When people ask about OpenAI income, they usually mean one of two things: how much money comes in, or how much is left after the bills. The first question has fuzzy answers. The second has even fuzzier ones, because the company is private and the figures come from leaks and reporting rather than public filings. Still, the picture that has emerged is consistent on one point: OpenAI spends enormous amounts of money, and the gap between revenue and cost is the big story.
This article separates what has been reported from what’s still guesswork, and explains the terms you’ll run into.
What’s been reported about OpenAI income and losses
All of these figures come from news reports citing company information, not from audited accounts, so treat them as provisional.
- 2025: Reports put the net loss at about $39 billion, but a large share of that was non-cash, linked to a restructuring charge and other accounting items. Excluding those, the adjusted loss was reported at around $8 billion.
- First quarter of 2026: The Information reported that OpenAI burned $3.7 billion in cash while generating $5.7 billion in revenue. It also reported a larger accounting net loss, again affected by non-cash charges.
- The future: The Financial Times reported, citing a company presentation, that OpenAI forecasts negative free cash flow of around $278 billion over the five years from 2026 to 2030.
Sources conflict on several numbers, and some forecasts floating around online are projections from analysts, not figures the company reported. Always check which one you’re reading.
Net loss, cash burn and why they differ
Three terms get mixed up constantly:
- Net loss is an accounting measure. It can include non-cash items, such as a one-time charge from restructuring or the way stock compensation is counted.
- Cash burn is the actual cash leaving the business after cash coming in. It’s what determines how long a company can keep operating before it needs more funding.
- Free cash flow is operating cash flow minus capital spending, such as buying data centre equipment.
That’s why a headline with a huge net loss can sound scarier than the cash picture, and why a modest quarterly burn can hide a very large long-term forecast. When you read a figure, check which of the three it is.
Why the costs are so high
The main driver is computing. Training advanced models takes vast amounts of specialised chips and electricity, and serving millions of users every day adds ongoing costs. On top of that, the company pays for research talent and data centre commitments. Revenue has been growing fast, as we covered in our look at OpenAI revenues, but costs have been growing alongside it. Growth in sales doesn’t translate into profit if each new user also adds computing expenses.
How the company pays for the gap
Companies in this position fund losses by raising money from investors, taking on partnerships and eventually, if they choose, selling shares to the public. Reports in 2026 said OpenAI had filed confidentially for a stock market listing, though I couldn’t find a confirmed public filing or listing date. A public listing would require audited statements, which would finally replace the leaked numbers with official ones. See our explainer on the OpenAI IPO for how that process works.
Should the losses worry you?
It depends who you are. For investors, large losses are a bet that scale will eventually bring lower costs per user and higher prices or volumes. History has examples both ways: some tech companies lost money for years and became very profitable, others never found the margin. For users and businesses, the practical risk is different: if a provider’s economics don’t work, prices, limits or product plans can change. That’s a reason not to build everything around a single vendor.
Quick answers about OpenAI income
Is OpenAI profitable?
Based on the reporting I found, no. Reports describe substantial losses and cash burn, though accounting measures and cash measures tell different stories.
Where did the numbers come from?
Mostly from leaked or privately shared company information reported by outlets such as The Information and the Financial Times, not from audited public filings.
Will an IPO change this?
A public listing would require audited statements, so we’d finally see the official figures, including costs, and could judge the business on firmer ground.
A simple way to track OpenAI income reports yourself
Keep a small note with four columns: the date of the report, the outlet, the metric it uses and whether it’s a run rate, a cash figure or an accounting figure. When a new number appears, add it instead of replacing the old one. After a few entries you’ll see which numbers move together and which just reflect different definitions. It’s also a handy way to notice when a headline claims something that goes beyond what its source actually said, which happens more often than you’d think.
What to watch
Three things will tell you more than any leaked figure: a public prospectus with audited accounts, the trend in cost per user and the share of revenue from higher-margin enterprise customers. Until then, read each OpenAI income report as a clue, not a conclusion. The Next Web’s summary of the first-quarter report is a good example of how the details are being reported.
Put simply, OpenAI income is a story about spending as much as earning. Whenever you read about OpenAI income, check whether the figure is a net loss, a cash burn or a forecast, because those three measures can differ by tens of billions of dollars.
Featured image: Laptop on desk book stacks (Unsplash).jpg by freddie marriage fredmarriage, CC0, via Wikimedia Commons.
