OpenAI Raises Planned Cloud Spending Through 2030 to $750 Billion

Jul 22, 2026

OpenAI is scaling up its computing budget to $750 billion through 2030 and launching a $20 billion data center in Georgia, amid internal leadership tensions over spending. Concurrently, SoftBank is raising 60 billion yen to fund its $65 billion OpenAI commitment. This massive spending occurs as the industry faces intense pressure to prove ROI, prompting OpenAI CFO Sarah Friar to introduce an outcome-based 'useful-intelligence-per-dollar' framework to address growing skepticism over AI economics.

AI spending growth, funding, and infrastructure

▪OpenAI has increased its projected spending on computing power to approximately $750 billion through 2030, up from an earlier projection of $600 billion.
▪SoftBank Group Corp. plans to raise approximately 60 billion yen ($369 million) through a bond offering to Japanese institutional investors during the week of July 27, 2026.
▪OpenAI's cloud-computing spending has caused tension between Chief Executive Officer Sam Altman and Chief Financial Officer Sarah Friar ahead of the company's planned initial public offering.
▪Worldwide spending on artificial intelligence is forecast by Gartner to total $2.59 trillion in 2026, representing a 47% increase year-over-year.
▪OpenAI announced plans to invest $20 billion to develop a 3.2-gigawatt data center called Project Camellia in Effingham County, Georgia.
▪SoftBank Group Corp. has committed nearly $65 billion to OpenAI and has considered a margin loan of around $6 billion backed by its OpenAI stake.
▪OpenAI hired Brent Mayo, a former xAI employee who helped build the Colossus supercomputer facility, to assist with its data-center build-out.

OpenAI useful-intelligence-per-dollar framework

▪The "useful-intelligence-per-dollar" framework evaluates artificial intelligence based on task performance, cost per successful task, output reliability, and value scaling with usage.
▪OpenAI Chief Financial Officer Sarah Friar introduced a "useful-intelligence-per-dollar" framework to measure artificial intelligence investments based on outcomes and task completion costs.

AI investment ROI pressure

▪A PwC study from January 2026 revealed that only 12% of chief executive officers reported that artificial intelligence delivered both cost and revenue benefits.
▪Investor concerns regarding returns on artificial intelligence spending have pressured semiconductor-related stocks, including SoftBank Group Corp.'s chip-technology subsidiary, Arm Holdings Plc.
▪S&P Global Ratings revised SoftBank Group Corp.'s outlook to stable from negative and affirmed its BB+ long-term issuer credit rating on July 16, 2026.
▪The PwC study indicated that 56% of chief executive officers reported seeing no significant financial benefit from artificial intelligence investments.

Token cost scrutiny

▪The issue of artificial intelligence investment returns has intensified due to increased scrutiny of token consumption and the processing costs paid by companies.
▪Business leaders are experiencing growing frustration over the economics of large language models, specifically questioning whether rising token costs translate into return on investment.

LLM economics debate

▪Palantir Chief Executive Officer Alex Karp argued that enterprises are growing tired of large language model economics and questioned their return on investment.
▪OpenAI Chief Financial Officer Sarah Friar argued that artificial intelligence value must be measured by work accomplished rather than traditional software adoption metrics like user seats.