial Hub, with Georgia Power contracted to deliver 3.2 gigawatts in phases between 2028 and 2032. According to Data Center Dynamics, citing Bloomberg, the Georgia project could eventually cost more than $30 billion.
This marks OpenAI's shift from renting capacity to behaving like an infrastructure owner. The company has hired Brent Mayo, formerly of xAI, to oversee data center construction and delivery. He reports to Uday Ruddarraju, OpenAI's CTO of computing capacity. This signals serious, long-term buildout.
Sacra estimates OpenAI hit a $25 billion annualized revenue run rate in February 2026, up from $20 billion at year-end 2025. That is substantial. The problem: compute commitments are rising faster than that growth, and unlike projections, contracts require cash.
Fortune reported in April that Friar was worried OpenAI was spending too much on data centers and might not generate enough revenue to cover already-signed contracts. Forbes reported the same source revealed OpenAI missed internal revenue and user growth projections. These are not trivial concerns. When you commit now to capacity arriving years later, a single missed forecast cascades across the entire spending plan.
The cost trajectory has turned sharp. Business Insider reported this week that building one gigawatt of AI capacity with common Nvidia systems has risen from about $29 billion to $35 billion, while newer configurations reach roughly $49 billion. A gigawatt-scale facility can take years to bring online. The arithmetic is straightforward: OpenAI is buying future compute against revenue it still must prove.
Oracle's exposure shows this is no longer just an OpenAI story. The Wall Street Journal reported that OpenAI agreed to buy $300 billion of Oracle compute over roughly five years beginning in 2027. Latest reporting adds a $138 billion AWS commitment and a $250 billion Microsoft Azure commitment. These are not software contracts with extra zeros. They bind the AI boom directly to the balance sheets and construction schedules of the largest infrastructure companies in the market.
Markets have been too casual about this dependency. A cloud provider typically sells capacity to customers whose spending aligns with predictable business models. OpenAI is different: extraordinary growth, substantial losses, and a valuation Bloomberg reported at $852 billion after a $122 billion funding round completed in late March. Amazon invested $50 billion; Nvidia and SoftBank each committed $30 billion. The valuation assumes demand arrives before the bills do.
Project Camellia offers a concrete measure of scale: 3.2 gigawatts, phased power delivery from 2028 to 2032, potential cost exceeding $30 billion. OpenAI states that residents won't subsidize the project and its closed-loop cooling system won't deplete local water. The harder question isn't whether Georgia can host the site—it's whether OpenAI can generate durable revenue from that infrastructure before commitments consume all available cash flow.
None of this suggests OpenAI faces immediate peril. Its revenue base is substantial, ChatGPT retains enormous reach, and enterprise demand is real. But the Journal's reporting reveals a genuine scenario where OpenAI has promised to buy more compute than its revenue trajectory can support. The CFO concern is the point. Admirable ambition and visible risk can coexist.
For investors eyeing an eventual IPO, the central question is whether this spending builds a moat or accumulates a liability. Right now, the contracts are more certain than the cash flows. That gap is what $750 billion looks like up close.