Your data. Your choice.

If you select «Essential cookies only», we’ll use cookies and similar technologies to collect information about your device and how you use our website. We need this information to allow you to log in securely and use basic functions such as the shopping cart.

By accepting all cookies, you’re allowing us to use this data to show you personalised offers, improve our website, and display targeted adverts on our website and on other websites or apps. Some data may also be shared with third parties and advertising partners as part of this process.

Shutterstock
News + Trends

Leaked figures show: OpenAI is making huge losses

Samuel Buchmann
19/6/2026
Translation: machine translated

The company behind ChatGPT spent US$21 billion more than it took in last year. The losses are significantly higher than those of previous tech companies during their growth phase.

OpenAI is racking up huge losses. This is reported by the independent financial journalist Ed Zitron reports, citing leaked documents that have also been verified by the «Financial Times».

In 2025, OpenAI took in around 13 billion US dollars and spent 34 billion. This resulted in an operating loss of around 21 billion. The net loss was as high as 60 billion. However, this figure includes one-off effects arising from the conversion from a non-profit to a for-profit organisation. In the process, OpenAI had to revalue the shares of existing external and internal investors.

Growth at any cost

Revenue more than tripled within a year – from 3.7 billion dollars in 2024 to just over 13 billion in 2025. However, costs also skyrocketed. In 2025, OpenAI spent 19.2 billion on research and development (2024: 7.8 billion), 7.5 billion on direct operating costs (2024: 2.6 billion) and 5.7 billion on sales and Marketing (2024: 1.1 billion).

A large proportion of expenditure goes towards computing power. The documents show that OpenAI transferred a total of 17.2 billion dollars to Microsoft in 2025, of which 10.6 billion was for training new models and 6 billion for processing prompts. The extremely high marketing costs are also noteworthy. At the end of the year, OpenAI had assets of just over 50 billion, around half of which was in cash. This cushion is absorbing the high annual losses for the time being.

OpenAI is also investing enormous amounts of capital in data centres to remain at the forefront of the AI race. By 2030 investments are expected to total a figure in the high three-digit billions. The lion’s share goes to chip manufacturers such as Nvidia and memory producers such as Samsung and SK Hynix. Because these companies are shifting their capacity to meet this demand, hardware for end consumers is becoming increasingly expensive and is making little progress.

Huge losses

High losses are not unusual during growth phases. Amazon and Tesla were in the red for years before becoming sustainably profitable. However, the AI industry’s infrastructure – comprising data centres, fibre-optic networks, energy supplies and specialised semiconductors – is vastly more expensive than the expenditure of earlier start-ups. Amazon lost a total of 2.8 billion over its first 20 years before turning a profit. Tesla took 17 years and incurred total losses of 6 billion.

That is pocket change compared to OpenAI’s 21 billion dollars in a single year. In the AI industry, spending is generally on a massive scale. The hyperscalers – Amazon, Alphabet, Meta, Microsoft and Oracle – together invest several hundred billion dollars a year in data centres. Estimates suggest that cumulative expenditure will soon reach 1.5 trillion. That is more than Switzerland’s gross domestic product.

Although these corporations are still posting profits, the majority of these come from other business areas. There is also criticism that they are currently whitewashing their figures in two ways: Firstly, they are extending the alleged lifespan of their infrastructure, thereby reducing depreciation in the short term. Secondly, they hold shares in private start-ups, which are being valued ever higher. However, these are merely paper profits that will vanish into thin air if these companies do not succeed in the long term.

OpenAI needs fresh capital

AI companies are hoping that the investments will pay for themselves in the long term through efficiency gains and new business areas. The costs of research and development could, in theory, fall, thereby paving the way to profitability. However, due to fierce competition, the opposite has been the case so far, as companies seek to outdo one another with ever-better models.

For the time being, therefore, OpenAI needs fresh capital soon. Having relied on private investors for years, this source appears to be slowly drying up. That is why the company is planning its IPO later this year – just like its rival Anthropic. Meanwhile, SpaceX celebrated an extremely successful first day on the stock market on 12 June and is currently valued at roughly the same level as Amazon. This is despite the fact that it, too, is racking up significant operating losses. In 2025, the AI division alone posted a loss of 6.4 billion dollars.

Header image: Shutterstock

13 people like this article


User Avatar
User Avatar

My fingerprint often changes so drastically that my MacBook doesn't recognise it anymore. The reason? If I'm not clinging to a monitor or camera, I'm probably clinging to a rockface by the tips of my fingers.


News + Trends

From the latest iPhone to the return of 80s fashion. The editorial team will help you make sense of it all.

Show all

These articles might also interest you

6 comments

Avatar
later