OpenAI expects to burn $278bn in negative free cash flow between 2026 and 2030 while revenue climbs from $36bn this year to $350bn in 2030, according to its own internal projections. The figures come from a presentation leaked to the Financial Times, and they offer a rare internal look at the wager under the entire AI build-out: spend like a nation state now, grow into it later.
OpenAI Watch analysis
What happened
The leaked deck, per the FT's report, sets two curves heading in opposite directions. Free cash flow runs deeply negative for five straight years, totalling $278bn from 2026 to 2030. Revenue is projected to rise nearly tenfold, from $36bn this year to $350bn in 2030. The report's summary gives the reason for the hole: heavy infrastructure investment combined with price pressures.
The arithmetic deserves a moment. A cumulative $278bn burn averages out at more than $55bn a year. The revenue line needs roughly 77% compound growth a year to travel from $36bn to $350bn in four years. And the document is a projection in presentation form, a genre in which ambition is traditionally rounded up.
The numbers
- $278bn negative free cash flow
The cumulative 2026 to 2030 projection in the leaked presentation. - $36bn revenue this year
The starting point of the growth curve. - $350bn revenue by 2030
Nearly a tenfold rise, about 77% compound growth annually. - Infrastructure and price pressures
The stated reasons cash flow runs negative while revenue climbs.
Why it matters
The argument over whether AI spending is sustainable has mostly run on outside estimates. This is the company's own arithmetic, and it puts a price tag on the strategy: OpenAI is projecting, internally at least, that a quarter-trillion-dollar hole is the cost of the position it wants. A planned burn of that size exceeds the total market value of most listed companies, financed in the expectation that 2030 revenue makes it all look like foresight.
It also hands both sides of the bubble argument fresh ammunition. If you think the build-out is reckless, here is the burn rate in the company's own hand. If you think it is a land grab, here is a $350bn revenue line to justify the digging.
Our read
Treat the deck as a statement of intent, not a forecast. Five-year revenue curves in leaked presentations are written by optimists, and projections are not results. But the intent is the story: this is what OpenAI believes the road ahead costs, and every future funding round, compute deal and price change will now be measured against these numbers. For anyone modelling the AI economy, the company's own stated trajectory is finally on the table, and it reframes every compute announcement as a down payment.
What to watch
- OpenAI's response
Whether the company confirms, disputes or reframes the leaked figures. - The revenue ramp
Whether reported results track towards the $36bn-to-$350bn curve. - The financing
How a burn of this scale gets funded: equity, debt, compute deals or all three. - Pricing moves
Whether API and subscription prices shift in line with the deck's price-pressure assumptions.
Discussion spark: Which line in the leaked deck is harder to believe: that OpenAI can burn $278bn and keep the money coming, or that revenue grows nearly tenfold to $350bn by 2030?
Sources and evidence
OpenAI Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by OpenAI.