OpenAI has raised its planned spending on computing power through 2030 to nearly $750 billion, up from roughly $600 billion at the start of the year, according to people familiar with the company’s plans. The increase, reported by the Wall Street Journal, reflects an accelerating race to secure chips, data centers and energy before rivals lock up the supply. The company also committed $20 billion to build a new data-center campus in Georgia, its largest single infrastructure investment in the United States to date.
The revised figure is more than the annual economic output of many countries. It signals that OpenAI, despite its ongoing talks with investors over a new funding round, intends to stay at the center of the AI infrastructure buildout that has come to define the technology industry this decade. The company has said its models require computing power that grows faster than the industry’s ability to supply it.
The Georgia project is the clearest sign yet that the plans are moving from spreadsheets to construction sites. OpenAI said the campus will house tens of thousands of specialized chips and will be powered in part by new energy infrastructure the company is helping to finance. State and local officials in Georgia have been courting the project for months, according to people familiar with the talks, attracted by the promise of construction jobs and a long-term economic anchor.
The spending increase comes as OpenAI’s rivals make similar commitments. Microsoft has pledged more than $200 billion in AI infrastructure this year alone, and Amazon and Google’s parent Alphabet are expanding their own data-center footprints at a record pace. The combined commitments have strained the supply of advanced chips, pushed power utilities to revisit grid planning and drawn scrutiny from regulators worried about the concentration of capital.
OpenAI’s path to financing the buildout is not without friction. The company is in talks with sovereign funds and private investors for a new round that would value it well above its last round, according to people familiar with the matter. The company has also been negotiating with cloud providers over the terms of its compute contracts, seeking flexibility as its needs change quarter to quarter.
The company’s appetite for computing power has become a subject of debate inside the industry. Some engineers and researchers argue that more efficient models could deliver similar capabilities with far less hardware; OpenAI executives counter that the demand for AI services is growing faster than efficiency gains can offset. The $750 billion figure is the company’s answer to that debate: scale first, optimize later.
For the broader market, the numbers reset the bar for what a serious AI company must spend. Startups raising money to compete with OpenAI now face a capital requirement that few venture funds can meet on their own, pushing them toward partnerships with cloud providers and national governments. The practical effect is that the AI industry is consolidating around a handful of companies that can write checks of this size.
The Georgia campus will take years to build, and the full $750 billion will not be spent all at once. But the trajectory is what investors are pricing in. OpenAI’s ability to raise, spend and then earn back the cost of its infrastructure will determine whether the AI boom of this decade is remembered as an investment cycle that paid off or as the largest capital misallocation in technology history. For now, the company is betting on the former, and it is placing that bet in Georgia.
The scale of the plan has raised questions about energy, which has become the binding constraint on AI expansion. OpenAI’s executives have said publicly that the company needs gigawatt-scale power for its campuses, and the Georgia project is designed with that requirement in mind. Utilities in the region are expanding transmission capacity, and the company has been in talks with power providers about dedicated supply, according to people familiar with the plans. Local communities, meanwhile, are weighing the jobs against the strain on grids and water.
The funding side is moving in parallel. OpenAI has held discussions with sovereign wealth funds and institutional investors about a new round, with terms that would value the company in the hundreds of billions of dollars, according to people familiar with the matter. The company has also expanded its existing partnership with Microsoft, which remains its largest cloud provider even as OpenAI signs deals with other suppliers. The arrangement lets OpenAI spread its demand across multiple clouds while keeping access to Microsoft’s capacity.
The broader effect of the revised outlook is on the industry’s capital math. If OpenAI alone spends $750 billion by 2030, and Microsoft, Google and Amazon spend comparable sums, the total AI infrastructure investment of the decade runs into the trillions. Suppliers of chips, power equipment and construction are already pricing in that demand. The risk, some economists and analysts argue, is that a slowdown in AI adoption would leave a vast inventory of underused data centers — the industry’s equivalent of overbuilt office towers. OpenAI’s answer to that concern is the same as its answer to everything else: demand will arrive, because the technology is moving fast enough to justify the spending.
