Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Yes: OpenAI is losing enormous sums while ChatGPT grows. The Information reported that OpenAI generated about $4.3 billion in revenue in the first half of 2025 while burning about $2.5 billion in cash and spending about $6.7 billion on research and development. But those figures describe OpenAI, not a separately reported ChatGPT business. And a reported projection of losses reaching about $14 billion in 2026 is a forecast, not a result.
What does it mean to say OpenAI is losing money on ChatGPT?
It is shorthand for a company-wide financial problem, not a published calculation of how much ChatGPT loses by itself. OpenAI has not demonstrated publicly that ChatGPT is profitable or unprofitable on a fully allocated basis, and the available reports do not provide a standalone ChatGPT profit-and-loss statement.
OpenAI’s finances combine revenue from consumer subscriptions, business and enterprise customers, and its API with spending on model research, product development, cloud and data-center capacity, inference (the computing used to answer prompts), sales, and employee compensation. Those costs support more than the ChatGPT app, so assigning the company’s entire loss to ChatGPT would be misleading.
Free tools Windows power users keep installed
One-click scans. No signup required.
The measures also answer different questions. Revenue is money earned; cash burn is cash used over a period; an operating loss compares revenue with operating expenses; and a net loss can include financing, tax, and accounting items beyond operations. Research and development is an expense category, while capital spending on long-lived infrastructure may be accounted for differently. Stock-based compensation is an expense but not necessarily an immediate cash payment.
#1 Best Overall
What do the reported figures actually show?
The numbers below are estimates or figures described in reporting—not a complete, publicly filed audited income statement. They should not be added together: revenue, R&D expense, cash burn, net loss, and a future projection measure different things.
| Period | Reported figure | What it means |
|---|---|---|
| 2024 | About $4 billion in revenue; roughly $5 billion in computing costs | Reuters Breakingviews cited these reported estimates while discussing OpenAI’s uncertain path to profit. They are not a full income statement and do not establish a total company loss or a ChatGPT-specific loss. Reuters Breakingviews |
| First half of 2025 | About $4.3 billion in revenue, $2.5 billion in cash burn, and $6.7 billion in R&D spending | The Information reported these figures from financial disclosures it viewed. It also reported about $2.5 billion in stock-based compensation for the period; that expense is not equivalent to an immediate cash outflow. The Information |
| 2025 reporting | Reports described substantially larger losses, including a headline figure of about $38 billion | That figure is not interchangeable with cash burn or a clean recurring operating loss. Ars Technica’s account discusses reported financial documents and accounting qualifications; a separate summary said the loss appeared closer to $8 billion after excluding a very large one-time charge and other non-cash expenses. Without the underlying statements, neither should be treated as a definitive measure of recurring operating performance. Ars Technica and State of Surveillance |
| 2026 | Losses could reach about $14 billion | This was reported as an internal projection, not an audited result. The Information described investor documents that reportedly projected losses nearly tripling from the preceding year. The Information |
The practical takeaway is that the most solidly quantified picture in these reports is fast-growing revenue alongside heavy cash use and research spending. The much larger reported loss figure and the 2026 projection need their accounting and forecast qualifications attached.
Where does the money go?
Serving prompts and other features
Each ChatGPT interaction uses computing capacity. A short text exchange is not equivalent to analyzing a large file, generating an image, using voice, browsing, conducting deep research, or running coding and reasoning workflows. Workloads differ, so a subscription price alone cannot reveal the cost of serving an individual customer.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Training models and doing research
Frontier-model development requires accelerator clusters, data-center capacity, electricity, researchers and engineers, data, and repeated experimentation. The Information’s reported $6.7 billion in first-half 2025 R&D spending is a company-level category, not a bill solely for training one model or running ChatGPT.
Securing infrastructure before demand is certain
OpenAI says its available computing capacity grew from about 0.2 gigawatts in 2023 to 0.6 gigawatts in 2024 and about 1.9 gigawatts in 2025. These are the company’s figures for available compute, not a measure of ChatGPT’s electricity use or an explanation of costs by itself. They illustrate the scale of capacity OpenAI says it is building toward as it pursues more use and revenue. OpenAI’s account of its scaling strategy
Compensation and commercial operations
Researchers, engineers, infrastructure specialists, product teams, and sales staff all contribute to costs. The Information reported about $2.5 billion in stock-based compensation in the first half of 2025. Such compensation may not require an equivalent cash payment at the time it is recorded, but it is still an economic cost to employees and shareholders and can substantially affect reported earnings.
Are free ChatGPT users the problem?
