Sam Altman’s company will have to raise new record funds while justifying its colossal losses, while it faces increasingly serious competition.
Make it or break it: As the new year begins, Sam Altman’s company finds itself at a crossroads, facing at least as many challenges as opportunities. It is entirely possible that it will end the year having consolidated its progress on generative AI, increased its revenues by developing its offerings to businesses and successfully completed its IPO. But for some skeptics, the idea of a loss of momentum, or even a bankruptcy of the company, cannot be ruled out either.
A new fundraising campaign in a difficult context
In 2026, one of the challenges for the parent company of ChatGPT will be to raise new funds. Figures leaked to the press indicate that it plans to burn $17 billion in cash this year, almost double last year’s $9 billion. And OpenAI has no intention of calming down. In 2025, it has committed to investing the colossal sum of $1.4 trillion over the next few years in several major projects, including the Stargate program, led by the Trump administration.
After raising $40 billion in March 2025, the largest sum ever raised from investors by a private company, OpenAI plans to beat its own record in 2026: this time the company would like to raise no less than 100 billion to meet its needs.
But OpenAI will not be the only one wanting to raise funds (xAI has just raised $20 billion), in a context where a growing number of mature start-ups (Databricks, SpaceX, Anthropic, etc.) are delaying their entry into the public markets.
A recent study from the University of Florida indicates that between 1980 and 2024, the average age of a company going public has more than doubled. Data collected by the investment bank Renaissance Capital indicates that the median age of a private company was 13 years in 2024 compared to 10 in 2018.
However, the abundance of private capital is not inexhaustible. After a sustained increase of 10% per year worldwide between 2012 and 2021, assets under private management have since stabilized at just over $20,000 billion. OpenAI has so far had no problems attracting investors, and Amazon is already in discussions to invest ten billion dollars in the company. But the amount she intends to raise is unprecedented and it is not certain that she will achieve it without difficulty.
“Since they are losing money and can’t stand on their own for long, they find themselves at the mercy of investors. If the situation continues to deteriorate, raising the next round of funding will be much more difficult than last time. Few investors have deep enough pockets to provide OpenAI with what it needs to stay afloat. If most, if not all, pull out, OpenAI’s valuation could fall, perhaps precipitously, as it did the case for WeWork, whose valuation quickly fell from its peak of $47 billion to near bankruptcy,” estimates Gary Marcus, an American AI expert, on his Substack.
Another option would be to go public, a possibility regularly mentioned by Sam Altman, but which also has its own difficulties. On the one hand, the company would have to manage to resolve the legal puzzle that opposes it to Microsoft. On the other hand, the IPO market is only just beginning to recover from several years of drought. An IPO would therefore not be without risk for Sam Altman’s company.
An economic model in question
Corollary to this first problem, OpenAI still has to answer questions about its economic model. Certainly, the company can boast impressive numbers over the past year: its number of weekly active users tripled, from 300 million to 900. Its annualized revenue rate increased from around $6 billion to $19 billion during the same period.
However, the company will have to wait until 2030 to generate profits, according to a recent study by HSBC bank. In addition, its computing power requirements, which represent by far the largest share of its expenses, remain closely correlated to its turnover, offering for the moment little prospect of a rapid increase in profits. OpenAI’s computing power needs have increased from 200 megawatts in 2023 to 1.9 gigawatts in 2025, and the company plans to add an additional 30 gigawatts over the next few years.
Even more problematic, OpenAI would lose money just by running its models: figures leaked to the press show that OpenAI’s inference costs exceeded its turnover in the first half of 2025. This means that the company will undoubtedly quickly be forced to increase its prices, which could slow down its growth and benefit its competitors, thus harming its future giant fundraising or its IPO…
Sam Altman’s company is studying several avenues to diversify and increase its revenues. A first channel consists of revolutionizing online commerce via generative AI. It now allows companies, including online craft sales company Etsy, e-commerce platform Shopify and retail giant Walmart, to sell products through ChatGPT in exchange for a commission.
Another is to accelerate the adoption of its technology in companies: the company has set up an entire division whose goal is to help large companies use its products, and deployed tools like AgentKit. Launched in October, it aims to help professionals adopt and deploy agentic AI.
Finally, the pioneer of generative AI has begun a vertical integration strategy. Following the path taken by Google with its TPUs, it signed an agreement with Broadcom to develop its own chips and recruited Sir Jony Ive, the former star designer at Apple, to study the design of a consumer device adapted to generative AI, which would make audio the main method of interaction, in front of the screen. 2026 will therefore serve as a test to determine whether the different parts of this strategy are paying off.
Too big to fail?
Finally, 2026 will also be an opportunity to determine whether OpenAI can resume its technological progress, after the solid offensive from Google, whose latest version of Gemini greatly impressed. Here again, if the company fails to reestablish its lead over its competitors, it will become difficult for it to raise ever larger sums from investors. At the same time, open models, particularly those from China, also display formidable performance, likely to worry OpenAI.
Note that the fact that a company loses a lot of money for years is nothing new across the Atlantic: from Netflix to Uber via Facebook and Amazon, many companies in the internet wave have accumulated losses for a long time before becoming cash machines. But with OpenAI, the volumes are such that the situation is unprecedented. Given the web that the company has woven around itself, it is also quite possible that OpenAI has simply become “too big to fail”.
