OpenAI Booked $5.7B in Q1—and Still Burned $3.7B in Cash

OpenAI’s commercial growth has accelerated, but it has not eliminated the company’s dependence on cash. A June 2026 Reuters dispatch said shareholder documents showed $5.7 billion in first-quarter revenue and $3.7 billion in cash burn; Reuters could not independently verify the figures.

What has changed since the story emerged in 2024 is the scale of OpenAI’s sales and the legal form of its commercial operation. What remains unresolved is the central financial question: revenue is rising rapidly, yet the business still consumes substantial additional capital while developing and operating its models.

The original $5 billion figure was a forecast

The early controversy concerned projected annual results, not a completed 2026 reporting period. Fortune’s September 2024 account placed OpenAI’s forecast at $3.7 billion in revenue and $5 billion in losses for that year, with equity-based compensation excluded from the loss estimate.

That distinction corrects two misleading impressions. The $5 billion estimate belonged to 2024 rather than 2026, and it was not a prediction that OpenAI would generate only $3.7 billion indefinitely. The first-quarter 2026 revenue figure alone was about 54% higher than the full-year sales forecast circulated in 2024, although comparing one realized period with an earlier forecast does not reveal whether either period was profitable.

The newer numbers therefore alter the scale of the story without reversing its direction. OpenAI has demonstrated far greater commercial traction than the old forecast suggested, but the simultaneous cash burn indicates that sales growth has not made expansion self-funding.

Cash burn is not the same as an accounting loss

The $3.7 billion figure should not be described as a quarterly net loss. Cash burn measures the reduction in cash over a period, while net income also incorporates non-cash expenses, accruals and other accounting items. Without a complete public income statement and cash-flow statement, the two measures cannot be treated as interchangeable.

The reported cash burn was equivalent to roughly 65% of the quarter’s revenue. That calculation illustrates the size of the funding gap, but it is not a profit margin: the numerator and denominator describe different financial measures, and the available documents do not provide all the information required to reconstruct standardized quarterly accounts.

This limited disclosure also sets a boundary on stronger conclusions. The figures establish that OpenAI generated billions in sales while consuming billions in cash, but they do not disclose how much of that spending was recurring operating cost, infrastructure investment, working-capital movement or another category. They cannot, by themselves, identify a break-even date.

“Profits first” oversimplifies OpenAI’s new structure

OpenAI has adopted a structure that gives investors conventional equity, but the nonprofit did not surrender formal control. OpenAI’s current structure page describes the October 28, 2025 update: the commercial arm became OpenAI Group PBC, the OpenAI Foundation retained exclusive governance rights and a 26% stake, Microsoft held roughly 27%, and current and former employees and other investors held the remaining 47%.

The public benefit corporation form matters because OpenAI Group is required to consider its stated mission and broader stakeholder interests rather than operating as an ordinary corporation focused solely on shareholder returns. At the same time, its shares now participate proportionally in increases in the company’s value. Commercial growth and investor wealth are therefore linked more directly than under the earlier capped-profit arrangement.

Formal nonprofit control remains significant. The Foundation appoints the commercial company’s directors and can replace them, so it would be inaccurate to say that outside shareholders took over OpenAI. The sharper question is how that governance authority will be exercised when the need for capital, product access, safety commitments and shareholder value pull in different directions.

Why rapid revenue growth has not settled the argument

OpenAI’s financial challenge is not simply a shortage of customers. AI services require computing capacity both when models are trained and whenever users generate text, images, audio, video or code. Growth can consequently increase revenue and infrastructure demand at the same time, especially while the company is also developing larger systems.

Several paths could narrow the gap: higher revenue per customer, greater use by businesses, lower model-operating costs, more efficient hardware or slower infrastructure expansion. These are possible mechanisms rather than established outcomes. The first-quarter figures do not show which combination, if any, will be sufficient.

For creators and small businesses that depend on OpenAI products, the immediate issue is not evidence of an impending shutdown. It is the company’s continuing reliance on financing and its incentive to direct scarce computing resources toward uses that can support higher prices or larger contracts. Product availability, usage limits and pricing may therefore reflect capital allocation as much as technical capability.

The defensible assessment is narrower than either “OpenAI is failing” or “growth has solved the problem.” The company is generating revenue at a scale that its 2024 projections did not anticipate, while the reported 2026 cash burn remains exceptionally large. Its new structure accommodates both nonprofit control and investor returns, but it does not answer when—or whether—the operating business will finance its own expansion.

Also read:

• OpenAI's New "Strawberry" AI Is Still Making Idiotic Mistakes
• OpenAI Pivoting From "Benefiting Humanity" to "Making Lots of Money"
• After Years of Chasing Money, OpenAI Reportedly Giving Up on Being a “Nonprofit”