OpenAI officially revised its long-term infrastructure roadmap, telling investors that it now expects to spend approximately $600 billion on compute by 2030.
This figure represents a dramatic recalibration—slashing its previous $1.4 trillion target by more than half—as the company shifts toward a more disciplined financial strategy ahead of a potential IPO.
The “Compute Reset”: $1.4 Trillion to $600 Billion
The adjustment signals a move from a “spend whatever it takes” mentality to one tied more closely to revenue reality.
- Strategic Recalibration: The earlier $1.4 trillion figure, famously touted by CEO Sam Altman for a massive global network of data centers and 30 gigawatts of power, had reportedly rattled some investors who questioned the return on such unprecedented capital.
- Efficiency Gains: OpenAI reportedly told investors that the forecast was cut because algorithmic improvements and inference optimization have progressed faster than expected. Newer models are becoming significantly cheaper to run and require less “raw” compute to achieve superior intelligence.
- Financial Discipline: The $600 billion target is now aligned with OpenAI’s goal of achieving positive free cash flow by 2030.
Financial Projections (2025–2030)
Despite the spending cut, OpenAI’s growth targets remain aggressive, positioning the company as one of the largest infrastructure spenders in history.
| Metric | 2025 (Actual) | 2030 (Projected) |
| Annual Revenue | $13.1 Billion | $280+ Billion |
| Annual Spending | $8 Billion | ~$100 Billion (Avg) |
| Adjusted Gross Margin | 33% | Targeting 50%+ |
| Revenue Split | Majority Consumer | 50% Consumer / 50% Enterprise |
Nvidia’s $30 Billion Equity Pivot
Coinciding with the revised spending plan, Nvidia has shifted its relationship with OpenAI from a hardware partner to a major strategic investor.
- Direct Equity Investment: Nvidia is in final talks to invest $30 billion directly into OpenAI as part of a funding round that could value the startup at $830 billion.
- Replacing the “Letter of Intent”: This $30 billion equity stake effectively replaces a previously discussed, more complex $100 billion “infrastructure agreement” that was never formalized.
- Funding Flexibility: This shift provides OpenAI with more flexible cash and gives Nvidia a clearer stake in its largest customer.
The “AI Economy” Context
Even at the reduced $600 billion mark, OpenAI’s planned infrastructure spend is nearly incomprehensible. For perspective:
- It is roughly 20 times what Tesla has spent on all its Gigafactories combined.
- It is larger than the entire GDP of Sweden.
- It averages out to $100 billion per year, which is nearly double the annual capital expenditure of hyperscalers like Amazon or Google.
“We are moving from an era of infinite scaling to one of intelligent scaling. Every dollar saved on infrastructure is a dollar that doesn’t need to be raised from investors at increasingly demanding valuations.” — Industry analysis on OpenAI’s 2026 Reset.
