OpenAI has dramatically scaled back its infrastructure spending ambitions, telling investors it now targets approximately $600 billion in compute investments by 2030, according to a report from CNBC. The revision represents a stunning 57% reduction from the $1.4 trillion figure CEO Sam Altman touted just months ago.
The numbers don't lie. When Altman made headlines promoting the $1.4 trillion infrastructure commitment, it was portrayed as evidence of AI's unstoppable momentum and insatiable demand for computing resources. Now, less than a year later, that figure has been quietly chopped nearly in half. The question investors should be asking is simple: what changed?
The massive downward revision suggests one of two possibilities, neither particularly comforting for AI bulls. Either the original $1.4 trillion figure was overly optimistic promotional material designed to attract capital and partnerships, or OpenAI has encountered significant constraints in capital availability, partnership commitments, or technical requirements that forced a reset. Both scenarios indicate the AI infrastructure buildout may face more friction than the hype cycle suggested.
For context, $600 billion by 2030 still represents massive capital deployment. But the gap between promise and reality matters in capital markets. When a company cuts its spending target by more than half, it raises questions about demand visibility, partnership stability, and whether the economics of AI training and inference are proving less favorable than anticipated.
The timing is particularly notable. This revision comes as questions mount about AI monetization, competition intensifies, and investors begin demanding clearer paths to profitability. The infrastructure gold rush that defined 2024 and early 2025 may be giving way to a more sober assessment of what's actually needed versus what sounds impressive in press releases.
OpenAI declined to provide additional comment beyond confirming the updated target to investors. What the company isn't saying may be as important as the number itself.
