OpenAI is sending a more cautious signal to investors. According to internal information obtained by CNBC, the company has set its total 2030 compute expenditure target at around $600 billion — a figure that is more than half less than the $1.4 trillion Sam Altman claimed last year.

Behind the adjustment are practical considerations. Insiders revealed that OpenAI achieved $13.1 billion in revenue in 2025, exceeding the original target, but still burned through $8 billion last year. In the face of external doubts about its expansion speed, the company now provides a more specific timeline and more conservative expenditure plan.

**Financing Structure Reveals Strategic Shift**

OpenAI is finalizing a financing round of over $100 billion, with about 90% coming from strategic investors. NVIDIA plans to invest up to $30 billion, and Amazon and SoftBank are also on the investor list. This financing structure indicates that OpenAI is sharing the risk of compute costs by deeply tying itself to key players in the industrial chain.

**Revenue Expectations Match Expenditure**

The company expects its total revenue in 2030 to exceed $280 billion, with consumer and enterprise businesses each accounting for half. The new $600 billion expenditure plan is precisely to match this revenue growth expectation.

In the second half of last year, OpenAI had already announced infrastructure cooperation worth billions of dollars with chip manufacturers and cloud computing companies. If this financing is completed, OpenAI’s pre-money valuation may reach $730 billion.

From $1.4 trillion to $600 billion, the change in numbers reflects the transition of the AI industry from frenzy to rationality. When compute costs become an unbearable burden, even the most aggressive companies have to recalculate the input-output ratio.