Free access can create inference demand without subscription revenue from that user, but it does not follow that every free account loses a known amount. OpenAI has not published a dependable per-user cost calculation that accounts for model choice, usage, features, and revenue allocation.
The free tier can also serve as a route to paid plans, product discovery, and broader adoption. OpenAI’s pricing page shows limits on access to some advanced capabilities for free users and expanded access on paid plans; limits and features can change. That can help manage costs, but it does not reveal the economics of an individual account. OpenAI ChatGPT pricing
Can paid ChatGPT plans still be unprofitable?
Potentially. A flat monthly subscription brings in a predictable amount from a customer while that customer’s use may vary considerably. A person who mostly asks short questions and a person who frequently uses compute-intensive tools do not necessarily cost the same to serve. Whether a subscription is profitable therefore depends on usage, features, model choice, infrastructure costs, and how OpenAI allocates shared expenses—not simply on the advertised monthly price.
OpenAI’s consumer pricing page showed Plus at $20 a month and Pro at $200 a month on August 16, 2026. Those are listed prices, not proof that either plan is profitable; availability, limits, and pricing can change. Check OpenAI’s current consumer pricing
API pricing is usage-based, so charges can track consumption more directly than a flat subscription, but that does not establish that API sales are profitable after serving, support, and other costs. OpenAI also offers team plans that combine seats with additional usage or credits for some advanced capabilities. Its Business page showed $20 per user monthly when billed annually or $25 billed monthly, with a two-user minimum, on August 16, 2026. Its flexible-pricing help page explains the additional usage mechanism. OpenAI Business pricing and OpenAI flexible pricing help
Recommended Free Tools
Why not charge more?
Higher prices could improve revenue per customer, but the company would have to weigh that against slower adoption, customer churn, and competitors offering cheaper alternatives. Business buyers may also resist higher bills unless they can connect the tools to measurable productivity gains. Heavy users might move to less expensive models, other providers, or local models.
Consumer pricing can function as distribution and adoption strategy as well as cost recovery. OpenAI may accept weak margins on some use in pursuit of future subscriptions, enterprise adoption, API demand, or other products. Even if the cost of answering prompts falls, that alone would not ensure company-wide profit while research, infrastructure commitments, hiring, and sales remain large.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What could improve—or worsen—the economics?
Potential paths to better margins
- More efficient models and inference systems could reduce the cost of serving a given workload.
- Higher enterprise and API revenue could make more of the underlying capacity generate income.
- Usage-based charges or feature limits could align revenue more closely with expensive workloads.
- Better data-center utilization could spread fixed capacity costs across more paid activity.
- Converting some free users into paying customers could increase revenue, though the value depends on how much those users consume.
Pressures that could keep losses high
- New model training and research may require continued investment even as older models become cheaper to serve.
- Infrastructure commitments made ahead of demand can leave capacity expensive or underused if growth disappoints.
- Price competition may limit how much OpenAI can charge consumers and enterprise customers.
- More compute-intensive features can raise usage costs faster than subscription revenue.
- Slower user or revenue growth would make it harder to absorb fixed investment and ongoing commitments.
Reporting based on Wall Street Journal coverage said OpenAI had fallen short of some internal revenue and user targets and that executives were concerned about future computing commitments. Those are attributed claims, not independently verified financial results in a public-company filing. Reuters summary carried by Yahoo Finance
Can OpenAI afford to keep losing money?
It may be able to continue funding losses if investors and strategic partners keep supplying capital, but funding is not profit. Financing can let a company spend ahead of revenue; it does not prove that the underlying business will eventually earn more than it costs. The terms, obligations, and ownership consequences of future financing matter, as do the company’s capacity commitments and actual demand.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteThe central business question is whether new spending creates enough future revenue, strategic position, or lower serving costs to justify it. OpenAI’s scale-up thesis depends on more than ChatGPT subscriptions: enterprise products, API usage, and tools embedded in valuable workflows could expand the addressable revenue. But that outcome is not guaranteed by user growth or a large funding round.
What the loss headlines do—and do not—tell you
A figure labeled “loss” can describe a cash shortfall, an operating result, or a net accounting result; a projection describes a possible future, not a completed period. Likewise, company-wide costs cannot be assigned to ChatGPT without segment-level reporting. OpenAI’s published materials and the cited reporting do not establish a clean monthly loss for ChatGPT, a verified cost per free user, or whether a specific subscription tier makes money after fully allocated costs.
So the defensible answer is that OpenAI is spending heavily and, by the reported measures, losing substantial sums as a company. ChatGPT is central to its revenue and computing demand, but the headlines alone cannot show that ChatGPT is inherently uneconomic—or that future growth will make it profitable.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →
