OpenAI to spend $280 billion on AI infrastructure by 2030, FT reports

September 19, 2026 at 07:10 AM1 min

This digest was compiled by AI from multiple sources — links to the originals are below.

OpenAI will spend nearly $280 billion on computing infrastructure by 2030, according to Financial Times projections. The company expects revenue to rise from $36 billion in 2026 to $350 billion by 2030, yet cumulative expenses will exceed income by about $280 billion.

Key Facts

OpenAI's cumulative computing costs from 2026 to 2030 are projected to reach about $856 billion, exceeding revenue by roughly $280 billion.
OpenAI expects revenue to grow from $36 billion in 2026 to $350 billion by 2030, with total revenue over the period reaching $840 billion.
The company plans to raise its valuation from $852 billion to $1.2 trillion in an upcoming funding round before its IPO next year.

Infrastructure Spending

OpenAI will allocate nearly $280 billion to computing infrastructure by 2030, according to Financial Times projections. The spending will cover data center construction and chip purchases for model training. Total computing costs from 2026 through 2030 are estimated at $856 billion, exceeding projected revenue by about $280 billion.

Revenue and Losses

OpenAI forecasts revenue of $36 billion in 2026, rising to $350 billion by 2030. Cumulative revenue over the period is expected to reach $840 billion, yet the company will still post significant losses. The negative cash flow from 2026 to 2030 is projected at $278 billion, reflecting the cost of aggressive scaling.

Funding and Valuation

OpenAI is preparing a funding round that would raise its valuation from $852 billion to $1.2 trillion. The round is expected to close before the company's planned IPO next year. Investors are being asked to accept a long horizon for financial returns, as the company prioritizes market share and technological leadership over near-term profitability.