OpenAI projects $278B cash burn through 2030 as compute spending surges

OpenAI expects to burn through nearly $280 billion in cash by the end of 2030 as the company pours hundreds of billions into computing infrastructure to support its AI models, according to a Financial Times report.

The ChatGPT maker projects negative free cash flow of $278 billion between 2026 and 2030, according to an internal presentation reviewed by the Financial Times. OpenAI expects computing power and infrastructure alone to cost roughly $856 billion over the period.

The spending comes despite expectations for rapid revenue growth. OpenAI forecasts revenue rising from $36 billion this year to $350 billion in 2030, with cumulative revenue reaching about $840 billion through the end of the decade.

OpenAI raised $122 billion in March, but its latest projections indicate that capital could be exhausted in 2028 if spending follows the planned trajectory. The company has been investing heavily in data centers and computing capacity while cutting prices to compete with Anthropic and lower cost open weight models.

Its cash burn forecast has nevertheless improved from an earlier projection in May, when OpenAI expected negative free cash flow of roughly $305 billion over the same period. New model launches also helped increase annualized revenue by about 20% in July, according to the FT.

The projections come as OpenAI enters early discussions over another major funding round. Investors have approached the company about financing at a valuation of around $1.2 trillion, while OpenAI is seeking an even higher figure.

The company’s funding requirements extend beyond OpenAI itself. Nvidia, Oracle, SoftBank backed infrastructure companies and other technology groups have entered large agreements tied to OpenAI’s future demand for computing capacity, making its ability to continue raising capital increasingly important across the broader AI infrastructure sector.

OpenAI had previously targeted an IPO this year after confidentially filing paperwork in June, but has since delayed those plans amid growing concerns around the pace of frontier AI development and questions over how public markets would value a company with such large projected losses.

Disclosure: This article was edited by Estefano Gomez. For more information on how we create and review content, see our Editorial Policy.