OpenAI Slashes Compute Spending Forecast to $600B
As OpenAI prepares for a potential Initial Public Offering later this year, the company is pivoting its infrastructure strategy to demonstrate more fiscal responsibility. OpenAI told investors in February it is now targeting roughly $600 billion in total compute spending by 2030, a significant reduction from more ambitious figures previously discussed by CEO Sam Altman. This strategic shift moves the company away from building and owning its own massive data centers and toward leasing capacity from cloud partners like Oracle, Microsoft, and Amazon.
The change in approach follows practical difficulties OpenAI encountered with its flagship Stargate data center initiative. After facing construction issues and challenges securing financing, OpenAI has stepped back from developing large portions of the project itself. Oracle is now leasing Stargate's primary campus in Abilene, Texas, and funding the buildout, underscoring OpenAI's retreat from direct ownership and its new focus on purchasing cloud capacity rather than building physical infrastructure.
Record $25B Revenue Fails to Cover $57B Annual Cash Burn
OpenAI’s strategic adjustment comes as its explosive top-line growth runs against a staggering cost structure. The company achieved a $25 billion annualized revenue run rate as of late February 2026, growing 17% in just two months from the $21.4 billion figure reported at the end of 2025. This growth trajectory, which saw revenue climb from $2 billion in 2023, is faster than any software company in history, including Google and Facebook.
Despite this record revenue, OpenAI remains deeply unprofitable and is not expected to break even until 2030. Its annual cash burn is projected to reach $57 billion by 2027, creating a complex financial narrative for Wall Street. The enormous cost of compute power required to train and run its AI models forces the company to generate revenue at an unprecedented scale just to cover its operational expenses, placing its path to profitability under intense scrutiny ahead of a public listing.
IPO Push Triggers Focus on Enterprise, Competition with Anthropic
With a potential IPO targeted for the second half of 2026 at a valuation up to $1 trillion, OpenAI has sharpened its internal focus to strengthen its business case. The company has redirected resources away from what one executive called "side quests"—such as its Sora video generator and Atlas web browser—to concentrate on enterprise and coding applications. This effort aims to rapidly scale its enterprise business, which currently accounts for $10 billion of its $25 billion annualized revenue.
This urgency is amplified by fierce competition and contractual obligations. Rival Anthropic has surged to nearly $19 billion in annualized revenue and is targeting profitability by 2028, two years ahead of OpenAI. This gives Anthropic a fundraising advantage with investors. Furthermore, investment deals, including Amazon's $50 billion commitment, contain clauses that make a portion of the funding contingent on an IPO, contractually accelerating the timeline. The combination of competitive pressure and investor expectations is forcing OpenAI to prove it can operate as a disciplined business, not just a research lab with a revenue stream.
