Altman Slashes OpenAI's Compute Targets
OpenAI CEO Sam Altman has reduced the company's near-term compute targets by more than half. The move is reportedly a response to investor concerns and serves to recalibrate the company's ambitions for scaling its artificial intelligence models.
- The new compute spending target is approximately $600 billion through 2030, a 57% reduction from the previous $1.4 trillion figure. - This revised spending plan is tied to a new revenue projection of more than $280 billion by 2030, with contributions expected to be split almost evenly between consumer and enterprise businesses. - The adjustment comes as OpenAI is finalizing a funding round expected to exceed $100 billion, with Nvidia, SoftBank, and Amazon reported as potential strategic investors. - Nvidia is reportedly in discussions to invest up to $30 billion as part of this funding round. - Rising operational costs are a key factor, as the expense of running AI models for users, known as inference costs, quadrupled in 2025. - This increase in inference costs caused OpenAI's adjusted gross margins to fall to 33% in 2025 from 40% in 2024. - Despite cost pressures, the company's revenue in 2025 reached $13.1 billion, surpassing its internal $10 billion target, while its cash burn of $8 billion was lower than the $9 billion it had forecasted. - User growth remains strong, with the flagship ChatGPT service now exceeding 900 million weekly active users.
