OpenAI Forecasts Nearly $280B Cash Burn
OpenAI reportedly expects nearly $280 billion of cash burn through 2030 as revenue and compute spending both scale dramatically.
OpenAI's internal forecasts reportedly show just how capital-intensive frontier AI could become even if revenue grows at extraordinary speed.
What happened
The company reportedly expects to burn nearly $280 billion in cash between 2026 and 2030.
Internal projections put revenue at about $36 billion in 2026 and as high as $350 billion in 2030, while compute-related spending over the period is projected at roughly $856 billion.
These figures are reported internal forecasts rather than realised spending or formal public guidance.
Why it matters
Frontier-model economics are unusual because rapid revenue growth can coexist with even faster infrastructure spending.
Training new models, serving inference traffic and securing long-term compute capacity all require enormous amounts of capital. That means access to financing may become as strategically important as model quality itself.
The bigger picture
The AI race is becoming an industrial-capital race. A small number of companies may be capable of spending hundreds of billions on compute, power and data centres before the economics fully mature. OpenAI's reported forecasts show why partnerships with chipmakers, cloud providers and capital markets are becoming central to competition at the frontier.