OpenAI Lost $21 Billion on $13 Billion in Revenue, Leaked Financials Show

A loss-to-revenue ratio of 1.6x at scale is the number that matters. The leaked figures put OpenAI's path to profitability under serious scrutiny.

Written by Simon Reed · Bureau Finance · Jun 18, 2026

The Number That Matters

OpenAI lost $21 billion on $13 billion in revenue, according to financial documents leaked and reported by Fortune. That is the figure to hold. Not the revenue — which is impressive for a nine-year-old AI lab — but the ratio: costs running at roughly 1.6x income at a moment when the company is trying to convince investors it can eventually turn a profit.

What the Ratio Reveals

A loss-to-revenue ratio above 1.0x is not unusual for high-growth technology companies in early scaling phases. Amazon ran losses for years. So did Uber. The question is always whether unit economics improve as volume grows — whether there is operating leverage hiding inside the cost structure.

For OpenAI, the concern is that its primary cost driver — compute — does not obviously get cheaper as usage scales. Training and inference costs are tied to chip prices, energy, and data centre capacity, none of which compress automatically with revenue growth. If anything, OpenAI's product ambitions (agents, video, voice, reasoning models) require more compute, not less.

The $21 billion loss figure, if accurate, suggests that cost discipline has not kept pace with revenue growth. $13 billion in revenue is not a small number. Losing more than that on top of it is.

Context: The For-Profit Restructuring

Timing matters here. OpenAI is in the middle of converting from a capped-profit structure — where investor returns were limited — to a conventional for-profit corporation. That restructuring has attracted regulatory scrutiny and legal challenges, including from co-founder Elon Musk.

Leaked financials showing a $21 billion loss complicate the narrative OpenAI needs to tell: that the business is on a credible trajectory toward self-sufficiency. Investors who participated in the company's most recent funding round at a valuation reported above $150 billion are now looking at a company that, on these numbers, would need to roughly double revenue while holding costs flat just to break even.

What OpenAI Wants You to Look At

OpenAI's public messaging has consistently emphasised revenue growth — and $13 billion is a genuine milestone. The company reportedly crossed $1 billion in monthly revenue in 2024 and has continued to scale enterprise contracts and API usage.

But revenue growth without margin improvement is not a business; it is a fundraising story. The leaked figures suggest the gap between the two narratives — the one OpenAI tells and the one the income statement tells — is $8 billion wide.

Caveats

These are leaked documents, not audited financials. OpenAI has not confirmed the figures. The loss number may include non-cash items, one-time charges, or accounting treatments that affect comparability. Until OpenAI publishes or confirms its financials, the precise figures carry uncertainty. The directional story — large losses at scale — is consistent with what has been reported previously.

Key takeaways

$21 billion in losses against $13 billion in revenue gives a loss ratio of approximately 1.6x — meaning costs exceed revenue by more than half again.
Revenue of $13 billion represents substantial scale for a company founded in 2015, but the loss figure suggests compute, staffing, and infrastructure costs are growing at least as fast.
The leak comes as OpenAI is restructuring from a capped-profit to a for-profit entity, making these numbers directly relevant to how investors and regulators assess the transition.
OpenAI has not publicly confirmed the figures; they are drawn from leaked documents reported by Fortune.
The operating leverage story here is negative: if margins are not improving as revenue scales, the business model faces a structural problem that fundraising alone cannot fix.

FAQ

How does OpenAI's loss ratio compare to other large technology companies at similar stages?

A 1.6x loss-to-revenue ratio is high even by the standards of capital-intensive tech scaling. For comparison, Amazon's worst loss years still saw losses well below revenue. Uber and Lyft ran significant losses during their growth phases but not at this ratio at comparable revenue scale. OpenAI's situation is partly structural: compute costs do not compress the way software margins typically do.

Are these figures confirmed by OpenAI?

No. The figures come from documents leaked to and reported by Fortune as of 16 June 2026. OpenAI has not publicly confirmed or denied the numbers. They should be treated as reported but unverified until the company issues a formal statement or publishes audited accounts.

Why does the for-profit restructuring make these numbers more significant?

Under its previous capped-profit structure, OpenAI's financial performance was primarily a mission question. As a for-profit entity, it becomes a valuation question. Investors and acquirers will apply standard discounted cash flow logic — and a company losing $21 billion on $13 billion in revenue needs a credible path to positive operating margins to justify a valuation above $150 billion.

What are the main drivers of OpenAI's costs?

Compute is the dominant cost — training large models and running inference at scale requires enormous GPU capacity and energy. Staffing is a secondary factor; OpenAI employs some of the highest-paid researchers in the industry. The company has also invested heavily in data centre infrastructure and safety research.